Final Project Reports

ESCMID AMS Certificate

Cohort 2022-2024

1
Table of Contents

An implementation of an Antimicrobial Stewardship Intervention regarding the use of antibiotic prophylaxis as

routine in Paediatric Patients after the insertion of an in-and-out urinary catheter .................................................. 3
Anastasia Izampella Papathanasiou

Optimization of the empirical antimicrobial therapy protocol in haematological patients with febrile neutropenia . 13
Ander González Sarria

Implementation of a bedside consultation for Staphylococcus aureus bacteremia ............................................... 22

Carlos Tarrazo Tarrazo

Antifungal stewardship using 1,3 Beta D Glucan and Candidemia guideline .......................................................... 27
Cătălina Hăpăianu

Reducing empirical Ceftriaxone use for CAP in the Internal Medicine department ................................................. 31
Francisco Miguel Lima Rosa Mendonça e Almeida

Third-day AMS interventions in a medical ward .................................................................................................... 37

Joana Catarina Gouveia Batista Garnel

Improving appropriate empiric antibiotic treatment through the use of antibiotic standardized treatment packages
.......................................................................................................................................................................... 46
Kristina Öbrink-Hansen

Implementation of local guidelines of Staphylococcus aureus bacteremia and ID consult as a standard of care in
University Hospital Bulovka ................................................................................................................................ 52
Kristýna Herrmannová

Antimicrobial de-escalation in patients with Enterococcus faecalis monomicrobial bacteremia ........................... 62

Marco Ripa

Optimizing blood culture collection practice ....................................................................................................... 79

Mona Mustafa Hellou

Culture matters: a multimodal behaviour change intervention to improve the quality of blood culture collection in
an academic emergency medicine department ................................................................................................... 85
Oana Joean

Blood culture diagnostic intervention in time of earthquake catastroph to improve diagnostics and therapy .......... 90
Osman Sezer Cirit

Appropriateness of antibiotic use at day 3-4 d of admission ................................................................................. 96

Shakil A. Shakar

Adequacy of d e - escalation therapy of bloodstream infections of a tertiary hospital: the impact of an Antimicrobial
Stewardship Intervention .................................................................................................................................. 109
Susana Ramos Oliveira

Improved usage of standard treatment packages for community-acquired pneumonia in the emergency department
......................................................................................................................................................................... 123
Susanna Deutch

Reducing the use of Piperacillin/Tazobactam for common infections – a Danish quality improvement study .........130
Trine Langfeldt Hagen

2
ESCMID AMS Certificate Program
My Project Report

Name
Anastasia Izampella Papathanasiou
Institution
Hippokration General Hospital of Thessaloniki
Country
Greece
Tutor
Dr Zoi Dorothea Pana
Title of Project
An implementation of an Antimicrobial Stewardship Intervention regarding the use of antibiotic prophylaxis as
routine in Pediatric Patients after the insertion of an in-and-out urinary catheter
Date of submission
17/03/2024

Abstract
Background
An antimicrobial stewardship intervention was implemented in a General Pediatric Department of 33 beds by a
resident Pediatrician and a Pediatric Infectious Diseases Specialist. The objective of this intervention was to
change current practice in prescribing antibiotic prophylaxis as routine after in-and-out bladder catheterization
for urine sample collection in infants and children older than two months, when a bacterial infection is ruled
out.

Methods
A before and an after intervention measurement was conducted in the same 4month period (November -
February). For the exploration of barriers and facilitators, interviews of the prescribers were performed, using
the Flottorp framework as reference. An intervention technique was chosen to address each barrier in
accordance with the COM-B model. The intervention was implemented during a 7 months period.

Results
The results of the measurement showed that before the intervention, the compliance to the recommendation
of not prescribing antibiotics as prophylaxis when performing in-and-out catheterization was 0% in a total of
10 cases, while after the intervention the compliance to the recommendation increased to 81,8% in a total of
11 cases. Before the intervention, the percentage of antibiotic prescription after urinary track catheterizations
was 100% in a total of 13 cases, while after the intervention the percentage dropped to 52,4% in a total of 21
cases.

Conclusions
A multifaceted antimicrobial stewardship intervention resulted in high compliance with the no routine
prescription of antibiotic prophylaxis in pediatric patients who undergo in-and-out bladder catheterization for
urine sample collection and had no clinical or laboratory findings of bacterial infection.

Introduction: description of the setting and the AMS program in place and identification of the problem
The hospital I work is a tertiary University hospital of 900 beds. At the beginning of my intervention I was a
resident Pediatrician in my last year of my residency in the 1st Department of Pediatrics. The department has 33
beds and is the center of reference for pediatric nephrology patients in northern Greece.
In the hospital, there is a Pediatric Ward which consists of two General Pediatric Departments, one Pediatric
ICU, two neonatal ICU, an Oncology Pediatric Department and a Pediatric Department for Infectious Diseases.
In the Pediatric Ward there are 7 Infectious Diseases Specialists Pediatricians. They are in charge of the
Pediatric Department for Infectious Diseases and do consultations in the Oncology Pediatric Department, the
Neonatal and Pediatric ICU on a daily to weekly basis but also as in the adults’ departments when consultation
is asked. Their routine includes infection prevention and antimicrobial stewardship during their consultation.

3
By the time I initiated my intervention, there was no organized antimicrobial stewardship program, as such. My
position as a resident Pediatrician made it impossible for me to form an AMS team, resulting in my deciding to
implement an intervention only in the Department, where I worked. I was also led to this decision due to the
fact that one consultant of mine, a Pediatrician and Infectious Diseases specialist, working both in the 1st
Pediatric Department and the Pediatric Department for Infectious Diseases was really keen to help me
implement a stewardship intervention in our department.
One of the problems regarding antimicrobial stewardship in my department is the standard use of antibiotic
prophylaxis, when performing in-and-out urinary catheterization for urine sample collection. This practice is
contradictory to guidelines.
In the emergency department or when a child is hospitalized, collection of urine sample for microscopic
examination or urine culture by in-and-out bladder catheterization is useful for the diagnosis or the exclusion
of a urinary tract infection in children who cannot urinate freely because of age or due to other comorbidities.
The catheterization is a more preferable method than the placement of a urine bag, since the technique is
sterile and therefore the contamination rates of the samples much lower. Catheterization is also used as an
alternative to suprapubic aspiration, which is the gold standard for the collection of urine sample in infants,
when the second one has failed.
In the case of the 1st Department of Pediatrics, when urinary tract infection is ruled out by the urine
microscopic examination and by additional clinical and laboratory findings, prophylaxis is always given after the
catheterization, even in a case of a sterile and non-traumatic one. The course of antibiotics lasts usually three
days, which means that the regimen is actually prescribed as empiric therapy for urinary tract infection.
Interestingly, in the other department of General Pediatrics and in the Pediatric Department for Infectious
Diseases, when an in-and-out urinary catheter is used and a urinary tract infection is ruled out, no routine
prophylaxis is prescribed. This difference in practice between the departments of the same hospital was one of
the main reasons which motivated me to choose this problem for my intervention, since I could share this
experience to persuade my consultants not to use antibiotics as routine prophylaxis.

Measurement
To evaluate the use of prophylaxis when performing in-and-out catheterization, I conducted a 6 month
retrospective audit and a 4 month prospective one, before the intervention. The population I included was
infants and children older than 2 months of age, who were catheterized by the use of an in-and-out urinary
catheter for the exclusion of a urinary track infection and there was no other indication for receiving antibiotics
(normal urine microscopic examination, no signs of sepsis, no signs of another bacterial infection, CRP <
30mg/L, sterile and non-traumatic procedure, possible detection of a viral pathogen (e.g. RSV, Covid-19) ,
possible documentation that the antibiotics were given only due to catheterization).
In our hospital, there is no electronic patient files system. In the Pediatrics Department, for every patient that is
discharged from the hospital, a report is written and given to the parents. A record of this report is saved in our
personal electronic files.
The retrospective audit was conducted for the period between May and October 2022. For the retrospective
part, I went through all the admissions we had in our department. If the reason of admission and the age of the
patient were relevant, I checked the report that was written at the time of the discharge of the patient and also
the laboratory findings of the patient in the electronic system. When the report was ambiguous about the
reason of prescribing antibiotics in cases of patients who were catheterized, I went manually through the
patient record files. Furthermore, I checked manually all the patients who were examined in the emergency
department during this six month period and there was a documentation of a urine sample collection.
The prospective audit was conducted between November 2022 and February 2023. For this phase, I informed
all my resident colleagues in regards of my measurement. In case a urinary track catheterization was
performed either in the emergency or in the Pediatrics department, I should be notified. Afterwards, I was
going through the patient record file and the report.
The initial plan was to conduct only a retrospective audit for the measurement before and a prospective audit
after the intervention. However, the period I chose for the retrospective phase is a period the incidence of
infections is low, so the need for differential diagnosis of infection and consequently the diagnostic
investigation occurs more rarely. Furthermore, I found out that in cases of a urine sample collection in the
emergency department, it was not always documented the technique used (urine bag, urinary track
catheterization, suprapubic aspiration), a fact that made the assessment of the appropriateness of antibiotic
prescription difficult. This led me to conduct an additional prospective audit. In this report, I include only the
results from the prospective audit, which can be comparable to the results after the intervention.

4
In the population of patients I mentioned, I assessed whether they received antibiotics after the urinary track
catheterization or not, the days of the antibiotic course, the choice of the antibiotic, the dose and the reason of
diagnostic investigation with urine sampling.
I used as primary quality indicator the compliance to the recommendation of not using antibiotic prophylaxis as
routine after urinary catheterization in patients with no indication for receiving antimicrobials. I
operationalized the quality indicator by using a fraction with numerator: the number of patients who did not
receive antibiotics after the urinary track catheterization and with denominator: the number of patients who
were catheterized and there was no indication for receiving antibiotics according to the criteria I mentioned
before. I used an additional quality indicator to avoid the bias in the measurement after the intervention,
which is the prescription of antibiotics in patients who underwent a urinary track catheterization. The
numerator of this quality indicator would be the patients who received antibiotics after the catheterization and
the denominator, all the patients who were catheterized. Consequently, patients who had a urinary tract
infection or sepsis were also included in this quality indicator.
Since antibiotic prescribing is a standard procedure in our Department after every urinary track catheterization,
the rate of not prescribing antibiotics was 0%, as expected, in a total of 10 patients (graph 1). Moreover, the
mean duration of prophylaxis was 3,35 days (graph 2). In 86% of the cases the prescription of antibiotics was
stopped by the time the results of the urine culture were available. In all cases the prescribed antibiotic was a
second generation cephalosporin. When given intravenously, it was cefuroxime and when orally cefaclor or
cefprozil, always in a therapeutic dosage (graph 3). The percentage of patients who received antibiotics after
urinary track catheterization including those with a bacterial infection was 100%, as expected (graph 4).

Barriers and facilitators

For the determination of the barriers and facilitators, I performed individual interviews with all the consultants
who prescribe antibiotics both in the emergency and the General Pediatric Departments, including the
Infectious Diseases Specialist, with whom I collaborated for the implementation of my intervention. The
interviews were not strictly structured and had more a character of a discussion, in which I was also presenting
the content of the ESCMID certificate I attend, the need of antimicrobial stewardship in Greece, the goal of my
stewardship intervention and the results of my measurement. I conducted in total 8 interviews. I used the
Flottorp framework as reference.
All of the barriers the consultants and the Infectious Diseases specialist addressed, referred to 3 domains of the
Flottorp checklist. That is factors related to guidelines, factors related to individual healthcare professionals
and factors related to professional interactions.
Regarding individual healthcare provider factors: (1) Prescribing antibiotics after every urinary tract
catheterization is a habitual behavior which is reproduced over years in the department and therefore difficult
to change.
Regarding professional interactions: (2) The standard use of prophylaxis was applied in the department by the
former chief, who was a Pediatric Nephrology Professor and is maintained by the succeeding team of Pediatric
Nephrology Professors with whom he was collaborating. This results in the hesitation of the consultants to act
differently, since this may lead to a conflict between them and the Pediatric Nephrologists. The Infectious
Diseases Specialist in fact mentions that in the past did argue with them regarding the prophylaxis, but never
managed to persuade them. (3) In the procedure of a urinary catheterization, following personnel is involved:
one or two nurses who perform the disinfection before the insertion of the catheter and a physician, in most
cases a resident doctor. However, the prophylaxis is always prescribed by the consultant. Some consultants
mentioned that they do not always trust the technique of their colleagues, if they are not present at the
procedure, mostly of the nurses. (4) There is no other consultant in the department who does not prescribe
prophylaxis.
And one last barrier related to guideline factors: (5) One consultant had doubts about the appropriateness of
the used disinfection agent and consequently about the compatibility of the guideline with the current
practice.
As far as the facilitators are concerned, following was determined: (1) there is no guideline supporting
prophylaxis. (2) All of the consultants were of the opinion that the current practice is wrong. They had
knowledge of the recommendation; some of them also mentioned that standard use of prophylaxis is not
indicated even when using indwelling catheters. All of them trusted the recommendation and all of them
agreed that using less antibiotics is beneficial both for the patients and for the environment. (3) The Infectious
Diseases Consultant could support the intervention by communicating the wrong practice in the ward among
the consultants. (4) All resident doctors where very willing to help in the implementation of the intervention.

5
Development and execution of the stewardship intervention
For the development of the intervention, I used the Capability Opportunity Motivation – Behavior (COM-B)
Model. At first I associated the barriers I found with the three components of the COM-B model and then
selected an appropriate intervention to address each barrier.
Before starting measuring, I asked permission from the chief of the Department. After Ι completed my
measurement, I communicated my results and reassured again that he consents to my intervention.
To address (1) the habitual behavior of prescribing prophylaxis after every urinary catheterization, during each
meeting we had after admitting patients in the department, in case of an antibiotic prescription due to
catheterization, a discussion was initiated by the ID specialist about the inappropriateness of using routine
prophylaxis (automatic motivation - persuasion). This intervention took place during a seven month period
between the measurements before and after.
To address (2) the belief that the Nephrologists oppose to the recommendation, I arranged a separate meeting
with each of the two Pediatric Nephrology Professors of the Department. I explicitly explained my project and
communicated the fact that I had permission from the chief of the department to implement my intervention. I
informed them that my intervention is supported by the Infectious Diseases Specialist and I also stressed that
the 3d Department of Pediatrics and the Department of Pediatric Infectious Diseases in our hospital follow a
different practice regarding the post catheterization prophylaxis. Furthermore, I explained that in my project
children with indwelling catheters or children undergoing other surgical or diagnostic procedures (e.g.
cysteography) are not included. Both of them stated, that they totally agree with not using routine prophylaxis
for in-and-out urinary catheterization. We also discussed the duration of prophylaxis in complicated cases,
traumatic or non-sterile catheterization and agreed that a shorter course should be considered (social
opportunity-restructuring of the social environment).
To address (3) the potential omissions in the technique of nurses and resident doctors, we organized an
instructional course on how to perform step by step a sterile urinary track catheterization. In the organization
of the course, the supervising nurse of the department and my ID consultant were involved (psychological
capability - education-training) The course took place right before the second measurement.
To address (4) the peer pressure of colleagues, the ID consultant started performing herself all of the urinary
tract catheterization and did not prescribe antibiotics as prophylaxis. (social opportunity – modelling).
To address (5) the possible inappropriateness of the disinfection agent, I looked thoroughly the literature and
found out that there was no evidence against the use of it. Subsequently, I shared the knowledge with the
consultant and also during the instructional course (psychological capability - education).
After my meeting with the Nephrologists, I organized a second round of discussions with the 7 consultants, in
which I reminded them of my project and shared the opinions of the Nephrology Professors regarding the use
of prophylaxis. I mentioned the attitude of the other two Pediatric Departments and also talked to them about
my intention to organize a course for the nurses and the resident doctors regarding the sterile technique for
insertion of a urinary tract catheter. I repeated a separate meeting with them after the conduction of the
course.
As far as my resident doctor colleagues are concerned, I organized three meetings with all of them, in which I
presented my project, the practice of the Department regarding the use of prophylaxis, the practice of the
other two Pediatric Departments and the problem with AMR. The meeting took place before and after the
initial measurement and before the post intervention one. (table 1)

Evaluation of the stewardship intervention

To evaluate my intervention, I conducted a 4 month prospective audit between November 2023 and February
2024. The percentage of cases in which no prophylaxis was given after a urinary track catheterization increased
from 0% to 81,8% (graph 5). From the total of 11 patients with no indication for receiving prophylaxis only 2
received antibiotics (graph 6). The percentage of cases in which antibiotics were given after the catheterization
divided to the total amount of patients who underwent a urinary track catheterization decreased from 100% to
52,4% (graph 7). From the total of 21 patients who underwent catheterization, 11 received antibiotics (graph
8). The prescribed antibiotics were for the one case cefuroxime iv and for the second cefaclor orally both for a
3 days course, in a therapeutic dosage.
Regarding the evaluation of the implementation, 85% of the resident doctors attended the educational course
and only 30% of the nurses.

6
Conclusions
A multifaceted antimicrobial stewardship intervention resulted in high compliance with the no routine
prescription of antibiotic prophylaxis in pediatric patients who undergo in-and-out bladder catheterization for
urine sample collection and had no clinical or laboratory findings of bacterial infection.
The implementation of the project was constant work during everyday practice. Having good relationships with
my colleagues was the most important facilitator. I found out that the lack of electronic patient files is a major
barrier for the extraction of data when conducting a retrospective audit. This means that measuring date is a
challenge in our setting.
My position as a resident Pediatrician hindered me from implementing a broader intervention in the hospital.
Nevertheless, through my engagement in the AMS Certificate, I had the chance to be involved in the
organization of the Drive-AMS course in Thessaloniki, Greece. This led to the formation of an Antimicrobial
Stewardship Team in the Hospital that I am a part of and we are already implementing an intervention
including both Pediatric and Adult wards.
As far as my Department is concerned, in collaboration with my Infectious Diseases Consultant we are planning
to have further educational courses regarding Infection Prevention, complementary to the course we organized
for the sterile technique of urinary track catheterization.

Acknowledgments
I would like to thank my Consultant Pediatric Infectious Diseases Specialist, Papadimitriou Eleni, for her support
and devotion to the project and Dr Zoi Dorothea Pana, who indeed supported the choice of my stewardship
intervention. I would also like to thank the resident doctors of my department who helped me with the
measurement and the implementation.

References
Raimund Stein a, *, Hasan S. Dogan b , Piet Hoebeke c , Radim Kocˇvara d , Rien J.M. Nijman e , Christian Radmayr f , Serdar
Tekgu¨l b, Urinary Tract Infections in Children: EAU/ESPU Guidelines, EUROPEAN UROLOGY 67 (2015) 546–558

James Diviney1 & Mervyn S. Jaswon1,2, Urine collection methods and dipstick testing in non-toilet-trained children,
Pediatric Nephrology (2021) 36:1697–1708

Guidelines for the Prevention of Catheterassociated Urinary Tract Infection Published on behalf of SARI by HSE Health
Protection Surveillance Centre 2011

Carolyn V. Gould, MD, MSCR 1; Craig A. Umscheid, MD, MSCE 2; Rajender K. Agarwal, MD, MPH 2; Gretchen Kuntz, MSW,
MSLIS 2; David A. Pegues, MD 3 and the Healthcare Infection Control Practices Advisory Committee (HICPAC) 4, GUIDELINE
FOR PREVENTION OF CATHETERASSOCIATED URINARY TRACT INFECTIONS 2009,
https://www.cdc.gov/infectioncontrol/guidelines/cauti/

Signe A Flottorp 1, Andrew D Oxman, Jane Krause, Nyokabi R Musila, Michel Wensing, Maciek Godycki-Cwirko, Richard
Baker, Martin P Eccles, A checklist for identifying determinants of practice: a systematic review and synthesis of frameworks
and taxonomies of factors that prevent or enable improvements in healthcare professional practice, Implement Sci. 2013

A guide on The COM-B Model of Behaviour, https://social-change.co.uk/files/02.09.19_COM-

B_and_changing_behaviour_.pdf

7
Graphs and Tables

Patients with no indication for

prophylaxis
10

No prophylaxis Prophylaxis

Graph 1. In a total of 10 children who were catheterized and had no indication for receiving prophylaxis, all of
them received antibiotics.

Days of prophylaxis
10
9
8
7
6
5
4
3
2
1
0 2 4 6 8 10 12
Graph 2. The median duration of antibiotic prophylaxis in patients who underwent a urinary catheterization
was 3,32 days.

8
 Choise of antibiotic
Cefuroxime Cefaclor Cefprozil

10%
10%

80%

Graph 3. In a total of 10 patients, 8 received cefuroxime intravenously, 1 cefaclor and 1 cefprozil orally.

Patients who underwent

catheterization
13

No antibiotics Antibiotics

Graph 4. In a total of 13 patients who underwent urinary catheterization, including patients with bacterial
infection, all of them received antibiotics.

9
 Sep Oct Nov 22– Mar Apr May Jun Jul Aug Sep Oct Nov 23-
22 22 Feb 23 23 23 23 23 23 23 23 23 Feb 24
Meeting with chief of x
the Department
Retrospective audit x
Prospective audit x
Interviews x
Meeting with resident x x x
pediatricians
Meeting with the x
Professors of the
Nephrology
Meeting with the x x
consultants
Intervention of the ID x x x x x x x
specialist
Educational course x
Prospective x
remeasurement
Table 1. The timeframe in which every intervention took place

Compliance to the recommendation

90%
80% 81.80%
70%
60%
50%
40%
30%
20%
10%
0% 0%
Nov 22-Feb 23 Nov 23-Feb 24

Graph 5.

10
 Number of patients
11
10
9
8
7
6
5
4
3
2
1
0
No prophylaxis Prophylaxis

Graph 6. In a total of 11 patients who underwent urinary catheterization with no indication for receiving
prophylaxis, only 2 received antibiotics.

Prescription of antibiotics
120%

100% 100%

80%

60%
52.40%
40%

20%

0%
Nov 22-Feb 24 Nov 23-Feb 24

Graph 7. Prescription of antibiotics in patients who underwent urinary track catheterization

11
 Number of patients
11
10
9
8
7
6
5
4
3
2
1
0
Antibiotics No antibiotics

Graph 8. In a total of 21 patients who underwent a urinary tract catheterization, 11 received antibiotics.

12
ESCMID AMS Certificate Program
My Project Report
1. Introduction
2. Measurements
3. Barriers and facilitators
4. Development and execution of the stewardship intervention
5. Evaluation of the stewardship intervention
6. Conclusion
7. Acknowledgements
8. Supplementary material
a. Definition of febrile neutropenia
b. Local protocol for empirical antimicrobial treatment in febrile neutropenia (2018)
c. MASCC Score
d. Bibliography
Name
ANDER GONZÁLEZ SARRIA
Institution
HOSPITAL UNIVERSITARIO PRÍNCIPE DE ASTURIAS
Country
SPAIN
Tutor
BHANU SINHA
Title of Project
OPTIMIZATION OF THE EMPIRICAL ANTIMICROBIAL THERAPY PROTOCOL IN HAEMATOLOGICAL PATIETS WITH
FEBRILE NEUTROPENIA
Date of submission
Wednesday, March 6, 2024

Abstract* (max 300 words)

Introduction: Febrile neutropenia is a common complication in onco-haematological patients. In our centre,
protocols for its empiric antimicrobial therapy include antipseudomonal agents as first-line choices.
Objective: To change the first-line options in the protocol for empirical antimicrobial treatment in patients with
febrile neutropenia hospitalized in Haematology, from antipseudomonal carbapenems to antipseudomonal
cephalosporins, based on local Pseudomonas aeruginosa epidemiology.
Methods: A lecture was offered to the Haematology department, including presentations about local
Pseudomonas aeruginosa epidemiology (cumulative antibiograms, isolation sources and rate of bacteriemias),
consumption of the most frequently used antipseudomonal betalactams (Meropenem, Imipenem, Piperacillin-
tazobactam, Ceftazidime, Cefepime) in DDDs/100 hospitalization days in Haematology, as well as the key
aspects of the protocol used in a different centre with a similar epidemiology. For the calculation of statistical
significance of differences in total mortality rates between study periods, Fisher’s exact test was used.
Results: Compliance to the updated protocol was measured comparing the DDDs/100 hospitalization days for
rd th
the mentioned antipseudomonal betalactams, at the end of the 3 month and at the end of the 6 month,
after the intervention. A descent was observed in the consumption of Meropenem (DDDs from 18.43 pre-
rd th
intervention, to 9.6 at the 3 month and 10.61 at the 6 month measurements), as well as a rise in Cefepime
rd th
(DDDs from 0.48 pre-intervention, to 3.29 at the 3 month and 4.69 at the 6 month measurements) and
rd th
Piperacillin-tazobactam (DDDs from 3.09 pre-intervention, to 0.15 at the 3 month and 4.33 at the 6 month
measurements). The cumulative susceptibility report for Pseudomonas aeruginosa showed no remarked
variations between study periods. Total mortality rates showed no statistically significant differences between
study periods.
Conclusions: Local epidemiology concerning antimicrobial susceptibility can vary over time, and it is important
to pass on updated data concerning optimal treatment choices to antimicrobial prescriber clinicians.

Stage 1 Introduction
The hospital where I work, Hospital Universitario Príncipe de Asturias, is a secondary hospital with around 500
beds, located in Alcalá de Henares, one of the mayor cities in the peripheral area of Madrid Capital City,

13
accounting for nearly 200.000 inhabitants in recent years. I work at the Department of Clinical Microbiology,
and my responsibilities include areas of bacteriology (urine and blood cultures, and exudates) as well as
antimicrobial susceptibility testing.
The Department of Haematology consists of 10-15 specialists and 8 medical residents, as well as nurses and
administrative staff. The hospitalization rate ranges between 5-15 beds (of a total of 20), for which two of the
specialists are the main responsible. However, during on-call shifts, other specialists (from a group of 8) and the
residents can be responsible for hospitalized patient care, from their admission to the Emergency Department.
The most frequent motives for hospitalization include chemotherapy, acute complications of haematological
diseases (oncological or not) and infections. The last most frequently occur in patients during of between
chemotherapy sessions. Only autologous bone marrow transplant is performed in this department (not
heterologous).
Many clinical guidelines agree that Pseudomonas aeruginosa is an important pathogen to be empirically
covered in febrile neutropenia in settings where its infections are epidemiologically important. Its cumulative
susceptibility report is a crucial standard to be considered in its empirical antimicrobial management.
During the observation period (September 2022), by routinely working with Pseudomonas aeruginosa strains
isolated from different sample types, coming from any medical or surgical department, I observed that I was
finding more strains susceptible to antipseudomonal cephalosporins than to carbapenems. Meanwhile, I also
observed that patients admitted to the Department of Haematology would be treated with carbapenems (with
or without other agents like teicoplanin, amikacin, …) most of the times. As this was only an observation, I
considered it as a potential problem that could be addressed with an AMS intervention, if the measurements
proved it right.

Stage 2 Measurement
For the description of the general frame, a retrospective study with data from June 2022 to May 2023 was
performed, extracted from the electronic patient file and microbiology databases of blood cultures collected
from any haematological patient either during hospitalization or at the Emergency Department. Although the
total number of patients hospitalized in Haematology during this full study could not be searched, the included
blood-culture sets belong to 68 different patients, 9 of which (13%) presented at least two different episodes in
which blood cultures were requested.
Total of blood-culture sets 185
Positivity rate 24%
Neutropenia 61% (83.6% of which was severe)

3% Isolates
1%
10% Coagulase negative staphylococci
Enterobacterales

29% Enterococci
57%
Listeria monocytogenes
Pseudomonas aeruginosa

The distribution of isolates in blood cultures taken at the Department of Haematology showed a notable
percentage of potentially contaminant species, for which a different intervention was conducted. Moreover,
Pseudomonas aeruginosa did not appear to be a frequent etiology in bactaeremias. No bactaeremias due to
Staphylococcus aureus were reported.
To understand the local epidemiology of Pseudomonas aeruginosa, a retrospective study with data from June
2022 to May 2023 was performed, extracting the information from the microbiology databases and containing
isolates from every department (n=533):

14
 3% Sample type
4% Urine
Non-respiratory exudates
18% 40%
Respiratory
35%
Blood
Rectal (MDR carriage)
-Cumulative susceptibility profile of routinely tested antipseudomonal betalactam antimicrobials in isolates of
Pseudomonas aeruginosa from the 12 months prior to intervention, excluding rectal swabs for MDR carriage.
Wild-type to all antipseudomonal betalactams 75.04%
Piperacillin/tazobactam 82.2%
Ceftazidime 87.2%
Cefepime 87.2%
Imipenem 85.03%
Meropenem 85.8%
Ceftolozane/tazobactam 100%
Co-resistance phenotypes
Piperacillin-tazobactam & Antipseudomonal cephalosporins 6.73%
Antipseudomonal carbapenems 9.42%
All agents excluding Ceftolozane-tazobactam 9.23%
Cefepime & Meropenem 5.96%
About 25% of isolates showed at least one resistance phenotype, and co-resistance rate was highest between
antipseudomonal carbapenems. Although no carbapenemase enzyme producing isolates were found in this
period, the cumulative susceptibility profile of the health area matched the recent Spanish nationwide surveys
on Pseudomonas aeruginosa antimicrobial resistance mechanisms and epidemiology (1), in terms of showing
that, on a national epidemiological base, it is common to find better cumulative susceptibility profiles for
antipseudomonal cephalosporins versus antipseudomonal carbapenems, suggesting that the first could be
considered as empirical options for infections that potentially include Pseudomonas aeruginosa as their
etiology.
Finally, to better understand the comparative use of different antipseudomonal agents
(piperacillin/tazobactam, ceftazidime, cefepime, imipenem, meropenem), cumulative data provided by the
Hospital Pharmacy department was studied. The following list corresponds to the use of the antibiotics in a 6
month-period previous to the intervention date (end of May 2023), measured in DDDs/100 hospitalization days
in Haematology. In this period, 170 patients were hospitalized in Haematology, with 17 deceases (10%).
Piperacillin/tazobactam 3.09
Ceftazidime 0.91
Cefepime 0.48
Imipenem 0.31
Meropenem 18.43
This data suggest a comparative overuse of Meropenem.

Stage 3 Barriers and facilitators

An individual approach strategy was first used, based on a short informative reunion with the two specialists at
the Department of Haematology who are most in charge of the hospitalized patients. The objective was to have
a better understanding of the reason why the protocol, designed by them and approved by the Infection
Committee at the hospital, was apparently not being followed. The approach was perceived quite negatively,
but we agreed that a formal reunion with the exposure of objective data would be helpful in evaluating if any
chage was needed. With that first step, I picked a “champion” inside the Haematology team, who also showed
concern over the high consumption of certain antimicrobials in the department, based on the knowledge they
had recently gained during a clinical stage rotation with an expert in Infectious Diseases and Antimicrobial
Stewarship in immunosupressed patients. A second date was agreed for a formal presentation (the
intervention), explained in Stage 4.

15
These reunions were also used as the chance to evaluate some of the barriers, here explained. Those barriers
were later classified using the COM-B Model of Behaviour, to have a better understanding of how to work with
them.
Barriers:
Organizational factors:
- A team of AMS existed at the start of the observation period, and it included several internal
medicine specialists (at least three), two clinical microbiologists, and two pharmacists. Since the beginning of
the COVID-19 pandemic, the team had lowered the active participation in the usual activities of AMS, mainly
due to work overload and not being able to meet in person regularly. Most of the patients treated with
antibiotics, including the haematological ones, would normally not be followed by a team of AMS. This could
translate into an over-use of broad spectrum antibiotics, empirically or/and in directed therapy. The
responsible haematologists did not believe they were choosing the “wrong” antibiotics since “nobody from the
AMS team had told them so”.- A local antibiotic guideline for febrile neutropenia (cf. supplementary material)
was openly available in the hospital intranet. No modification has ever been made since 2018.
- I, as the main responsible for this project, was not part of the AMS team or the Infection Committee
at the hospital, which potentially reduced trust.
Prescriber-related factors:
-Haematologists trusted it, and stated that “they followed it correctly”. Having a carbapenem as a first
line option for most of their patients was sensed as “relieving”, under the impression that empirical therapy
should be “as broad as possible” and cover Pseudomonas aeruginosa.
-The responsible haematologists disliked being “observed” or “judged” on how they were using
antibiotics.
-The residents were taught to take the same decisions as their specialists. Moreover, usually they
would not know a different perspective unless they did a rotation at a different hospital.

Based on the COMB Model of Behaviour theory, I classified them as it follows:

- Capability barriers: it was important for the prescribing haematologists to understand basic concepts of
antimicrobial resistance phenotypes that show in cumulative susceptibility reports, regarding what they intend
to cover with an antimicrobial therapy. Furthermore, having the example of a protocol that was being used in a
different hospital, with a similar epidemiology, was also an important component.
- Motivation barriers: data on the benefits of Antimicrobial Stewarship and adverse effects of antimicrobial
therapy were also addressed, in order to surpass the automatic motivation of the maintained habits in
antimicrobial prescription. The content of the intervention aimed at a more reflective motivation, to make
them part of the decision of adapting the protocol to their own epidemiology and patient profile.
- In terms of opportunity, the main fact to be addressed was to periodically update the microbiological data
that could help them adapt their protocols as needed.

Facilitators:
Prescriber-related factors:
- All haematologists showed their will to try to improve patient care, from the very beginning. They
were open to discussion if I could prove the facts regarding Pseudomonas aeruginosa epidemiology and
antimicrobial resistance mechanisms.
th
-One of the 4 year residents in the Department of Haematology had recently done a rotation with a
national expert on ID management in immunosuppressed patients (including Haematology). Her support was
crucial in organizing a general reunion with the Haematology team, in order to explain her gained perspective
and experience during the rotation.
Organizational factors:
- The Department of Clinical Microbiology traditionally has close and friendly relationship with
haematologists, due to the fact that we share laboratory space, among other reasons. This close relationship
often leads to checking data about haematological patients for collaborative and research reasons, including
not only their microbiological isolates and results, but also antimicrobial management.
-Many colleagues in the Department of Microbiology openly expressed their will to participate in data
collection and interpretation of results. Our data is informatized, and accessible to collect.
-The Hospital Pharmacist responsible for antibiotic consumption helped and collaborated by providing me with
the DDDs needed for this study.

16
Stage 4 Development and execution of the stewardship intervention
An initial reunion was held with several members of the Haematology team, including the two responsible for
in-hospital patient care and two residents, including one who had recently done a rotation in ID consultations
for immunosuppressed patients in a different hospital.
Based on the previously explored data, an update session on febrile neutropenia management, Pseudomonas
th
aeruginosa epidemiology in our setting and its resistance mechanisms was offered in collabration with the 4
year Haematology resident, for the whole Department of Haematology. Other aspects such as cumulative
antibiograms for Enterobacterales and the high rate of contaminants in blood cultures were also highlighted,
these being only indirectly related to the main topic. The main objective of this session was not only
informative, but also a try to make them feel part of a potential improvement.
Pseudomonas aeruginosa considered a priority pathogen to be empirically covered in severe neutropenia, it
was explained that considering carbapenems a better and/or safer option for its coverage can be a
misconception. Analyzing the data on cumulative antibiograms for Pseudomonas aeruginosa, it was proven
that antipseudomonal cephalosporins have, in fact, a better susceptibility rate in our setting. Furthermore, the
incidence in Pseudomonas aeruginosa bacteremia was low in the pre-intervention study period.
Moreover, some facts were pointed out regarding the protocol in use:
-The protocol did not contemplate any situation where antipseudomonal cephalosporins or
piperacillin/tazobactam could be used, directly escalating to antipseudomonal carbapenems.
-Given that meropenem was very frequently used in these patients, even in the absence of its main indication
according to the protocol, compliance was low
-More therapeutical options should be included as part of an antimicrobial stewardship intervention.
- The proposed doses for the included antimicrobial agents are susceptible to be revised in future
updates.
The decision to update the febrile neutropenia empirical management protocol was made by the Haematology
team itself, and to begin the re-introduction of other antipseudomonal agents empirically (including
ceftazidime, cefepime, and piperacilin/tazobactam).
Nevertheless, it was also highlighted that, besides the cumulative report for Pseudomonas aeruginosa, other
factors had to be taken into account: other cumulative reports (like Enterobacterales), specific clinical features
(including the suspected type of infection) and previous history of MDR bacteria (including betalactamases like
ESBL, AmpC producers or carbapenemases, MRSA or VRE, for example).

Stage 5 Evaluation of the stewardship intervention

In the post-intervention period, 138 patients were hospitalized in Haematology (versus 170 in the pre-
intervention), with 11 deceases (8%), making no statistically significant differences in total mortality (p=0.538).
Measurements were equally made, and results shared with the Haematology group.
-Percentage of patients with febrile neutropenia among all haematological patients with blood cultures (151
sets of blood cultures - two sets per episode, very rarely just one): 68,8% had neutropenia and fever at the time
when the cultures were collected (defining neutropenia as <1500/μl). Among them, 75,9% had severe
neutropenia (<500/μl).
Pre Post
Neutropenia (among collected BC) 61% 68%
Severe neutropenia 83.6% 75.9%

-Positivity rate of blood cultures in all haematological patients: 18.5% of all blood cultures (151) were informed
as a positive result. This improvement is possibly related to a different Diagnostic Stewardship project was held
at the same time (with different interventions and measurements).
Pre Post
Positivity rate in BC 24% 18.5%

17
 Anaerobes; 14,20% Pseudomonas
aeruginosa; 21,40%

Enterobacterales;
28,50% Coagulase negative
staphylococci;
Enterococci; 14,20% 21,40%
A different distribution was observed in the isolates from blood cultures during this study period. Pseudomonas
aeruginosa bacteriemia cases accounted for 6 patients, which was significantly higher than the previous study
period the reason being unknown (as previously mentioned, a higher total percentage of neutropenia was
found in this period, but among them, less were severe) and probably related to the low number of isolates
included in this study. No other non-fermenting bacilli were isolated during this period.
The epidemiological study for Pseudomonas aeruginosa was also repeated in the second period:
Blood; 6,50% Rectal (MDR
control);
3,14%
Respiratory; 19,20%
Urine; 34,40%

Non-respiratory
exudates; 39,50%
Cumulative susceptibility profile of routinely tested antipseudomonal betalactam antimicrobials in isolates of
Pseudomonas aeruginosa from the 6-month post-intervention study period (n=380), excluding rectal swabs for
MDR carriage (n=12). Wild type only refers to antipseudomonal betalactams.
Wild-type to all antipseudomonal betalactams 73.7%
Piperacillin/tazobactam 87.6%
Ceftazidime 89.3%
Cefepime 90.3%
Imipenem 83.4%
Meropenem 85.7%
Ceftolozane/tazobactam 99.4%
Co-resistance phenotypes
Piperacillin-tazobactam & Antipseudomonal cephalosporins 2.7%
Antipseudomonal carbapenems 11.9%
All agents excluding Ceftolozane-tazobactam 5.7%
Cefepime & Meropenem 6.1%
About 26% of isolates showed at least one resistance phenotype, and co-resistance rate was highest between
antipseudomonal carbapenems. Only one metallobetalactamase enzyme (VIM) producing isolate was found in
this period. As in the first period, these data suggest that, even though the general susceptibility for
antipseudomonal betalactams could probably be considered acceptable for empirical treatment options,
cephalosporins continued showing a better profile compared with carbapenems (in this period, even
piperacillin/tazobactam did).
Finally, to better understand the comparative use of different antipseudomonal agents, cumulative data
provided by the Hospital Pharmacy department was studied. The following list corresponds to the use of the
antibiotics in a 6 month-period post the intervention date, measured in DDDs/100 hospitalization days in
Haematology. In this period, 138 patients were hospitalized in Haematology, with 11 deceases (8%), showing
no statistical differences with the first period (p=0.538). The measurements were divided in two periods,
corresponding to 3 months each, to have a better understanding of the impact of the intervention:

Pre Post 1 (3m) Post 2 (6m)

Piperacillin/tazobactam 3.09 0.15 4.33
Ceftazidime 0.91 0 0

18
Cefepime 0.48 3.29 4.69
Imipenem 0.31 0 0
Meropenem 18.43 9.6 10.61
Overall, as the las table shows, a re-arrangement of antibiotic use was observed, especially with meropenem
(initial decrease of 47.7%, followed by a little increase to 42.4% lower than in the pre-intervention period) and
cefepime (initial increase of 660%, followed to reaching 940% of its use in the pre-intervention period). The
reason why Ceftazidime and Imipenem went from low DDDs to not being consumed at all is believed to be
related to systematically trusting the other options instead, and does not reflect a change in patient profiles or
epidemiological variables. An initial drop in the consumption of Piperacillin-tazobactam was also observed,
followed by a rise at the last measurement point, and is believed to be related to a similar reason, for which
initially, during the first post-intervention weeks, a need for prescribing habits was well accepted and the
preferred option was Cefepime, except from the cases when Meropenem was still being prescribed.
The following image is a graphic illustration of the evolution of antimicrobial consumption over time during the
study, measured in DDDs/100 hospitalization days in Haematology.
20

15 Piperacillin-tazobactam
Ceftazidime
10
Cefepime
5
Imipenem
0 Meropenem
Pre-intervention 3 months post- 6 months post-
intervention intervention
Conclusions
Based on the observation of how the Department of Haematology aimed at covering infections potentially
involving Pseudomonas aeruginosa with antimicrobial agents that routinely showed to be less susceptible than
other betalactam options, an intervention was designed in coordination with other colleagues from
Microbiology, Haematology and Hospital Pharmacy. The intervention focused on local epidemiological aspects
of cumulative antibiograms (of Pseudomonas aeruginosa and other common aetiologic microorganims),
resistance mechanisms and empirical antimicrobial therapy in febrile neutropenia, learnt from a team of ID
specialists from a different hospital with a similar epidemiology to ours.
Although the study period is short, the generated feedback so far has been used to prove that no big
differences should be expected in cumulative profiles in short periods of time, which can help in gaining
confidence for the election of an empirical antimicrobial treatment. A new protocol for empirical antimicrobial
treatment in febrile neutropenia is under development. Furthermore, notable changes have already been
observed in antimicrobial consumption data, and are expected to continue similarly in the near future.
Given all the barriers and facilitators present in this project, mutual trust and goodwill were essential parts. As
of today, we continue collecting the same data about Pseudomonas aeruginosa and other aetiologic
microorganims in Microbiology; antimicrobial consumption data are regularly studied in Hospital Pharmacy;
and patients continue developing febrile neutropenia in the Department of Haematology. A very satisfying
difference with the pre-intervention era is that, many times when I have the chance to meet in person with the
haematologists, they will mention things like “We started our patient on Cefepime, because we remembered
what you told us”.
The studied data also discovered other possible interventions to be considered in the future, such as the
importance of protocols for blood-culture collection and its impact on the rate of contaminant species, as well
as the doses for the antimicrobial agents included in the protocol (betalactams and other groups).
Thanks to the ESCMID AMS Certificate Program, I have developed a deep insight into the variability of aspects
that have an influence on antimicrobial and diagnostic stewardship.

Acknowledgements
This project could have not been possible without the help and support from my involved colleagues in
Microbiology, Haematology and Hospital Pharmacy, as well as the guidance and mentorship provided by Dr.
Bhanu Sinha, Dr. Jeroen Schouten and of the rest of professionals involved in the courses of the AMS
Certificate.

19
Supplementary material

Definition used for febrile neutropenia

Febrile neutropenia is described as Tº 38ºC maintained for at least 1 hour, in patients with severe
neutropenia (<500/μl), or mild neutropenia (500-1000/μl) if it is expected to continue lowering until a severe
range. It is considered to be a medical emergency and prompt empirical antimicrobial coverage is a priority,
with previous blood culture extraction (always including peripheral sets, and adding sets from peripheral or
central catheters if present), as well as any other microbiological sample based on the presenting syndrome.

Local protocol for empirical antimicrobial treatment in febrile neutropenia (2018). Written by the
Departments of Haematology and Medical Oncology. Approved by the hospital Infection and Antibiotic Use
Committee.
This guideline stratifies the risk of the febrile neutropenia episode and indicates an antibiotic regimen based on
that risk:
 Low risk:
Patients starting fever while outside of the hospital, with every criteria included in the following list:
o Severe neutropenia (<500/μl) with an expected duration of <7 days.
o MASCC score (Multinational Association for Supportive Care in Cancer) ≥ 21 points (which implies <5% of
complications).
o Haemodynamic stability
o No comorbidities (CrCl > 50mL/min, liver funtion tests <3x normal range)
o No Acute Myeloid Leukemia
o No radiological signs of pneumonia
o No signs of severe skin & soft tissue infection

This is the only scenario where a further differentiation is made upon criteria for treating the patient in an in-
hospital versus out-hospital setting.
o Out-hospital: adequate oral tolerance, absence of/low grade mucositis, 24h family support. Patients are
kept in observation for at least 4h before discharge, and re-evaluated in 48-72h, at which point
hospitalization will be indicated if fever persists.
st
1 line: Amoxicilin/clavulanate 875/125 mg /8h oral
nd
2 line: Levofloxacin 500mg /24h oral
o In-hospital: inadequate oral tolerance, ≥2 grade mucositis, no family support
st
1 line: Ceftriaxone 2g iv /24h
nd
2 line: Levofloxacin 1g iv /24h

 High risk:
Patients with at least one of the following criteria:
o Severe neutropenia (<500/μl) with an expected duration of >7 days.
o MASCC score <21
o Hemodynamic instability
o Comorbidities (ClCr < 50mL/min, liver funtion tests >3x normal range)
o Acute Myeloid Leukemia
o Radiological signs of pneumonia, severe skin/soft tissue infection, mucositis grade 3-4

A further differentiation is made if the patient has previous records of MDR bacteria colonization:
No previous record of MDR bacteria:
st
1 line: Imipenem 500mg iv /6h (+Teicoplanin 6mg/kg iv /12h if catheter related infection is
suspected).
Or Meropenem 1g iv/8h if high risk of seizures
nd
2 line: Amikacin 15mg/kg iv /24h (+Teicoplanin 6mg/kg iv /12h if catheter related infection is
suspected)

Previous record of MDR bacteria (including methicillin resistant S. aureus, glycopeptide resistant enterococci,
ESBL or carbapenemase producing Enterobacterales:

20
 st
1 line: Imipenem 500mg iv /6h (+Teicoplanin 6mg/kg iv /12h if catheter related infection is
suspected) AND:
Vancomycine 1g iv/8h if methicillin resistant S. aureus
Linezolid 600mg iv /12h if glycopeptide resistante enterococci
Amikacin 15mg/kg iv /24h if ESBL/carbapenemase producing Enterobacterales

Other clinical scenarios are considered separatedly:

Typhlitis (neutropenic enterocolitis):
st
1 line: Metronidazol 500mg iv/8h (coverage for Clostridium septicum) + Imipenem 500mg iv/6h

Invasive fungal infection, CMV and other specific agents are considered separately.

MASCC Score:
The MASCC (Multinationa Association for Supportive Care in Cancer risk index) score is a predictive tool to
assess the risk of complications in patients with chemotherapy related febrile neutropenia (2). It identifies
patients with febrile neutropenia who are low risk for complications and may be suitable for outpatient
management. It is calculated based on the sum of the following variables:
-Burden of illness
-Hypotension
- Active chronic obstructive pulmonary disease
- Type of cancer
- Dehydration requiring IV fluid therapy
- Status at onset of fever
- Age
A score of ≥21 predicts a low risk of medical complication and such patients may be suitable for outpatient
management with oral antibiotics.

Bibliography:
1. Del Barrio-Tofiño E et al for the GEMARA-SEIMC/REIPI Pseudomonas study Group. Spanish nationwide
survey on Pseudomonas aeruginosa antimicrobial resistance mechanisms and epidemiology. J
Antimicrob Chemother. 2019 Jul
2. https://www.cancercalc.com/MASCC.php

21
ESCMID AMS Certificate Program
My Project Report

Name
Carlos Tarrazo Tarrazo
Institution
San Agustin University Hospital
Country
Spain
Tutor
Petros Ioannou
Title of Project
Implementation of a bedside consultation for Staphylococcus aureus bacteremia
Date of submission
March 16, 2024

Abstract

Staphylococcus aureus bacteremia (SAB) is a common infectious disease with high rates of morbidity
and mortality. Recently, studies have shown that consultation with a clinician with expertise in
infectious diseases may reduce mortality, and several other quality indicators have shown to improve
outcomes in these patients. Given the higher-than-expected mortality of SAB in our hospital and low
adherence to some of these identified quality indicators, we implemented a bedside consultation for all
patients with SAB with the objective to improve management and outcomes. We analyzed the data
from the post-implementation period (12 months) and found an increase in adherence to the selected
quality indicators and a benefit in survival when compared with the pre-implementation period (12
months).

Stage 1 Introduction

I work as an Internal Medicine specialist and member of the Antimicrobial Stewardship (AMS) team in a
general hospital with 450 beds. Our AMS team is formed by two clinicians who are Internal Medicine
specialists with expertise in infectious diseases, two microbiologists, one pharmacist and one intensivist.
Given the absence of an official Infectious Diseases (ID) specialty in Spain, we internists are also in
charge of managing patients with complex infectious diseases, answering ID consultations from other
physicians, and following these patients in the outpatient setting.
In our daily practice, we frequently encountered patients with SAB, and we had a subjective feeling of
higher-than-expected mortality when compared to the mortality described in the literature. We also
observed an inappropriate antibiotic choice or duration and absence of source control in a lot of these
patients. We were only consulted as requested by the leading physician, sometimes after complications
as persistent fever and/or metastatic infections were identified, so we were only involved in a small
proportion of SAB episodes.
For this reason and given recent publications in the last years that showed improved outcomes in
patients with SAB associated with identified quality indicators and an ID consultation, we hypothesized
that implementing an early unrequested bedside consultation for all patients after Staphylococcus
aureus isolation on blood cultures could improve adherence to these quality indicators and decrease
mortality and/or readmissions.

Stage 2 Measurement

To evaluate management and outcomes of SAB episodes prior to our intervention, we selected a pre-
intervention period of 12 months, and we performed a retrospective analysis from the electronic

22
medical record (EMR) of the characteristics of all episodes of SAB (n=43) from February 1, 2022, to
January 31, 2023. All SAB episodes were extracted from the Microbiology electronic database and were
provided by the microbiologist in the AMS team after reviewing all positive blood cultures during that
period.
To assess the management of each of these episodes, we did a literature review of identified quality
indicators associated with improved outcomes in patients with SAB and selected the ones that showed
to have more impact in outcomes and seemed feasible to improve in our institution (described in Table
1).

Table 1. Quality indicators and definitions

Quality indicators Definition
Follow-up blood cultures Follow-up blood cultures drawn in the first 72h after initiation of antimicrobial
(BC) therapy in all patients with SAB
Transthoracic Transthoracic echocardiography performed in the first 7 days after isolation of
echocardiography (TTE) Staphylococcus aureus in blood cultures in all patients with SAB
Transesophageal Transesophageal echocardiography performed in patients with complicated
echocardiography (TEE) SAB
Adequate empirical Empirical antibiotic treatment against MSSA and MRSA (vancomycin or
antibiotic treatment daptomycin) initiated in the first 24h after isolation of Staphylococcus aureus
in all patients with SAB
Adequate directed Antibiotic treatment with an antistaphylococcal penicillin or first-generation
treatment cephalosporin in the first 24h after susceptibility results for MSSA
Antibiotic treatment with vancomycin or daptomycin in the first 24h after
susceptibility results for MRSA
Adequate duration of Duration of antibiotic therapy of at least 14 days after bacteremia clearance
treatment for uncomplicated SAB
Duration of antibiotic therapy of at least 28 days after bacteremia clearance
for complicated SAB
Source control Source control performed, in SAB patients with an identifiable source requiring
drainage and/or removal of prosthetic material

We found that a very small proportion of patients received adequate duration of therapy, source
control when needed or were performed a transesophageal echocardiogram to rule out endocarditis,
even in the presence of complicated bacteremia, prosthetic material, or cardiac devices. The rate of
follow-up blood cultures, adequate empirical and directed antibiotic choice and transthoracic
echocardiography was not optimal but better than we anticipated (results depicted in Table 2).

Table 2. Adherence to quality indicators in the pre-intervention period

Quality indicator Adherence
Follow-up BCa 31/43 (72.09%)
TTEa 27/43 (62.79%)
TEEb 6/31 (19.35%)
Adequate empirical therapya 32/43 (74.41%)
Adequate directed therapya 35/43 (81.40%)
Adequate duration of therapy a 12/33 (36.36%)
Source controlc 11/20 (55%)
a. Proportion of patients from the total of SAB episodes
b. Proportion of patients with complicated SAB who underwent TEE
c. Proportion of patients with an identified source of infection requiring source control

To analyze the clinical outcomes of each SAB episode, we measured death at 6 weeks from
Staphylococcus aureus isolation in blood cultures and readmission attributable to the infection or its
complications at 6 weeks after hospital discharge. We found that 14/43 patients died (32.56%), a

23
mortality higher than the one described in the literature, and that 7/43 patients were readmitted
(16.27%).

Stage 3 Barriers and facilitators

Before implementation, we identified several determinants of practice that acted as barriers in three
different Flottorp domains, related mostly to guidelines and health professional factors, but also
influenced by lack or resources. During the implementation period we encountered a barrier related to
patient factors because of increased length of antibiotic therapy.
• Regarding guidelines, the first barrier we encountered was the lack of a SAB guideline in our
institution.
• The second barrier was related to health professionals and consisted of lack of knowledge
about SAB management, that led to a low adherence to the quality indicators.
• The third barrier was related to the difficulties in the Cardiology department to perform
echocardiography in all SAB episodes, due to both lack of knowledge about the indication and
underestimation of the prevalence of infective endocarditis in this population, and to lack of
human resources that resulted in a significant delay to perform an echocardiography in
inpatient patients.
• The fourth barrier we identified was the cause of the low rate of source control in SAB
episodes, and was related to health professional factors, due to the lack of knowledge and
awareness of surgical departments of the importance of source control, and lack of motivation
to perform the procedures and follow-up required by the procedures to drain abscesses or
removed prosthetic material.
• During the implementation period we encountered a fifth significant barrier related to patient
factors. With the improvement in adequate duration of therapy and increase recognition of
complicated bacteremias, treatment duration and hospital stay increased, mostly in those
patients with MSSA bacteriemia that were not suitable for OPAT treatment with daptomycin.
• The main facilitator we had was related to professional interactions and communication. We
already had a great communication system between microbiologist and clinicians, so it was
easy to communicate immediately with a phone call to the other members of the team every
time Staphylococcus aureus was isolated in a blood culture and that led to evaluation of the
patients in the same morning. The modest size of the hospital also was a facilitator to the
communication with other health professionals, allowing us to personally discuss SAB cases on
the first day of bacteremia diagnosis with the leading team and provide a set of
recommendations that led to an early improvement in management.

Stage 4 Development and execution of the stewardship intervention

First, in the pre-implementation period, we tried to overcome the barriers we had previously identified.
• Regarding the lack of a SAB guidelines in our hospital, we did a literature review of
international and national guidelines and developed our own hospital guidelines for the
management of SAB. As part as the Infectious Diseases Committee of the hospital we
presented the completed guideline to the rest of the committee in December 2022 and was
finally approved in January 2023, before the implementation date.
• Regarding the lack of knowledge in other healthcare professionals about SAB management, we
did an educational session about SAB management and presented the new guidelines in one of
the weekly educational meetings we have with all the other medical departments. That led to a
greater understanding of the importance of the different quality indicators and improved the
adherence to the recommendations.
• We overcome the barrier of the lack of knowledge and the limited resources in the Cardiology
department by emphasizing the importance of echocardiography in SAB during the
presentation in the general educational session described previously and by doing an individual
meeting with members of the Cardiology department in which we reviewed the literature,
reached a consensus and wrote an algorithm to perform TTE in all patients with SAB and avoid
TEE in patients with SAB with low risk of endocarditis based on recently validated scores (we
used the VIRSTA score) and in those in which TEE or endocarditis diagnosis wouldn´t change

24
 management as in those patients not suitable for surgery or invasive treatment because of
frailty or comorbidity who were going to be treated for at least 4 weeks anyway because of
complicated bacteremia. In the algorithm we also agreed to prioritize early studies in patients
with high risk of endocarditis as in those with cardiac devices or prosthetic valves.
• Regarding the lack of knowledge and motivation in surgical departments to perform source
control, we also tried to do individual educational meetings with each department, but we
didn’t find the same success. We found and increased acceptance and predisposition to
remove prosthetic material like cardiac devices from Cardiac Surgery, mainly because of
increased use of Nuclear Medicine studies that led to an improvement in infection detection
and reduced diagnostic uncertainty, but we didn’t have enough success with other surgical
departments like Orthopedic Surgery, and the rate of source control in osteoarticular infection
was very low.

From February 1, 2024, we implemented an unrequested bedside consultation for all patients with SAB.
After the blood cultures isolation, the Microbiology laboratory directly reported the results to the AMS
clinician, and we went to evaluate the patient and discuss the case with the lead physician, all within the
same morning of SAB diagnosis. We helped with the management plan and provided a set of
recommendations for each patient based on the different quality indicators. This set of
recommendations was also recorded in the electronic medical record. We followed-up the patients
during the hospital stay until discharge.

During the post-implementation period we found and unexpected barrier regarding patient factors and
discomfort because of increased hospital stay to complete an adequate duration of treatment. We
overcame this problem using the evidence recently presented in the SABATO trial to implement an early
oral switch for patients with low-risk SAB, and discussed with members in the OPAT team and our
pharmacist the possibility to buy elastomeric pumps, so we started treatments with cefazolin in
continuous infusion at home for those patients with complicated bacteremia who needed to stay on IV
antibiotics and fulfilled criteria for OPAT treatment, which ultimately led to an overall decrease in
hospital stay in the post-implementation period.

Stage 5 Evaluation of the stewardship intervention

During the 12 month post-intervention, from February 1, 2023 to January 31, 2024 we evaluated the
adherence to the quality indicators previously described, the mortality at 6 weeks and the readmissions
at 6 weeks because of clinical failure in 42 SAB episodes, and compared them to the 12 month pre-
intervention period (depicted in Table 3 and Table 4).

Table 3. Comparison between adherence to quality indicators in the pre-intervention and post-
intervention periods
Adherence to quality indicators
Pre-intervention Post-intervention p
period (n=43) period (n=42)
Follow-up BC 31/43 (72.09%) 40/42 (95.24%) 0.0068
TTE 27/43 (62.79%) 34/42 (80.95%) 0.0913
TEEa 6/31 (19.35%) 17/36 (47.22%) 0.0213
Adequate empirical therapy 32/43 (74.41%) 41/42 (97.62%) 0.0034
Adequate directed therapy 35/43 (81.40%) 41/42 (97.62%) 0.0296
Adequate duration of therapy 12/33 (36.36%) 31/39 (79.49%) 0.0003
Source controlb 11/20 (55%) 19/25 (76%) 0.2047
a. Proportion of patients with complicated bacteremia who underwent TEE
b. Proportion of patients with an identified source of infection requiring source control

25
Table 4. Comparison between clinical outcomes in the pre-intervention and post-intervention periods

Pre-intervention Post-intervention p
period (n=43) period (n=42)
Death at 6 weeksa 14/43 (32.56%) 5/42 (11.9%) 0.0358
Readmission at 6 weeksb 7/43 (16.27%) 3/42 (7.14%) 0.3134
a. Death at 6 weeks after diagnosis of SAB
b. Hospital readmission (because of the infection) at 6 weeks after discharge

Conclusions
After implementation of a bedside consultation for all patients with Staphylococcus aureus bacteremia,
we observed a significant increase in the adherence to some quality indicators, including the
performance of follow-up blood cultures, transesophageal echocardiography for patients with
complicated bacteremia and the adequation of duration and choice of empirical and directed
antimicrobial therapy. We failed to significantly improve the performance of transthoracic
echocardiography and source control, mainly because of inadequate source control in patients with
osteoarticular infections.

We also found a significant decrease in mortality in the post-intervention period. In contrast with the
high mortality in the pre-intervention group, the mortality in the post-intervention period was low
(11.9%) compared to the mortality described in the literature for patients with SAB, which may be
influenced by our low rate or MRSA bacteremia (16.7%). There were no significant differences between
the pre-intervention and post-intervention groups in the proportion of patients with MRSA bacteremia
(20.9% vs 16.7%) or the proportion of patients with complicated bacteremia (79.07 vs 80.95%), so the
decrease in mortality seen between groups seems to be attributable to the improvement in
management. It must be noted that a small proportion of patients in the pre-intervention group were
evaluated by the AMS team as requested by the lead physician, so the impact of this intervention in a
naïve population could be bigger.

The AMS Certificate Program has given me the tools to implement a successful intervention in our
hospital. The good results after our intervention and the interaction with colleagues from different
countries during the course and their different projects, have motivated me to continue to implement
interventions in our institution to improve the use of antibiotics and management of infectious diseases.

26
ESCMID AMS Certificate Program
My Project Report

Introduction

Name
Cătălina Hăpăianu, MD
Institution
Infectious Diseases Hospital, Cluj Napoca, Romania
Tutor
Dr. Sophie Schneitler/Nina Bühler B.C.
Title of Project
Antifungal stewardship using 1,3 Beta D Glucan and Candidemia guideline
Date of submission
15 March 2024

Abstract
Antifungal resistance along with antibiotic resistance poses a growing threat to patient care,
necessitating effective stewardship programs. Over 1,5 million people die every year from fungal infections.
Compared to antibiotics, there are only four classes of systemic antifungal medicines (azoles, echinocandins,
pyrimidines and polyenes), and increasing antifungal resistance poses a growing threat to patient care.
Analyzing antifungal consumption in our hospital across pre-pandemic, pandemic, and post-pandemic
periods revealed a consistent upward trajectory in the post pandemic period. At the national level, there is an
observed increase in antifungal consumption, as indicated by the European antimicrobial consumption report.
We implemented the 1-3 beta D glucan test in the laboratory to improve antifungal consumption.
Following the implementation of the test, I conducted an educational session to ensure comprehensive
understanding of its utilization. Consequently, we developed a candidemia guideline based on existing literature
to ensure appropriate antifungals administration.

Stage 1 Introduction

The hospital I work at, as a Laboratory Medicine Doctor, on the microbiology and serology department,
is a Teaching Infectious Diseases Hospital with 190 beds. The hospital has 4 adult wards with 28 beds each, one
children ward with 48 beds, an HIV ward with 22 beds and an ICU ward with 9 beds.
When I started the AMS Certificate Program, the Hospital didn't have any Antimicrobial Stewardship or
Antifungal Stewardship team in place. In Romania, there isn't a legal requirement for hospitals to have an
Antimicrobial Stewardship (AMS) or Antifungal Stewardship (AFS) team, even though our country has one of the
highest resistance rates in Europe. This lack of regulation makes the implementation of a stewardship program
challenging.
The Hospital was designated as a COVID-19 hospital for two years between 01 March 2020 – 31 March
2022. Every COVID-19 case from the town was redirected here in this period of time, and any other infectious
diseases cases were sent to other internal medicine or ICU wards from other hospitals around.
Although there isn't currently an established AMS-team, certain categories of antibiotics and antifungals
are subject to usage restrictions in the hospital. For instance, fluoroquinolones have been restricted for five
years, and more recently, since mid’ 2023, carbapenems have also been subjected to restrictions due to the
alarming rate of resistance, particularly among Klebsiella pneumoniae strains.
As a laboratory doctor, I sought to enhance good practices within my workplace. Consequently, we
introduced a new test, 1-3 Beta D Glucan, aimed to optimize antifungals utilization and potentially discontinuing
unnecessary treatments when feasible.
I had a late start with the project, because I had to choose something I could do from my position and
other project ideas before was not widely accepted. Testing started in March 2023 after a lengthy process of
obtaining approvals and the validation procedure. In Romania there is only one distributor for this test, so we
didn’t have the opportunity to choose other. We are the only laboratory in town offering this test, that is why
my database consists of patients belonging to our hospital and other hospitals in town as well. I had no acces to
the data files of the patients from the other hospitals.

27
 Hospital protocols for fungal infections are currently not available. While culture remains the preferred
method for diagnosing fungal infections, it is time-consuming and yields a positivity rate of 50-70%. In contrast,
serologic diagnostics helps in providing a faster and simpler assessment, helping to avoid unnecessary and
prolonged antifungal therapy. Despite the low incidence of candidemias, the implementation of a guideline could
optimize the utilization of antifungal medications.

Stage 2 Measurement

To assess the utilization of antifungals, we examined four distinct time periods. The first period
encompasses the year prior to the onset of COVID-19 (March 1, 2019, to February 28, 2020). The second period
spans the first year of the COVID-19 pandemic (March 1, 2020, to February 28, 2021). The third period
corresponds to the second year of the COVID-19 pandemic, while the fourth period encompasses the year
following the pandemic of COVID-19 (April 1, 2022, to March 31, 2023). Data were extracted from electronic
patient records and analyzed utilizing the antimicrobial consumption tool (AMC Tool).
Fluconazole is the sole antifungal medication permitted for use without specific approval on the clinical
wards of the hospital. The administration of other antifungals (Anidulafungin, Caspofungin, Micafungin,
Isavuconazole, Posaconazole, Voriconazole, and Amphotericin B) requires approval from the medical director
prior to their utilization. In the ICU ward, echinocandins are always readily available when required.
The graphical representations of the antifungals consumption, expressed in DDD and DDD/100 bed-
days, indicate an upward trend in our hospital. However, this trend remains lower than that observed in the
pre-COVID-19 year.

Antimycotics for systemic use DDD

3000,000
2000,000
1000,000
0,000
Pre COVID COVID Year 1 COVID Year 2 Post COVID

Antimycotics for systemic use DDD/100 bed-days

8,000
6,000
4,000
2,000
0,000
Pre COVID COVID Year 1 COVID Year 2 Post COVID

Considering that the total consumption (community and hospital sectors) of antimycotics for systemic
use (ATC group J02), in EU/EEA countries, 2018–2022, are expressed as DDD per 1 000 inhabitants per day, I
cannot do a real comparison. As we can see from the graphs, the trends are ascending in our hospital. The
consumption trends are ascending in Romania according to the European report.

Stage 3 Barriers and facilitators

Using the Flottorp framework as a reference, the following barriers and facilitators were found.
Barriers: (1) Lack of knowledge or understanding of the new test (1-3, beta-D-glucan). (2) Lack of
guidelines for systemic fungal infections. (3) Opinions and communication among professionals may hinder
adherence to the guideline.
Facilitators: (1) Throughout the entire process, I received support from the Infectious Diseases (ID)
professor at our hospital and our pharmacy clinician. The ID professor assisted me also in reviewing the
guideline. (2) My international tutor, Dr. Sophie Schneitler, and Nina Bühler, who oversees the Serology

28
department at Dr. Schneitler's hospital, conducted a review of the guideline and aided me in validating the 1-3
Beta D Glucan test. Despite the delayed initiation of my project, they assisted me closely to finalize it. (3) The
hospital has an intranet system that provides easy access to internal documents and guidelines.

Stage 4 Development and execution of the stewardship intervention

To address the initial lack of knowledge and comprehension of the test, I disseminated an informative
email detailing the test description to each ward's head physician at the outset of the testing process.
Furthermore, upon receiving test requests from external hospitals, I supplied the results accompanied by an
explanatory email attachment. Additionally, whenever feasible, I personally phoned each physician who
requested the test to discuss the clinical rationale. Because the test informative email wasn’t the best method
to explain the test, on 23rd of November 2023 I did a presentation of serologic diagnostic for fungal infections
that we do in our laboratory. The presentation is also available on the intranet for anyone in our hospital who
wants to access it.
The testing commenced in the latter half of March, with a total of 86 patients tested, 58 of whom
were from our hospital and 28 from other hospitals. We function as the designated Reference laboratory for
this test, as it is not conducted elsewhere in our city. Although I have included these patients in the database,
complete data for them is not available.
1-3 Β-D-GLUCAN TESTS PER MONTH AND GENDER
8
7
6 2
5 2
5 3
4
3 5 7
3 2 2 2
3 5 5
2 2 2 4 4 2 4
1 3
1 2 2 2 2 2 2
1 1 1 1 1 1
0
F M F M F M F M M F M F M F M F M F M
Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Other Hospital Infectious Diseases Hospital

Since March 2023, we tested six patients/month from our hospital. In November and December the
number decreased because there were some renovations in the area and half of the ICU rooms were closed. The
graphs show that the number of tests from others hospitals is increasing since November 2023.
Most of the patients tested from our hospital were males older than 50 years old.

1-3 Β-D-GLUCAN TEST RESULTS BY AGE AND SEX

AT INFECTIOUS DISEASES HOSPITAL
12
10 1

8 1 2
6 1
1 10
4 7 7
6 1
2 3 4 4 3 1
1 1 1 2 1
0
M M M F M F M F M F M F M
00 - 10 20 - 30 30 - 40 40 - 50 50 - 60 60 - 70 70 - 80 80 - 90

Nonreactive Reactive

29
 To address the lack of local guideline we developed a Candidemia guideline following the latest
Candidemia guideline (IDSA 2016), and other international guidelines to enhance the diagnostic and the quality
of care. On 9th February 2024 I presented the Candidemia guideline to the clinicians and now it is also available
on the intranet.
To address adherence to the guideline, I coopted in the team more people. Me and the pharmacy
clinician worked to the guideline and the ID Professor and the head of the Laboratory reviewed the guideline.

Stage 5 Evaluation of the stewardship intervention

As a result of the informative email distributed at the beginning of the test, only a handful of doctors
contacted me to request further information.
On the first educational session where I presented the serologic diagnostic for fungal infections attended
40 doctors. The session was held in the morning, after the daily medical report and that meeting almost
everybody attends it.
At the second educational session when I presented the Candidemia guideline I used Mentimeter to get
some feedback from the audience. There were 40 participants who answered the feedback questions, 95% of
them considered that they will use the presented protocol.
The antifungals consumption and the adherence to the guideline will be evaluated after a period of 6
months.

Conclusions

Throughout my participation in the ESCMID AMS Certificate Program, I have gained a profound
understanding of antimicrobial stewardship principles and their application in clinical practice. Specifically, I've
learned about the importance of antimicrobial prescribing guidelines, the role of multidisciplinary teams in
stewardship efforts, and strategies for optimizing antibiotic use while minimizing resistance.
The ESCMID AMS Certificate Program has been a transformative experience for me personally and was
a great opportunity to be part of a great international network, to have a great international tutor, to do the
international observership and to see how other hospitals deal with this global problem. I am grateful for the
knowledge and skills gained through the program and remain committed to advancing antimicrobial stewardship
efforts within our institution and beyond.
As a laboratory medicine doctor, engaging in stewardship efforts presents unique challenges, yet it's
essential to recognize the valuable tools at our disposal for enhancing antibiotic prescribing practices. Through
courses like this, I've come to understand that with dedication and the right approach, every healthcare
professional can contribute to improving prescribing habits and fight antimicrobial resistance. While some of us
may take significant strides in this journey and others may make smaller ones, every effort counts towards our
collective goal.

30
ESCMID AMS Certificate Program
My Project Report

Introduction

Name
Francisco Miguel Lima Rosa Mendonça e Almeida
Institution
Centro Hospitalar de São João
Country
Portugal
Tutor
Jaap ten Oever
Title of Project
Reducing empirical Ceftriaxone use for CAP in the Internal Medicine department
Date of submission
16/03/2024

Abstract

Introduction
Ceftriaxone is widely used for the treatment of community-acquired pneumonia (CAP) even where local
recommendations suggest narrower-spectrum options. We designed an intervention aiming to reduce
empirical ceftriaxone use for CAP by 50% over 6 months in a tertiary hospital’s internal medicine (IM)
department with 200 beds.

Methods
A local CAP guideline has existed since 2019 and recommends aminopenicillins + azithromycin for patients
requiring hospitalization. In February/2023, we discussed barriers and facilitators in a focus group with key
prescribers and analyzed them using the COM-B model. The intervention took place from May-October/2023:
short sessions in each sector directed to the identified barriers and focusing on education, persuasion and
modelling, along with follow-up with data feedback.
Outcome indicators were the percentage of CAP patients empirically treated with ceftriaxone and ceftriaxone
LOT for CAP/1000 discharged patients. Percentage of CAP empirically treated with broad-spectrum antibiotics,
CAP patients’ in-hospital all-cause mortality and length of stay (LOS) were chosen as balanced indicators.
Indicators were assessed in February (baseline) and May-October. All patients with CAP-specific ICD-10
diagnostic codes were selected and data on LOS, in-hospital mortality and antibiotic use was extracted from
the institutional data warehouse, followed by manual review.

Results
In February (baseline) and May/23 (while the first sector visits were still taking place), the proportion of
empirical ceftriaxone was 21%, and we saw a decrease from June-October towards a median of 6,8%.
Ceftriaxone LOT showed a similar trend. Balanced indicators showed median values similar to baseline.

Discussion
Our intervention achieved the goal of a 50% decrease of ceftriaxone use for CAP and did not show signs of
increasing adverse outcomes. This project allowed us to test an effective and well received way of designing
AMS interventions driven by data and focused on behavioral changes and was the pilot for new similar projects
already underway.

31
Stage 1 Introduction: description of the setting and the AMS program in place and identification of the
problem
I work in the Infection Control and Antimicrobial Stewardship Department (UPCIRA – “Unidade de Prevenção e
Controlo de Infeção e Resistência aos Antimicrobianos”) in a tertiary hospital with 1100 beds. Our internal
medicine (IM) department is comprised by approximately 200 beds divided between 4 sectors. Before this
project, AMS interactions with the IM department consisted almost entirely in validation of restricted
antibiotics (the list of restricted antibiotics does not include ceftriaxone), automatic stop-orders for all
antibiotics after 7 days and sporadic discussion of clinical cases.

Regarding the antibiotic treatment for CAP, both national (2011) and local (2019) guidelines are available. Both
these guidelines recommend the use of ampicillin or amoxicillin, with or without clavulanic acid and either in
monotherapy or combined with azithromycin according to the severity of the infection. Ceftriaxone is featured
only as an alternative for penicillin-allergic patients. The local guideline was produced by a multidisciplinary
team including elements from UPCIRA, Microbiology, IM, intensive care and emergency departments.
The choice of our project was influenced by our local data. A previous study performed by us (Gorgulho A et al,
Antibiotics, 2023) analyzed 450 consecutive ceftriaxone empiric prescriptions from January and May 2021 and
found that half were inadequate and that most of that inadequate use was for CAP and in the IM department.
This was consistent with PPS data – for example, we used global-PPS to assess antibiotic use in the IM and
Intensive Care departments in February 2023 and ceftriaxone corresponded to 15% of all prescriptions (the
second highest antibiotic in proportion in the WATCH category).
Therefore, we decided to perform an intervention that aimed to reduce empiric ceftriaxone use for CAP by 50%
over 6 months in the IM department.

Stage 2 Measurement
To assess the results of our intervention, we chose the following qualitative (1) and quantitative (2)
indicators:

(1) Percentage of patients with CAP admitted to the internal medicine department empirically started on
ceftriaxone:

* A patient was considered “empirically started on ceftriaxone” if ceftriaxone was started or maintained on first
evaluation by a practitioner from the IM department
** Patients who were already on antibiotics for CAP on admission (for example if they were transferred from
other institutions) were excluded from the denominator. Patients who were not treated with antibiotics (for
example if they were admitted for palliative care or if viral pneumonia was suspected) were included in the
denominator.

(2) Ceftriaxone LOT for CAP /1000 discharged patients

We also collected data for balanced indicators to try to assess the safety of our intervention:

(3) Percentage of patients with CAP admitted to the internal medicine department empirically started on
broad-spectrum antibiotics (using the same definition of “empirically started on” and the same criteria for
inclusion in the population as for indicator 1):

*** Broad-spectrum antibiotics: piperacillin/tazobactam, anti-pseudomonal cephalosporins, fluoroquinolones

or carbapenems

32
(4) In-hospital all-cause mortality among all CAP patients admitted to the IM department during the study
period (using the same criteria for inclusion in the population as for indicator 1);

(5) Median hospital length of stay in days for CAP patients admitted to the IM department during the study
period (using the same criteria for inclusion in the population as with indicators 1).

CAP patients were identified by post-discharge diagnostic codes. Data was extracted automatically from our
hospital’s data warehouse and included local patient ID, diagnostic code, date of admission, length of stay, use
of ceftriaxone (yes/no), use of broad-spectrum antibiotics as defined above (yes/no), date of first and last
administration of ceftriaxone and outcome (death, discharge or other)

All identified patient files were reviewed manually according to the following scheme:
(1) Does the patient belong to the study population (admitted to the IM department with CAP without having
previously started treatment) » If positive, include in denominator and proceed to further evaluation; If
negative, exclude from study
(2) Were ceftriaxone or broad-spectrum antibiotics prescribed for CAP? » If negative, exclude from numerator
(3) Were ceftriaxone or broad-spectrum antibiotics started or continued by the IM team on first evaluation » If
negative, exclude from numerator

A one-month audit was performed (February/2023) and found empiric ceftriaxone use in 21% of the 60
patients identified with CAP (table 1)

Stage 3 Barriers and facilitators

We used a focus group approach with specific elements of each of the four IM sectors designated by the head
of department. After a brief presentation of our local data on ceftriaxone use, on local epidemiology and on
our hospital’s CAP protocol, two questions (transcribed below) were asked which were answered in turns by
each of the four participants. The answers were written down during the interview, summarized and sent back
in the same week via e-mail. All participants confirmed that this summary reflected their opinion and their
answers during the focus group.

(1) “In your understanding, which are the main barriers to using protocol first-line antibiotics for CAP over
ceftriaxone?”
- Most prescriptions are made in the emergency department and attending teams are reluctant to change
them;
- Some practitioners do not trust our protocol’s first line choices for CAP;
- Ceftriaxone can be given once per day and has no need for renal adjustment;
- The protocol is not widely known inside the IM department;
- Some IM practitioners are not aware of the disadvantages of using ceftriaxone over protocol first-line
antibiotics.

(2) Which resources can be used to improve the use of protocol first-line antibiotics for CAP over ceftriaxone?
- Most practitioners in the department are motivated to improve their practices regarding antibiotics;
- All four sectors hold weekly reunions where good clinical practices are often discussed;
- All prescriptions on admission to the IM department are reviewed by and IM practitioner (either in the ED by
IM practitioners in emergency duty or immediately after admission to the department by the internal residency
team).

We analyzed the barriers according to the COM-B model:

- Psychological capability (lack of awareness of protocol or of disadvantage of using ceftriaxone)
- Social opportunity (reluctance to changing prescription started in ER)
- Physical opportunity (access to guidelines)
- Automatic motivation (lack of habit of changing prescription started in ER)
- Reflective motivation (lack of trust in protocol options, posology convenience)

33
Stage 4 Development and execution of the stewardship intervention
Our intervention consisted on:

(1) An initial presentation in each of the four sector meetings focusing on.
- Education (to address reflective motivation, psychological capability): local CAP epidemiology and protocol
first-line options were discussed
- Persuasion (to address reflective motivation, automatic motivation): we approached disadvantages of using
ceftriaxone over first-line options, local epidemiology showing increasing gram-negative resistance and plans
for future audit of ceftriaxone prescriptions with data feedback. We proposed early switch from protocol first-
line options to oral therapy as an alternative to the posology commodity of ceftriaxone.
- Modelling (automatic motivation, social opportunity): we invited senior prescribers present in each meeting,
namely those who had been involved in the elaboration of the local protocol, to give their opinion and to
reinforce protocol-according practices and to encourage active revision of prescriptions started in the ED.

(2) We performed new brief presentations in the sector meetings in September and November with data
feedback using the indicators described above

Our intervention was also designed according to the facilitators proposed in the focus group:
- We used the IM sector meetings as a platform for our intervention
- We took advantage of the team’s motivation for good practices in deciding the tone of our message and in
asking role models to provide their opinion.
- Since all ED prescriptions would be reviewed on admission by IM practitioners, we were able to focus our
resources in intervening in the IM department and did not need to approach and train the ED teams, which
would pose logistic problems

Stage 5 Evaluation of the stewardship intervention

We used the same indicators as described in stage 2 for follow-up measurements.
The intervention was initiated in the last week of April and the first two weeks of May. We collected data
between May and October (figure 1 and figure 2).
In May/23, while the first sector visits were still taking place, the proportion of CAP patients empirically started
on ceftriaxone remained at 21%. From June to October, we saw a steady decrease towards a median of 6,8%
(with an outlier in September – 17,5%, which was also below the results from February and May). Ceftriaxone
LOT for CAP/1000 discharged patients showed a similar trend. Although we saw an initial increase in all-cause
mortality and proportion of CAP patients empirically treated with broad-spectrum antibiotics, the median from
June to October was 17,5% and 19,6% respectively, which were similar to the initial audit.

Conclusions
We were able to achieve our goal of decreasing empiric ceftriaxone use for CAP in the IM department by 50%
over 6 months. Except for September, we found a significant decrease in ceftriaxone use for CAP. It is possible
that the second round of participation in the sector meetings in September contributed to the absence of
ceftriaxone prescriptions for CAP in October, which would be indicative of the impact of our message but
would raise questions on the duration of the effect and would suggest the need of a continued mechanism of
discussion and feedback.
This model of data-driven behavioral intervention was well received by the IM department. During our visits to
the sector meetings we were open to listen to individual practitioner’s opinion of our intervention and it was
mostly seen as positive and as more effective than relying only on sporadic communication and antibiotic
restriction. Practitioners discussed feeling increasingly aware of the problem and more motivated to changing
antibiotics initially prescribed in the ED when necessary. Also, they felt that ceftriaxone use for other infections
was likely reduced due to spillover effect from this intervention and encouraged us to collect data and to
intervene in respiratory infections other than CAP. In general, the possibility of receiving feedback on adequacy
of antibiotic use for specific infections was highlighted as a particularly positive aspect.
Finally, we were able to identify other opportunities for improvement in CAP management during the project
(namely delayed or absent IV to oral switch and a high proportion of patients managed with broad-spectrum
antibiotics such as piperacillin-tazobactam).

34
As next steps:
- Our results will be presented at ECCMID 2024 (Poster number: P2698)
- We aim to present the results of our intervention to the IM department as a written report sent to the head
of department and as an oral presentation to the whole department.
- We have assembled a team with elements from different departments who are motivated for quality
improvement in antibiotic use and who received training in a local course included in the DRIVE-AMS project.
This team, with our coordination, has started to design an intervention in the IM department focused in
improving IV to oral switch for CAP and a separate intervention for diminishing empiric ceftriaxone use for CAP
in the ICU department.
- We are planning to provide regular feedback on ceftriaxone use for CAP from here on, which will be a part of
a broader plan of giving feedback on antibiotic use to all hospital departments
- For the two above objectives, we aim to make data on antibiotic use by focus of infection easier to collect. We
are working with the epidemiology and codification departments in order to simplify the process and we will
try to suggest developments to the IT team that manages our AMS electronic platform.

As limitations:
- We were not able to perform a formal structured interview according to one of the frameworks proposed by
the study due to logistical reasons.
- Making the CAP protocol more accessible to practitioners by simplifying our hospital’s document
management system (environmental restructuring) would likely have helped our intervention. However, it was
not possible to articulate with the IT department to proceed with this part off the project.
- We decided not to intervene in the ED because we did not have the resources to do so and because, as
pointed out in the focus group, all prescriptions would be reviewed by an IM practitioner on admission to the
department. However, extending our intervention to the ED would probably have reinforced our positive
results.
- Despite automatic acquisition of data from our data warehouse, data processing implied manual review of
each patient file and was cumbersome. This likely had a negative repercussion on the quality of data and the
number of analysis we were able to perform. For example, we were only able to audit one month of
prescriptions to establish a baseline, although we are convinced that collecting data from other pre-
intervention months would have given us better baseline data.

As a final and personal note, acquiring the skills to develop this project and having adequate support and
mentorship for it was, in my view, the most valuable aspect of the AMS Certificate (in par with the social and
networking components of the certificate, both with colleagues and faculty members, and the unforgettable
experience of developing bonds of friendship with such a diverse group along different European cities).
I opted to join the certificate instead of giving the first steps towards a PhD because at the time I prioritized
acquiring new skills over obtaining academic achievements, and my expectations were absolutely fulfilled. I
have indeed learned a new and better approach to AMS which has the potential of filling many old gaps in my
local context and which is already shaping the way I work.
I would particularly like to thank my tutor, Jaap ten Oever, for the quality input and support he gave me during
the project and for the constant availability and friendship. I would also like to thank Dr. Nuno Pereira, UPCIRA
coordinator, for his central role in designing and adjusting the intervention, for participating in the
implementation and for allocating the resources to make it possible, and Prof. Dr. José Artur Paiva for key
inputs in designing the intervention and for being instrumental in linking our department to the DRIVE-AMS
project. Finally, I would like to acknowledge the facilitating role of the head of the Internal Medicine
Department (Dr. Jorge Almeida) and of the IM department practitioners that participated in the focus group
and helped reinforce our message during the sector meetings (Dr. Paula Dias, Dr. António Vieira Lopes, Dr.
Maria João Lima and Dr. Edite Pereira)

35
Table 1 – Baseline audit

Indicator Baseline (Feb/23, n=61)

Empiric ceftriaxone use for CAP 13/61 (21,3%)
Ceftriaxone LOT/1000 discharged patients 86,1
Empiric broad-spectrum use for CAP 14/61 (23%)
All-cause in-hospital mortality (CAP patients) 18%
Median LOS of CAP patients 7 days

Figure 1 – Evolution of result indicators

Figure 2 – Evolution of balanced indicators

References
Gorgulho A, Cunha F, Branco E et al. Appropriateness of Empirical Prescriptions of Ceftriaxone and
Identification of Opportunities for Stewardship Interventions: A Single-Centre Cross-Sectional Study. Antibiotics
(Basel). 2023 Feb 1;12(2):288. doi: 10.3390/antibiotics12020288.

36
ESCMID AMS Certificate Program
My Project Report

Name
Joana Catarina Gouveia Batista Garnel
Institution
Hospital Prof. Doutor Fernando Fonseca
Country
Portugal
Tutor
Pilar Retamar-Gentil
Title of Project
Third-day AMS interventions in a medical ward
Date of submission
17th March 2024

Abstract
Introduction
De-escalation (ADE) and IV-oral switch are antimicrobial strategies that reduce broad-spectrum agents use and
have impact in quality of care, reducing emergence of resistance, length of stay and morbidity related to
antibiotics. In my institution, baseline evaluation showed high broad-spectrum antibiotics use and low adherence
to IV-oral switch and de-escalation practices in amoxicillin-clavulanate, ceftriaxone and piperacillin-tazobactam
prescriptions. The aim of the project was to explore and implement AMS interventions in antimicrobial de-
escalation (ADE) and IV-oral switch in an internal medicine department.
Methods
Baseline evaluation included a 3-month audit of prescriptions of study antibiotic using quality indicators defined
as number of patients on antibiotic who were streamlined or switched to oral (numerator)/ total number of
patients on antibiotic (denominator). Previous global adherence to strategies was 17,79%. Barriers to adherence
were evaluated using a focus group with IM prescribers and evaluated using COM-B and Flottorp models.
Interventions included: construction of IV-oral criteria tool and empirical therapy and de-escalation guidelines,
clinical sessions to present guidelines and benefits of AMS strategies and 72-hours post-prescriptional audit and
feedback for amoxicillin-clavulanate, piperacillin-tazobactam and ceftriaxone prescriptions.
Results
Overall adherence to de-escalation and oral switch raised to 61,72%. Quality indicators: ADE was performed in
54,7%, 52,8% and 64,3% of amoxicillin-clavulanate, ceftriaxone and piperacillin-tazobactam prescriptions. IV-oral
switch was performed in 45,28%, 45,28% and 50% of amoxicillin-clavulanate, ceftriaxone and piperacillin-
tazobactam prescriptions. Clinical sessions and meetings were attended by more than 80% of staff. Secondary
adverse outcomes (readmission, mortality or adverse effects due to interventions) were not observed.
Conclusions
The interventions resulted in substantial increase in quality of antibiotic use. Implementation based on baseline
evaluation and behavioural change provided good results and was a driver to the development of our AMS
program and to the expansion of interventions to other departments.

Stage 1: Introduction

I work at a secondary hospital with 804 beds, where current AMS interventions consist mainly of post-
prescription validation of restricted use antimicrobials, including carbapenemes and fluoroquinolones.
Here, patients are admitted thourgh the emergency room, where they remain for one or two days, even if
clinically stable. The ER unit has rotating medical staff and high patient volume (75- total bed capacity) and
severity. Most empirical antibiotics are started under this high pressure environment, where collecting adequate
prior cultures is also a challenge. The main consequence is empirical therapy based on broader-spectrum
antimicrobials, that isn’t reassessed after patients are admitted in internal medicine wards. A 48-hour
reassessment of antibiotics is considered good practice to check for usefulness (whether it is infection or other
diagnosis in which antimicrobials are unnecessary) and effectiveness, including defervescence and other signs of
improvement. Most cultures have a turn-around time of 48 hours, accelerated by MALDI-TOF®, and reasoned

37
decisions can then be made earlier. There isn’t a culture or policy of antimicrobial reassessment among doctors
in my institution and antimicrobial de-escalation or IV-oral switch are infrequent.
Prior to implementation, antimicrobial use analysis highlighted a higher number of ceftriaxone and piperacillin-
tazobactam prescriptions (2466 and 2254 patients in 2021, respectively), less amoxicillin-clavulanate and
ampicillin prescriptions (1625 and 249 patients in 2021) and even less oral alternatives prescriptions (702
cefuroxime, 704 oral amoxicillin-clavulanate and 227 oral amoxicillin patients). The aim of the project was to
explore antimicrobial stewardships interventions and investigate its effectiveness in optimization of antibiotic
use in antimicrobial de-escalation and IV-oral switch. After retrospective measurement, a quasi-experimental
prospective study in patients admitted to one internal medicine department was implemented.

Stage 2: Measurement

To evaluate de-escalation and IV-oral switch practices in this internal medicine department, I performed a 3-
month retrospective audit of amoxicillin-clavulanate, ceftriaxone and piperacillin-tazobactam prescriptions using
the patients’ electronic clinical file, in a total of 208 prescriptions, and evaluated: (A) acquisition of infection
(community, nosocomial or long term care facility - LTCF), (B) site of infection, (C) if empirical or directed therapy,
(D) causative agent when isolated in microbiological study, (E) if therapy was appropriate according to guidelines,
site and required antimicrobial spectra, (F) if de-escalation was performed and (G) if IV-oral switch was
performed. To overcome potential bias, the period of evaluation is analogous to the period of the project launch
in 2023, but in 2022.

For amoxicillin-clavulanate, 78 patients’ prescriptions were evaluated, for community-acquired infections (n=
70, nosocomial n=3, LTCF n=5). 61 were respiratory infections, the remaining were: 8 urinary infections, 5 sepsis,
2 vaso-occlusive crisis, 1 bacteremia, 1 SSTI. Antimicrobial therapy was empirical in 73 prescriptions and was
considered directed in 5 (2 Streptococcus pneumoniae, 1 Klebsiella pneumoniae, 1 E. coli, 1 Enterococcus
faecalis). Amoxicillin-clavulanate was adequate in 55 patients. Therapy was de-escalated in only 2 patients but
we found 55 patients with motif to ADE (adherence rate to ADE in amoxicillin-clavulanate prescriptions was
3,63%) and IV-oral switch was performed in 16 patients (but possible in 63 patients).
Quality indicator as defined by: numerator: number of patients on amoxicillin-clavulanate whose therapy was
streamlined according to guidelines and/or microbiological result (n=2) and denominator: total number of
patients on amoxicillin-clavulanate in the internal medicine ward (n=78) is 2,56%. Quality indicator as defined
by: numerator: number of patients on amoxicillin-clavulanate whose therapy was changed from iv to oral (n=16)
and denominator: total number of patients on amoxicillin-clavulanate in the internal medicine ward (n=78) is
20,51%.

For ceftriaxone, 65 prescriptions were audited. Source of infection was mainly community-acquired (n=56,
86,15%). Respiratory infections represented 38,46% (n=25, of which 3 by S. pneumoniae [pneumococcal invasive
disease n=1]), other sites: urinary 36,92% (n=24), bacteremia (n=5), intra-abdominal (n=5), SSTI (n=2),
prophylaxis in esophageal varices (n=2) and unknown in 1. Therapy was empirical in 54, and directed in the
remaining (mostly Enterobacterales (n=7) and 2 Streptococcus pneumoniae). Ceftriaxone was deemed
appropriate in 31 patients and inappropriate in 34 (52,31%). ADE was performed in only 9 patients, wasn’t
considered adequate to streamline in 6 and the remaining 50 completed therapy only with ceftriaxone. IV-oral
switch was performed in only 8 patients.
Quality indicator as defined by: numerator: number of patients on ceftriaxone whose therapy was streamlined
according to guidelines and/or microbiological result (n=9) and denominator: total number of patients on
ceftriaxone in the internal medicine ward (n=65) is 15,25%. Quality indicator as defined by: numerator: number
of patients on ceftriaxone whose therapy was changed from IV to oral (n=8) and denominator: total number of
patients on ceftriaxone in the internal medicine ward (n=65) is 12,31%.

For piperacillin-tazobactam, I evaluated 65 prescriptions, of which 35 were directed at community- acquired

infections, 23 were of nosocomial source, 7 were LTCF-acquired. It was empirical in 45 cases (69,23%) and
directed in 20 cases (Enterobacterales ESBL n= 9, non-ESBL in 3, Pseudomonas aeruginosa in 4, Streptococcus
pneumoniae n=3, Haemophilus influenzae n=1). Site of infection was: respiratory in 33 (50,77%), urinary in 17,
intra-abdominal in 5, bacteremia in 4 (1 ESBL producer, 1 P. aeruginosa, 1 E. coli and 1 Haemophilus influenzae
both susceptible to ampicillin), febrile neutropenia syndrome in 3 (none had positive cultures) and SSTI (n=1).
Piperacillin-tazobactam was considered adequate in 32 patients (49,23%). ADE was performed in 2 patients and
IV-oral switch in 0. Quality indicator as defined by: numerator: number of patients on piperacillin-tazobactam

38
whose therapy was streamlined according to guidelines and/or microbiological result (n=9) and denominator:
total number of patients on piperacillin-tazobactam in the internal medicine ward (n=65) is 15,25%. Quality
indicator as defined by: numerator: number of patients on piperacillin-tazobactam whose therapy was changed
from IV to oral (n=0), denominator: total number of patients on piperacillin-tazobactam in the internal medicine
ward (n=65) is 0%.

Stage 3: Barriers and facilitators

To analyse barriers and strategies to overcome them, I used Flottorp Behaviour Change Techniques and the
COM-B to identify main domains and design the matching interventions. After identifying low adherence to
the studied AMS objectives, I arranged a first meeting with junior and senior prescribers these results were
presented and a discussion about streamlining and IV-oral switch were discussed before and after the results
presentation. 77,78% of senior doctors (7 out of 9) and all residents (but one, 96,3%) were present. All the
audience was unanimous saying they usually remember to switch to oral therapy when discharge is due and
seldomly before that. Junior doctors consistently explained their low adherence as a “lack of culture or habit”,
particularly regarding IV-oral switch. Regarding physician barriers to IV-oral switch, besides aforementioned
lack of habit, individual prescribers identified “lack of awareness” of defined switch criteria and “lack of
confidence” to apply the criteria. Patient-related factors were mainly “fear of aggravation” after switch,
“potential reduced efficacy of oral route, especially in elderly patients and bacteremic patients”. Organizational
barriers were the aforementioned “lack of culture” in department, “absence of specific IV-oral switch
guidelines” and “shorter time to make decisions or re-evaluate antimicrobial therapy”. After presentation, the
group manifested ignorance of low adherence rates (they thought they would be higher, despite the
mentioned barriers) and unawareness of benefits of IV –oral switch in patient and organizational points of view
(examples: phlebitis, length of stay, human resources demands, costs). Prescribers barriers to streamlining
were related to “practice regarding interpretation of antibiogram and choosing appropriate antimicrobials with
shorter spectrum” and “lack of knowledge of the benefits of de-escalation”, organizational barriers were also
“lack of guidelines”, “lack of peer support to de-escalate” and “lack of culture in evaluating antimicrobial
therapy when patient is improving” (which we found to also apply to IV-oral switch). Patient-related factors
was streamlining in bacteremic or more severe patients or in those patients who were still febrile or in whom
inflammatory markers had not reduced markedly reduced.
To address lack of habit and culture in de-escalation practices, action planning and habit formation, I promoted
discussion and reassessment of antimicrobial therapy with the prescribers, both orally and through a written
audit and feedback report in the patients’ file. Audit and feedback also acted as a modelling behaviour –
shaping knowledge by putting recommendations for switch in practice and gaining confidence in such changes.
Prescribers were also incentivised to have a written plan for every patient on antibiotics including re-evaluation
day, planned duration and potential oral drug to switch to. This acted as individual training. Also, to restructure
the lack of antimicrobial changing culture and enable clinicians to make decisions easily, simple tools to assess
IV-oral change criteria were available for all prescribers. These criteria were previously discussed in the first
meeting, so that all prescribers had agreed on them ahead.
To promote capability, small and large group clinical discussions addressed the difficulties identified by the focus
group: interpretation of microbiology results and appropriateness of antimicrobials. Also, to facilitate de-
escalation, direct rules of de-escalation, available as a list of preferred antimicrobials (from most to least
recommended, according to spectrum of activity and our current cumulative antibiogram report) in empirical
and directed therapy. For the identified patient-related factors, interventions were audit and feedback at
individual level, enabling prescribers to see how changes didn’t fail, promoting training sessions to discuss
benefits of de-escalation and IV-oral switch at patient and organization level (avoiding the natural consequences
of the unwanted behaviour) and evidence of bioavailability of drugs and non-inferiority of oral therapy in specific
infections. Individual audit and feedback was the main strategy used to induce repetition of behaviour and
creation of a habit of evaluation of antimicrobial at 72 hours. Also, I gave oral or written positive feedback when
de-escalation or switch were done by the prescribers without my intervention and incentivisation like “the
patient is better on oral therapy, maybe you can discharge him today”. Incentivisation and motivation is also
planned to be addressed by providing metrics per prescriber (individually) and per department and
benchmarking.

39
All these strategies were directed at making reassessment of antimicrobial prescriptions a policy in the internal
medicine department. Identified barriers, Flottorp Techniques and COM-B domains and planned interventions
are thoroughly described in Table 1 in Suppl. Material.

Stage 4: Development and execution of the stewardship interventions

To address changes in prescription, timely re-evaluation of antimicrobial prescriptions and early IV-oral switch
and de-escalation, I developed these interventions:
1. IV-oral criteria guidelines
2. Antimicrobial appropriateness guidelines for empirical therapy and de-escalation recommendations
according to site of infection (when empirical), microbiological results and current antimicrobial.
3. Clinical sessions to present IV-oral switch criteria and empirical and directed therapy guidelines.
4. Antimicrobial training rounds with real and simulated cases with junior and senior prescribers
5. Post-prescriptional audit and feedback
1. IV-oral switch criteria were developed according to best evidence available and made available for all
prescribers as a digital printable tool (see Supplementary Material). The choice of antibiotic would be guided
according to the microbiological isolate and antibiogram, if it existed. If there wasn’t a microbiology result to
guide therapy, the IV – oral switch was performed according to the site of infection. The following infections
were excluded: endocarditis, CNS infections including meningitis, necrotizing fasciitis, Staphylococcus aureus
bacteremia. Although there already exists favourable data concerning IV-oral switch in Staphylococcus aureus
bacteremia and streptococcal endocarditis, these were excluded due to the need to exclude complications
before switch, that was considered beyond the scope of the current project.
2. Local guidelines for appropriate empirical therapy according to site of infection were developed according to
national (when available) or international guidelines.
De-escalation was performed in three different settings, which had specific recomendations for prescribers and
antimicrobial stewards: 2.1 Stopping combination therapy: if combination therapy is being used, the second
agent can be reassessed at the third day of therapy. This will include discontinuation of: macrolides in
community-acquired pneumonia if Legionella pneumophila is not suspected or diagnosed; vancomycin if MRSA
nasal colonization is negative in pneumonia, and is not suspected nor the causative agent; aminoglycosides were
started as adjunctive therapy2.2 If directed therapy is possible: if microbiology-guided therapy is possible due to
positive microbiology results, with adequate samples according to site of infection. The narrower spectrum
antimicrobial would be used (see list in Supl. Material). Oral switch can also be considered. 2.3 If empirical
therapy is to be continued without positive microbiology results: de-escalation depends on these factors:
confirmation of infection (no diagnosis other than infection can explain the patient’s clinical findings); the patient
is not deteriorating; initial empirical antimicrobial therapy had broader or inappropriate spectrum and/or
empirical antimicrobial therapy is not according to guidelines.
3. There was a clinical session to present AMS objectives and scientific evidence and one describing the planned
interventions. The first one happenned on the 16th of March of 2023, joining the weekly meeting of the
department and was attended by 91,18% of prescribing doctors and foundation year residents in rotation.
Overall assistance number was 31 doctors.
4. The second formal moment happened on the 7th of september of 2023 and guidelines for IV-oral switch, de-
escalation and appropriate empirical therapy were presented and discussed. Attendance rate was a bit lower
due to holidays of staff – 83,87%, so I then gathered more individually with smaller teams to discuss their cases
and specific antimicrobial prescription strategies (in substitution of the clinical round). Smaller group strategies,
though time-consuming, were more fruitful and had more visible impact in prescription practices.
5. Audit and feedback was performed at 72 hours of intravenous therapy for community-acquired infections –
mostly amoxicillin-clavulanate, ceftriaxone and piperacillin-tazobactam. For each prescription, a
recommendation was written in the electronic patient file with the preferred antimicrobial, route of
administration and duration. Acceptance rate (as defined by IV-oral switch or de-escalation performed by doctors
autonomously as per guidelines or proposed as part of audit and feedback intervention) and outcomes were
recorded 7 days and 60 days after recommendation.

40
Stage 5: Evaluation of the stewardship intervention

A total of 128 prescriptions of amoxicillin-clavulanate, ceftriaxone and piperacillin-tazobactam were evaluated

from September/2023 to January/2024. Acceptance rate as defined as IV-oral switch and/or de-escalation (as
defined) raised from 17,79% (total of patients who were de-escalated or switch to oral therapy in evaluation
phase) to 61,72%, n=79 (additionally, partial acceptance of switch, such as stopping combination therapy or
switching to oral/ de-escalation but not with the recommended antimicrobial per guidelines was seen in 12
patients, 9,375%). IV-oral switch was recommendable or recommended in 86 patients and performed in 62
patients (72,09%) and de-escalation was proposed in 60 patients and performed in 47 (78,33%). Median duration
of antibiotherapy was 7 days (IQR 7; 8.75). Median number of IV therapy-saved days was 3 days (IQR 0-5; mean
3,27 days). No readmissions, adverse effects or deaths were recorded due to interventions.
As for amoxicillin-clavulanate (n=60, of which 46 were IV and 14 were already oral), prescribed for respiratory
(n=46), urinary (n=11), SSTI infections (n=2) and 1 E. coli bacteremia. IV-oral switch (of the 46 IV prescriptions)
was recommendable in 39 and performed in 31 (79,49%). Therapy was empirical in 50 (83,33%) and directed in
remaining (E. coli, n=5, Klebsiella spp. n=3, Streptococcus pneumoniae n=1, Proteus n=1, Enterococcus faecalis
n=1, Haemophilus influenzae n=1). Therapy was considered appropriate in 51 (85%). De-escalation per guidelines
(in directed therapy or when lesser spectrum was recommended) was adequate in 19 prescriptions and
performed in 14. Quality indicator as defined by: numerator: number of patients on amoxicillin-clavulanate
whose therapy was streamlined according to guidelines and/or microbiological result (n=14) and denominator:
total number of patients on amoxicillin-clavulanate in the internal medicine ward (n=60) is 23,33%. Quality
indicator as defined by: numerator: number of patients on amoxicillin-clavulanate whose therapy was changed
from IV to oral (n=53, includes previous oral prescribed by physician) and denominator: total number of patients
on amoxicillin-clavulanate in the internal medicine ward (n=60) is 88,33%.

There were 53 prescriptions of ceftriaxone, mostly empirical (73,58%), others directed (24,53%) or prophylactic
(1,88%, high-risk variceal bleeding patients). Directed therapy was aimed at: E. coli, n=5, Streptococcus
pneumoniae n=2, other Streptococcus n=2, Klebsiella spp. n=2, Proteus n=1, Enterococcus faecalis n=1
(microbiologically inadequate). Site of infection was respiratory (n=23), urinary (n=16), sepsis and septic shock
(n=5), intra-abdominal (n=4), bone and joint (n=2), SSTI (n=1), upper respiratory (n=1).
It was deemed appropriate in 35 prescriptions (66,04%). IV-oral switch was reasonable in 31 patients and de-
escalation adequate in 35 patients. Quality indicator as defined by: numerator: number of patients on ceftriaxone
whose therapy was streamlined according to guidelines and/or microbiological result (n=28) and denominator:
total number of patients on ceftriaxone in the internal medicine ward (n=53) is 52,83%.
Quality indicator as defined by: numerator: number of patients on ceftriaxone whose therapy was changed from
IV to oral (n=24) and denominator: total number of patients on ceftriaxone in the internal medicine ward (n=53)
is 45,28%.

There were 14 prescriptions of piperacillin-tazobactam (66 prescriptions were excluded due to clinical
deterioration, nosocomial infection or isolates not susceptible to the antibiotics included). Therapy was directed
in 6 (Pseudomonas aeruginosa n=2, E.coli n=3, MSSA n=1, Listeria monocytogenes n=1, Morganella morganii
n=1). and empirical in 8 patients. It was considered appropriate in 71,43%. Site of infection was respiratory n=6,
urinary n= 6, SSTI n=1, bacteremia n=1. ADE was adequate in 6 (42,86%) and IV-oral switch was recommended
in 8 (57,14%). Quality indicator as defined by: numerator: number of patients on piperacillin/tazobactam whose
therapy was streamlined according to guidelines and/or microbiological result (n=9) and denominator: total
number of patients on piperacillin-tazobactam in the internal medicine ward (n=14) is 64,29%. Quality indicator
as defined by: numerator: number of patients on piperacillin-tazobactam whose therapy was changed from iv to
oral (n=7) and denominator: total number of patients on piperacillin-tazobactam in the internal medicine ward
(n=14) is 50%.
Attendance to meetings and clinical sessions was reported above. Guidelines/tools are available in Suppl. Mat.

41
 70 64,29
60 54,72 52,8
50
50 45,28 45,28

40
30
20,51
20 15,25 15,25
12,31
10 2,65 0
0
AAC IV-oral AAC de-escalation Ceftriaxone IV oral Ceftriaxone de- P/T IV-oral P/T de-escalation
escalation

Graphic 1: Baseline (orange) and post-implementation (green) quality indicators for de-escalation and IV-oral switch practices
(values in %, AAC: amoxicillin-clavulanate; P/T: piperacillin-tazobactam).

Conclusions

I present preliminary results which are favourable to the interventions in place. Future actions for the project
include: continuing active audit and feedback for other three months to consolidate results, give written and oral
feedback and benchmarking to prescribers with current results and check for new unidentified barriers and have
AMS champions in the department to keep the team motivated. This work will be submitted in ECCMID next year
with total results. Expansion to other Internal Medicine Units is currently being worked on, measuring the same
quality indicators.
Specific limitations to the current project were the difficulty to assess patients on their second or third day
because of days-off (monthly medical strikes in 2023, weekends, holidays…) and longer ED stays, EUCAST new
MIC determinants for oral beta-lactams (which didn’t allow IV-oral switch in some cases) and absence of
guidelines to IV-oral switch in sickle cell disease syndromes. As in other IV-oral switch interventions, antimicrobial
therapy was sometimes extended for more 1-2 days, conflicting with the idea of shorter is better that we tried
to put in practice as a secondary effect of the project. Very low baseline adherence to ADE and switch acted as a
facilitator but also as a barrier, especially for the more senior doctors.
The process of designing the intervention was very didactic, as I started off seeing piperacillin-tazobactam as the
only issue but, by using adequate measurements and audits, realized that others emerged and lower hanging
fruits were reachable such as IV-oral switch and ADE. Understanding barriers and designing tailored interventions
and clinical discussions were new tools that I developed with the project and used for other interventions in
place, such as S. aureus bacteremia program and empirical therapy guidelines. Also, the implementation of a
project as this, and the good results it had so far, made other departments eager to collaborate with us on their
issues and goals.
With the AMS Certificate, our AMS Program is now in a more mature phase, with more interventions in place as
mentioned and a more structured way of functioning, using basic implementation science that I wasn’t familiar
with before. I have to reckon the exceptional mentoring of Pilar Retamar-Gentil in the project, whose expertise
gave simple solutions to my implementation problems and also helped tremendously on the basics of designing
a scientific project, competencies from which I now benefit daily. Developing this project, getting to know so
many different settings with our faculty members in which AMS prospered and making friends with peers with
such diverse geographical and cultural backgrounds but who envision and work on a future with better use of
antibiotics were the absolute positive aspects of the Program.
As a final thank you note, I would like to acknowledge the Internal Medicine Department (Medicina 2), especially
its Head Dr. Fernando Aldomiro, for embracing the project and inciting his team to do better. They did!

42
Supplementary material

Flowchart 1: Design of planned audit and feedback intervention regarding audit and feedback at 72 hours of intravenous
antimicrobial therapy for community acquired infection.

IV-oral switch
If directed therapy is
possible
Streamline to narrow spectrum
Day 3 intervention
if oral switch is not feasible
Empirical antimicrobial
therapy is started in ER
Planning of therapy Adjust to guidelines according to
duration site of infection
If empirical therapy is to
be continued
IV- oral switch if possible

Table 1: Barriers, framework of behaviour change techniques according to Flottorp and Michie (COM-B), planned interventions.
Flottorp’s Behaviour Change
Barriers COM-B Interventions
Techniques
Use every strategy (audit and
feedback, guidelines, oral
Habit formation
8. Repetition Motivation positive reinforcement) to make
Lack of habit (also
and third day interventions a habit.
organizational) Feedback on
substitution Discuss patients on their third
behaviour
Opportunity day to remind prescribers to re-
evaluate therapy.
2.Feedback Feedback on Modelling: review prescriptions
and behaviour and suggesting oral switch and
Lack of knowledge
monitoring Capability de-escalation at patient level
of IV-oral switch
Instructions on Motivation (audit and feedback).
criteria
6.Shaping how to perform Promote training sessions for
knowledge behaviour prescribers.
4. Shaping Instructions on
Prescriber-related

knowledge how to perform Education, availability of criteria

Unawareness of IV- behavior Capability list for prescribers, audit and
oral switch benefits 12. Restructure the Enablement feedback for prescribers
Environment environment of discussing current criteria
change prescription
Infrequent
re-evaluation of
Promote daily evaluation of
antimicrobial 1. Goals and
Action Planning Motivation antimicrobial therapy, especially
therapy (when Planning
at its third day.
patient is
improving)
Train prescribers in
Difficulty in
interpretation in microbiological
interpreting
results, development of de-
microbiological
4. Shaping Behavioural escalation simple rules (such as
results and Capability
knowledge practice list of recommended antibiotics
antibiograms to
according to isolates).
streamline
Promote training sessions for
antibiotics
prescribers.
Elderly patients 2.Feedback Feedback on Review prescriptions and
Bacteremia and behavior suggesting oral switch and de-
Patient- Severe conditions monitoring Capability and escalation at patient level (audit
related Regarding Instructions on motivation and feedback).
streamlining: 6.Shaping how to perform Promote training sessions for
patient still septic knowledge behaviour prescribers.

43
 (febrile and/or
inflammatory 8. Repetition Behavioural
markers not and practice
markedly lower) substitution
Fear of aggravation
Lack of culture of
Use every strategy (audit and
re-evaluation of 8. Repetition Habit formation
feedback, guidelines, oral
antimicrobial and Opportunity
positive reinforcement) to make
therapy (also substitution Motivation
third day interventions a habit.
prescriber-related)
Alert for adverse effects of broad
spectrum antibiotics, such as
Clostridium difficile infections,
Information on multidrug-resistant agents risk,
5.Natural
Unawareness of IV- health and of prolonged IV therapy such
consequences Capability
oral switch or consequences as phlebitis or salt-water
streamlining when the disturbances.
12. Reward Opportunity
benefits behavior is not Incentivisation: realization that if
and threat
performed prescribers do more the
Organizational factors

behavior, shorter lengths of stay

and less complications will arise
for each patient.
Give positive feedback to
Lack of peer 3.Social Motivation
Social support prescribers when they perform
support support Capability
streamlining and IV-oral switch
Development of guidelines
4. Shaping
regarding IV-oral switch and de-
knowledge Behavioural
escalation (according to
practice
Capability microbiological results and, when
Unavailability of 12.
these aren’t available,
guidelines Environment Restructure the
Opportunity recommended antimicrobials
change environment of
according to site of infection).
prescription
Promote training sessions for
prescribers.
Provide metrics of prescription
Low dissemination
6. Comparison habits and objectives for
of current practices Demonstration of
of the Opportunity prescribers and benchmark with
and no the behaviour and
behaviour (Persuasion) other prescribers and other
benchmarking to social comparison
internal medicine wards in the
prescribers
institution.

Table 2: IV-oral switch criteria checklist

Oral alternative antibiotic with adequate oral bioavailability exists (see table 3)

Existence of signs of clinical improvement

Vital signs are stable (no hypotension, no need for inotropic support or fluid resuscitation)

Apyrexia for 48 hours (without antipyretics) and no hypotermia

Oral ingestion is secured (no vomit)

No signs of intestinal dysfunction (no diarrhea, ileus, enteric drainage; no malabsorption is suspected)

The infection is not one of the following: endocarditis, CNS infections including meningitis, necrotizing
fasciitis, Staphylococcus aureus bacteremia

44
Table 3: IV-oral switch: first-choice oral antibiotics according to initial antibiotherapy and site of infection.
ANTIBIOTICS SITE OF INFECTION ORAL ALTERNATIVES
Amoxicillin
Respiratory Amoxicillin-clavulanate
If allergy is detected: doxycicline
Amoxicillin
Amoxicillin-clavulanate
Amoxicillin-clavulanate
Urinary Cefuroxime
Trimethoprim-sulfamethoxazol
Ciprofloxacin
Flucloxacillin
Skin and soft tissue
Amoxicillin-clavulanate
Amoxicillin
Amoxicillin-clavulanate
Respiratory
Cefixime
If allergy is detected: doxycicline
Amoxicillin
Ceftriaxone Amoxicillin-clavulanate
Urinary Cefuroxime
Trimetoprim-sulfametoxazol
Ciprofloxacin
Flucloxacillin
Skin and soft tissue
Amoxicillin-clavulanate
If directed therapy is possible:
Amoxicillin
Respiratory Amoxicillin-clavulanate
If therapy remains empirical:
Levofloxacin
Amoxicillin
Piperacillin-tazobactam
Amoxicillin-clavulanate
Urinary Cefuroxime
Trimetoprim-sulfametoxazol
Ciprofloxacin
Flucloxacillin
Skin and soft tissue
Amoxicillin-clavulanate

Table 4: De-escalation when cultures are negative: first-line and alternative antimicrobial therapy according to site of
infection
Site of infection Appropriate antimicrobial therapy if negative cultures
Community-acquired First choice Amoxicillin/ ampicillin
pneumonia Alternative Doxycicline
Reassessment of macrolide if present in initial therapy
Skin and soft tissue First choice Flucloxacillin
If abscess is suspected: amoxicillin-clavulanate
Alternative Clindamycin
Urinary Directed therapy guided by cultures
Intra-abdominal First choice Amoxicillin-clavulanate
Ceftriaxone ± metronidazole
Bone and joint First choice Flucloxacillin + ceftriaxione

Table 5: Preferable antimicrobials for de-escalation when infectious agent was isolated in microbiology.
Microbiology results Preferred antimicrobials (from most to least preferable, as listed)
Gram positive cocci Penicillin, amoxicillin/ ampicillin, cefuroxime, trimethoprim-sulfamethoxazol, ceftriaxone,
ciprofloxacin
MSSA with bacteremia Flucloxacillin (only IV), cefazolin
MSSA without bacteremia Flucloxacillin, amoxicillin-clavulanate
Gram negative Fosfomycin (cystitis in female patients only), amoxicillin/ ampicillin, cefuroxime, amoxicillin-
clavulanate, trimethoprim-sulfamethoxazol, ceftriaxone, ciprofloxacin
Note: Antimicrobials in this list and respective order were chosen according to narrow spectrum and the institution’s annual
cumulative antimicrobial susceptibility report.

45
ESCMID AMS Certificate Program
My Project Report

Name
Kristina Öbrink-Hansen
Institution
Department of Infectious Diseases, Internal Medicine, Gødstrup Regional Hospital
Country
Denmark
Tutor
Nathan Peiffer Smadja
Title of Project
Improving appropriate empiric antibiotic treatment through the use of antibiotic standardized treatment
packages
Date of submission
07.03.2024

Abstract* (max 300 words)

To help improve choosing the correct antibiotic treatment, guidelines are of great importance. At the Central
Denmark Region, guidelines for empirical antibiotic treatment are available, including standardized treatment
packages (STP), to help physicians choose the right drug, dose, dosing interval and length of treatment.
However, these STPs have not been routinely used. The aim of this study was to increase appropriate antibiotic
treatment for suspected pneumonia, erysipelas, sepsis and urinary tract infections (PESU) through the use of
STPs.
The project was conducted at the Department of Internal Medicine, Gødstrup Regional Hospital, Denmark,
between October 2022 and October 2023. All medical doctors (MDs) in the department received information
about the project, training in using the regional guidelines and prescribing an STP. A pocket card with a
checklist for prescribing antibiotic treatment was developed and posters with a reminder to use STPs were put
up in all wards. Data regarding antibiotic treatment and the use of STPs was extracted for weekly analysis
through the regional business intelligence platform. All MDs received a monthly update on the project as well
as repeated information about the regional guidelines and how to use STPs. A point prevalence study (PPS)
regarding patients with PESU, the use of STPs and adherence to regional guidelines was performed before, six
and twelve months after the interventions.
During the project period, the percentage of patients with PESU receiving an STP and the percentages of
patients with PESU receiving antibiotics according to regional guidelines increased substantially.
In conclusion, teaching sessions, data on usage and repeated information regarding guidelines for antibiotic
treatment and STPs, as well as pocket cards and posters with information about the project led to a substantial
increase in appropriate treatment and the use of STPs for patients with PESU.

Stage 1. Introduction

I work as an ID physician in a general hospital with 300 beds. There are 90 beds at the Department of Internal
Medicine, divided in 5 separate wards. The number of internal medicine patients admitted often exceeds 90,
and patients are then located to other departments at the hospital. There is no AMS program, nor an AMS
team available at the hospital. There is, however, treatment guidelines for antibiotic therapy available. For
example, a regional treatment guideline regarding empirical antibiotic treatment for community-aquired
infections in immunocompetent individuals. In this guideline, there are so called Standardized-Treatment-
Packages (STP) available, to help physicians choose the right drug, dose, dosing interval and length of
treatment. The physician can prescribe these packages in the electronic medical record (EMR). For example,
there is an STP for mild pneumonia (CURB65 score 0-2) with iv. Benzylpenicillin 0,6g q6h for 48 hours, followed
by oral V-penicillin 800 mg q6h for another 72 hours.

46
Many patients admitted to the Emergency Department (ED) and from there, transferred to the Department of
Internal Medicine, are started on broad spectrum antibiotics (primarily piperacillin/tazobactam), that are not in
line with regional guidelines, and very few physicians use the STPs available. Patients admitted to the ED are
initially seen by an ED physician. If the patient is transferred to the Department of Internal Medicine,
a physician from the Department of Internal Medicine has to approve the treatment plan made, including
indication and choice of antibiotic treatment prescribed.

The aim with my project was to increase the prescription of appropriate empirical antibiotic treatment
according to the regional guideline for suspected pneumonia, erysipelas, sepsis and urinary tract infections
(PESU), through the use of STPs.
All antibiotic prescriptions in the hospitals and clinics throughout the region where I work are documented in
the EMR and may be extracted for further analysis through the regional business intelligence platform (the BI-
platform).

Stage 2. Measurement

To evaluate appropriate empirical antibiotic treatment for suspected PESU, a point prevalence study (PPS) was
performed one month before the interventions were initiated. All patients at the Department of Internal
Medicine on October 4th 2022 were evaluated. I assessed the number of patients on antibiotic treatment, and
for these patients I registered indication and the antibiotic prescribed. For patients with PESU I registered (1)
the appropriateness of the antibiotic prescribed (in accordance to regional guidelines or not) and (2) the use of
an STP if possible. Data was extracted from the EMR.

Quality Indicators (1)

Numerator: The number of patients with PESU with appropriate antibiotic treatment prescribed.
Denominator: The total number of patients with PESU.

Quality Indicators (2)

Numerator: The number of patients with PESU prescribed an STP.
Denominator: The total number of patients with PESU.

Results from the pre-intervention PPS are depicted in Table 1. Nearly 80% of patients with PESU received
antibiotic treatment according to the regional guidelines, but only 20% received an STP.

Furthermore, in October 2022, the weekly number of patients admitted with suspected PESU and the number
of these patients prescribed an STP were extracted from the BI-platform. Results, depicted in Figure 1, show
that on average, STPs were prescribed for approximately 11% and 16% of patients with pneumonia and
erysipelas. However, no patients with sepsis and UTI received an STP.

47
 Department of Internal 04.10.2022
Medicine (five wards)
Patients admitted (n) 93

Patients treated with 65 (70)

antibiotics (n(%))
Patients admitted with 35
PESU (n)
Patients with PESU 7 (20)
receiving an STP (n(%))
Patients with PESU 27 (77)
receiving appropriate
treatment, according to
regional guidelines
(n(%))
Table 1. Data from the point prevalence study before the intervention.
PESU: pneumonia, erysipelas, sepsis, urinary tract infection

Figure 1: Pre-intervention data from the BI-portal (October 2022). The weekly number of patients admitted
with PESU and the number of these patients receiving an STP.

48
Stage 3. Barriers and facilitators

Information about barriers and facilitators were obtained from the EMR during the PPS and from interviewing
my colleagues, using the COM-B model as a reference.

The following barriers were identified:

- Lack of knowledge on regional guidelines.
- Lack of confidence in making/selecting a diagnosis (primarily junior doctors)
- Physicians were not familiar with STPs.
- Physicians did not trust the STPs with regards to the choice of antibiotic and length of treatment (primarily
senior doctors).

The following facilitators were identified:

- Regional antibiotic guidelines are available, with link to appropriate STP.
- Using an STP will help the physician choose the right drug, dose, dosing interval and length of treatment.
- Using STPs is time saving.

Stage 4. Development and execution of the stewardship intervention

All medical doctors (MD) and nurses in the department received written and oral information about the
project. All MD received training in using the regional guideline and prescribing an STP. A pocket card with a
checklist for prescribing antibiotic treatment was developed and posters with a reminder to use STPs were put
up in all wards. Furthermore, junior doctors received training in decision-making regarding diagnostic work-up
and antibiotic treatment for patients with potential infections.
Data regarding antibiotic treatment for patients with PESU and the use of STPs was extracted weekly through
the BI-platform. All MDs received a monthly update on the project with presentation of the STP usage data
from the BI-platform as well as repeated written information about the regional guidelines and how to use
STPs.
Details about the timeframe for the interventions are depicted in Table 2. Intervention functions according to
the COM-B model are presented in Table 3.

49
 Intervention Timeframe
Information about the project to all doctors and October 2022
nurses in the department
Pre-intervention PPS November 2022
Presentation of data from the pre-intervention PPS December 2022
Development of pocket-card (checklist for December 2022
prescribing antibiotics)
Education and training of all MDs in using the January 2023
regional guidelines and prescribing an STP
Nudging posters regarding STPs put up in all wards March 2023
6 months post-intervention PPS April 2023
Junior doctor training: decision-making regarding June 2023
diagnostic work-up and antibiotic treatment for September 2023
patients with potential infections
12 months post-intervention PPS October 2024
Monthly oral update about the project and February – October 2023
presentation of STP data from the BI-platform
Monthly e-mail with repeated information regarding February – October 2023
regional guidelines and how to use STPs
Table 2. Interventions initiated during the project period.

Main Barriers Lack of knowledge

Lack of training
COM-B Component Capability - psycological
Intervention function Education
Training
Environmental restructuring (posters)
Table 3. intervention functions according to the COM-B model

Stage 5. Evaluation of the stewardship intervention

The PPS regarding patients with PESU, the use of STPs and adherence to regional guidelines was repeated after
six and twelve months (April 2023 and October 2023). The percentages of patients with PESU receiving an STP
increased from 20% to 63% and the percentage of patients with PESU receiving antibiotics according to
regional guidelines increased from 77% to 98% (Table 4).

Data from the BI-platform regarding the monthly use of STPs for patients with PESU are depicted in Figure 2.
During the study period, the percentage of patients with PESU receiving an STP increased substantially, with
the highest numbers registered for pneumonia and erysipelas (approximately 50% and 80%, respectively).

A questionnaire with 10 questions was sent to all MDs in May 2023. The doctors were asked about the AMS
project, the regional guidelines, the STPs and the pocket-card – if they felt they had received enough
information and if the interventions initiated helped them in their daily work. The feedback was overall very
positive. The majority of MDs felt the continuous information and the interventions initiated were sufficient
and helpful.

50
 Department of Internal 04.10.2022 04.04.2023 04.10.2023
Medicine (five wards)
Patients admitted (n) 93 114 117

Patients treated with 65 (70) 76 (67) 74 (63)

antibiotics (n(%))
Patients admitted with 35 55 43
PESU (n)
Patients with PESU 7 (20) 32 (58) 27 (63)
receiving an STP (n(%))
Patients with PESU 27 (77) 48 (87) 42 (98)
receiving appropriate
treatment, according to
regional guideline (n(%))
Table 4. Data from the point prevalence studies before, six and twelve months post the interventions.
PESU: pneumonia, erysipelas, sepsis, urinary tract infection.

Figure 2. The percentage of patients with pneumonia, erysipelas, sepsis and urinary tract infections, who were
prescribed an STP. Data are presented as monthly results. The dashed line represents the linear regression line.

Conclusions

The project has overall been very rewarding. It has made doctors and nurses aware of AMS; what it is, how
important it is and the consequences it may have. My biggest challenge has been trying to induce behavioral
change, especially among senior colleagues. I was a bit surprised about the initial resistance from this group of
colleagues, while on the other hand, junior doctors were very interested, eager and willing to learn about initial
antibiotic treatment and the STPs. Rome wasn´t built on one day, but I believe in the power of continuous
education and training as well as persistence. This project has given me the opportunity to continue my work
with AMS in the department, and the AMS certificate program has given me a solid foundation and the tools I
need to continue my work with this very important topic.

51
ESCMID AMS Certificate Program
My Project Report

Name
Kristýna Herrmannová
Institution
University Hospital Bulovka
Country
Czech Republic
Tutor
Petros Ioannou (Heraklion, Greece)
Title of Project
Implementation of local guidelines of Staphylococcus aureus bacteremia and ID consult as a standard
of care in University Hospital Bulovka
Date of submission
17.3.2024

Abstract* (max 300 words)

Background: Staphylococcus aureus bacteremia (SAB) constitutes a significant source of bloodstream

infections associated with high levels of morbidity and mortality. This investigation explores the
inaugural AMS initiative at University Hospital Bulovka, Prague. The facility has 850 beds and employs
565 physicians across various medical specialties. Despite the presence of an Infectious disease clinic
(approximately 100 beds), the tradition of consulting ID specialists in the hospital was previously
absent.

Methodology: Initially, a retrospective study was performed, and the magnitude of the problem of SAB
was identified. Analysis showed a high 30-day mortality rate among patients with SAB who received
treatment outside the ID clinic. Objectives were established using the SMART criteria. An online survey
was conducted to ascertain the barriers, utilizing the Flottorp framework for guidance. Upon
identifying these barriers, interventions were planned and executed over two sequential phases. The
efficacy of the intervention was assessed via process quality indicators, with 30-day mortality rates
serving as the outcome indicator. Pre- and post-intervention periods were analyzed using Fisher´s
exact test.

Results: The evaluation of the interventions was conducted in two phases. The initial intervention (the
creation and presentation of local guidelines for SAB) had no impact on patient mortality in the target
group or the overall patient cohort with SAB after three months. However, there was a significant
increase in adherence to a bundle of quality indicators in both observed groups. Due to the insufficient
effect, two additional interventions (physician education and unsolicited ID consultations) were
implemented over the next six months. Data analysis revealed a reduction in mortality in both groups,
with the initial goal being met for adherence to bundle and mortality rate in the target group.

Conclusion: Implementation of local guidelines and ID consult in SAB patients decrease 30-day
mortality in both studied groups.

52
Stage 1 Introduction: description of the setting and the AMS program in place and identification of the
problem

I serve as an Infectious Disease (ID) specialist at the University Hospital in Prague, Czech Republic,
which accommodates nearly 900 beds and encompasses most clinical departments except for
cardiothoracic surgery, neurosurgery, and hemato-oncology. The hospital lacked an AMS program and
a team when initiating the Antimicrobial Stewardship (AMS) certificate program. However, it houses
essential departments for an effective AMS program, including clinical pharmacy, hospital
epidemiology, microbiology, and infectious disease special. Previously, the hospital had not
implemented any AMS-related activities. There were no fundamental data for AMS, such as antibiotic
consumption data and good-quality surveillance reports of hospital-acquired infections (HAI).
Additionally, there were no established local antibiotic guidelines or local antibiograms for
antimicrobial resistance. The concept of Antimicrobial Stewardship (AMS) is not widely recognized. For
my project, I selected SAB, a diagnosis with high mortality that transcends the expertise of infectious
diseases and necessitates collaboration among all four key specialties for AMS activities. My hypothesis
was that due to the absence of local guidelines for SAB management and the lack of infectious disease
consultations across the hospital, the mortality rate for this diagnosis would be higher than the one
reported in the literature. I verified my hypothesis by conducting a six-month retrospective study (May
to October 2022) of patients with SAB throughout the hospital with the agreement of the ethical
committee, which confirmed my initial assumption. Through the analysis of collected data, high 30-
day mortality among SAB patients in the hospital was recorded, and especialy in SAB patients
treated outside the Infectious Disease Clinic. Outside the ID clinic, an ID specialist is not involved in
managing patients with SAB. This correlated with low adherence to selected quality indicators I had
established based on the local guidelines for SAB management that I created and were inspired by
international guidelines. The characteristics of patients in the retrospective group did not significantly
differ in basic attributes between patients with SAB in the entire hospital and those treated outside
the infectious disease clinic. In the group of all SAB patients, there were 56 patients, the gender
distribution was: males 51.8% vs. females 42.8%, age median (Q1-Q3): 77 (63-83) years, IQR (20) years,
IQR = Q3-Q1 and 66.7% of patients had a risk factor for complicated SAB. In the target group (patients
with SAB treated outside the infectious disease clinic), there were 33 patients, the gender distribution
was: males 54,5% vs. females 45,5%, age median (Q1-Q3): 77 (62-83) years, IQR (21) years, IQR = Q3-
Q1 and 100% of patients had a risk factor for complicated SAB. The adherence to the guidelines is
presented in Graph 1 – adherence to the whole bundle of indicators was 56% in all patients with SAB
and 50% in SAB patients treated out of ID clinic. (A description of process quality indicators is written
in the measurement part). The analysis of the retrospective study showed the overall 30-day mortality
in the hospital to be 41% (51% in all departments but the ID department vs. 26% in the Infectious
diseases department). During this project, many related steps became apparent, driven by the
project's activities. The project highlighted the absence of multidisciplinary problem-solving capability
and, secondly, the inability and sometimes resistance to cooperation between specialties. At the
project's outset, I felt isolated, lacking support from any quarter. Another major barrier was the lack
of dedicated time for the project, which at the beginning of the project was a maximum of 0,1 FTE.

Stage 2 Measurement
To evaluate my project, which is focused on the standard of care of the management of SAB, I choose
qualitative indicators to evaluate the management process and develop interventions to improve it. I
have determined 4 structure indicators: the presence of local or national guidelines for SAB and the
presence of infectious diseases physicians, microbiologists, and clinical pharmacists. I have created
local guidelines for SAB with the cooperation of clinical pharmacy and epidemiology to define the
standard of care for SAB in the hospital. I have chosen a bundle of 6 process quality indicators: Q1 -
adequate therapy started within 24 hours knowing the diagnosis SAB, Q2 - control blood cultures taken
within 48-72 hours after starting the adequate therapy, Q3 - performed transthoracic/transesophageal

53
echocardiography, Q4 - for patients with vancomycin therapeutic drug monitoring (TDM) performed,
Q5 - source control done if applicable and Q6 – Infectious diseases consultation written in the patient´s
file. All those were collected from every patient´s file manually. To evaluate the adherence to those
process indicators, I established as the numerator for the QI 1,2,3, and Q6, the number of patients in
whom adequate therapy was started within 24 hours/control blood cultures were taken, ECHO was
performed, and ID consult was written in the file. As a denominator for those, I used the total number
of patients with SAB in the described group. For Q4, as a numerator, I used the number of patients
with TDM performed for vancomycin and, as a denominator, the number of all patients treated with
vancomycin. For Q5, as a numerator, I used a number of patients with a control source and as the
denominator, the total number of patients where source control was applicable. As an outcome
indicator, I chose 30-day mortality. I set a goal using the SMART goal method. For the quality indicators
Q1-Q3 and Q5, my target was to achieve a 15% increase within 9 months. For the Q4 indicator, the
goal was to maintain adherence at the same level within 9 months. For the Q6 indicator, the goal was
to reach a 50% increase compared to the baseline value within 9 months. For the whole bundle, I set
up a goal to reach a 20% increase. I aimed to reduce mortality by 15% in both groups (all SAB patients
and the target group – patients treated outside the ID clinic) under observation within 9 months. (Table
3 to see the goal for each indicator).
I did the prospective part in 2 steps. I evaluated the data 3 months after implementing the first
intervention and compared them with the retrospective data using Fisher´s exact test to see if there
was a statistically significant increase in adherence to six of the process indicators, the bundle of them,
and a statistically significant decrease in 30-day mortality of the patients. After that, I performed
another 2 interventions for 6 months and compared the data at the end of the prospective part with
retrospective data the same way as after 1st intervention. In every phase of measurements, I looked at
all patients with SAB and on the subgroup of the patients treated with SAB outside the Infectious
disease clinic, which for this project was the target group for improvement. To describe the groups’
demographics, I used gender differences, the median age, the interquartile range, and the risk for
complicated SAB.

Stage 3 Barriers and facilitators

The main barrier to my project was my lack of time to manage it effectively. Despite my efforts to
consistently explain the project's importance to my supervisor, I succeeded only in October 2023, at
the end of the project's implementation phase.
Before sharing the local guidelines with the hospital, I conducted an online survey targeting the
hospital staff. This survey aimed to gather insights on the clinical importance of SAB and their
understanding of the recommended practices according to the international guidelines. Given my
circumstances, an online survey was the most feasible option. Delivering the survey to the doctors was
not straightforward due to a lack of direct contact. Therefore, I sent it to the heads of departments
and asked them to forward it to their doctors. The survey was conducted in November 2022, about 2-
3 months before the guidelines were presented. The response rate was disappointingly low (45 doctors
out of 565, 8%), and I didn't have the opportunity, time, and space to conduct another survey. Looking
back, conducting interviews would have been a more effective way to understand the challenges and
what could help, but I couldn’t do so then. I used the Flottorp framework as a guide to identify my
barriers and facilitators.
I still used the survey responses, even though there weren't enough to fully represent the hospital
professionals' opinions. A full analysis of the barriers can be found in Table No 1. I chose the main
ones to focus on from all the barriers identified in my project. The first major barrier is the doctors'
lack of knowledge about managing SAB and their limited understanding of its clinical importance,
presented by the absence of local guidelines. The second major barrier is the hospital's lack of
teamwork across different departments (especially microbiology with others) regarding the proper use
of antibiotics for managing SAB, together with the lack of ID consults in the hospital. The third major

54
barrier is that the hospital's leadership does not prioritize the use of antibiotics as a key issue to address
– the absence of an antibiotic hospital plan or AMS team. As a facilitator for my project, I have found
the presence of all the specializations that were supposed to be members of the AMS team, the
presence of the pharmacology laboratory for TDM testing and at the beginning of my project, also
financial support from the hospital leadership to be able to study AMS certificate program, but as the
main facilitator I have found was my motivation to run this project the best way possible.

Stage 4 Development and execution of the stewardship intervention

Table No. 2 describes all the interventions I have performed to achieve the goal of my project. As the
main ones, I identified 3 of them I will describe in more detail. The first step involved was developing
local guidelines for SAB with the assistance of a clinical pharmacist and other infectious disease
specialists to fulfill critical structure indicator. These guidelines were presented to the entire hospital
9. January 2023 through a hospital-wide seminar, which was also recorded and made into a video
available to all hospital staff on the intranet accompanied by the recommended local guidelines and
all the presentations in pdf-form to be downloaded if wanted (1st intervention). Following the
presentation of the guidelines and their incorporation into the hospital's official documentation (done
30th January 2023), space was allowed to observe whether this intervention would influence the
involvement of an infectious disease specialist in the care of patients with SAB, which had previously
been rather random. Specifically, it aimed to see if hospital physicians would begin to utilize the
guidelines, which include a component advocating for an infectious disease consultation (in
attachment: first page of guidelines to show). It happened from February to April 2023. The effect was
just minimal. In all SAB patients, the adherence to bundle increased from 56% to 69%, mostly thanks
to the increase of Q6, and the 30-day mortality remained the same (Table 1). For the target group, the
adherence to the bundle increased from 50% to 62% (also thanks to the increase in adherence to Q6),
and the 30-day mortality decreased insignificantly (Table 2).
Consequently, two interventions were established to contribute to physicians' education and directly
increase the presence of an infectious disease specialist in managing SAB. First, an educational seminar
on the existence and application of the SAB local guidelines was conducted sequentially at all hospital
clinical departments during the time following the morning meeting, when most physicians are present
(minimum 80% of participation). Those seminars were performed from May 2023 to August 2023. The
second intervention was unsolicited ID bedside consultation, which I always performed. It was
possible to perform it thanks to the newly established system of electronic reporting of Staphylococcus
aureus in blood culture from microbiology to me (established in the middle of March 2023). The consult
was always performed in coordination with the attending physician. This intervention started at the
beginning of May 2023 and lasted 6 months until October 2023 (included). After this period, the
measurement was done again. (to see in the Evaluation part). The aim was to enhance adherence to
the created guidelines. From November 2023, unsolicited consultations were discontinued to reassess
whether the implemented interventions were effective. Based on the outcomes of this period, the
interventions will be re-evaluated in May 2024, and further actions/interventions will be planned.

Stage 5 Evaluation of the Stewardship Intervention

At the beginning of the project, I found 4 structure indicators, from which three were present (ID
specialist, microbiologist, and clinical pharmacist). Initially, I created local guidelines for SAB with the
cooperation of other specialists. With this action, I fulfilled the last defined structure indicator. My
project was evaluated in two phases for process and outcome indicators. Subsequent analysis of the
1st intervention's impact revealed minimal effects on the 30-day mortality rate across the hospital,
which remained unchanged. The adherence rate to the treatment bundle increased significantly from
56% to 69% (p=0,0375) (see Table 1). In the group of all SAB patients, there were 27 patients, the
gender distribution was: males 77,8% vs. females 22.2%, age median (Q1-Q3): 65 (48-77) years, IQR

55
(28) years, IQR = Q3-Q1 and 76% of patients had a risk factor for complicated SAB. Focusing specifically
on patients with SAB treated outside of the Infectious Diseases (ID) department—the primary target
group of this project—the mortality rate saw a marginal reduction from 51% to 47% (p>0,9999).
Adherence to the bundle improved from 50% to 62% (p=0,0124) (see Table 2). There was a statistically
significant increase in adherence to the Q6 indicator. In the target group, there were 17 patients; the
gender distribution was: males 76.5% vs. females 23.5%, age median (Q1-Q3): 70 (62-73) years, IQR
(11) years, IQR = Q3-Q1 and 82,3% of patients had a risk factor for complicated SAB. All the data are
presented in Tables 1 and 2. Observing that the initial intervention did not yield the anticipated
outcomes, I implemented two further interventions. The first aimed at educating physicians about the
guidelines' existence, application, and importance for patient outcomes to use it. I successfully
engaged most physicians (at least 80%) by conducting sessions in each department following morning
meetings. Concurrently, I introduced unsolicited ID consultations for SAB patients. These combined
interventions led to in all patients with SAB increase in bundle adherence from 56% (before 1st
intervention) to 75% by the end of 9 months after presenting the guidelines, which was a statistically
significant increase (p <0,0001), also almost reaching the goal for the bundle. There was a statistically
significant increase in adherence to quality indicator number 6, together with achieving the goal for
this quality indicator. There was no significant increase for the rest of the quality indicators, and the
goal wasn’t achieved (see Table 3). The 30-day mortality rate decreased from 41% to 28%, not
statistically significant (p=0,2275). In the group of all SAB patients, there were 53 patients, the gender
distribution was: males 69.8% vs. females 30.2%, age median (Q1-Q3): 68 (53-80) years, IQR (27) years,
IQR = Q3-Q1 and 62,3% of patients had a risk factor for complicated SAB. Among the targeted patient
group, adherence to the guidelines increased from 50% to 71%, which was statistically significant (p
<0,0001), and reached the planned goal of the increase of 20%. More critically, the 30-day mortality
rate in this group decreased from 51% to 32%, which wasn’t statistically significant, but it had reached
the planned goal of my project. Again, there was a statistically significant increase in adherence to the
quality indicator number 6 (p <0,0001), also achieving the goal for this indicator. In the target group,
there were 41 patients; the gender distribution was: males 65.9% vs. females 34.1%, age median (Q1-
Q3): 67 (54-79) years, IQR (25) years, IQR = Q3-Q1 and 70% of patients had a risk factor for complicated
SAB.

Conclusions

Though it may sound overly sentimental, participating in the AMS certificate program has been life-
changing, overwhelmingly in a positive sense. For years before the program was introduced, I had
longed for something exactly like it. However, such educational opportunities were not available in the
Czech Republic. The real-world outcome of my project has been multifaceted. Firstly, I hope I have
planted a strong and healthy seed for AMS activities at the University Hospital Bulovka in Prague,
where, as a side effect of my project, an AMS team was established and, from March 2024, will be
officially integrated into the hospital's structure, with me becoming its coordinator. It is the first official
AMS team in the country. Still, it will require a lot of work, so the team works as it should, but the
people around me seem as motivated as I am. Secondly, the project forced me to learn to stand up for
myself and what I believe is right despite an incredible amount of small and large obstacles in the form
of technical, time, and especially human barriers. Thirdly, I was taught the methodology of measuring
the activities we undertake in our work, and I applied this methodology in practice for the first time. I
see this as crucial for the future. I am just beginning and have much to learn, but thanks to the AMS
certificate program, I have gained a stable foundation I can build going forward. The last thing I want
to highlight is creating an amazing supportive group of professionals, people, and friends who believe
in the exact cause as I do. I believe that we will maintain this supportive network in the future. Of
course, the past 2 years have not been easy, sometimes almost unbearable, but the perseverance is
now paying off manifold. The project itself highlighted a significant problem within the hospital, and
what is presented here is just the beginning of the process. I am already encountering many other

56
barriers that prevented me from achieving the goals, and I plan to continue working on this project.
Despite not achieving all goals, I feel proud that even a few interventions realistically reduced the
mortality of patients with SAB at the hospital where I work. Although there was no statistically
significant reduction in mortality, I perceive that the decrease was significant in the real world for each
patient who did survive.

Acknowledgment

I want to extend my deepest gratitude to all the instructors of the AMS certificate program. Without
their guidance, I wouldn't have acquired the essential knowledge for the AMS program. My thanks also
go to my peers, who were always there to offer support and motivation. Additionally, I want to express
my appreciation for everyone around me who believed in me. Specifically, I want to thank my tutor,
who was with me through the project. I'm also thankful for those who made my journey through this
program challenging, as they perhaps inadvertently pushed me to improve in many aspects and helped
me overcome obstacles.

Attachments – Graphs and Tables

Graph 1

Retrospective data (05/22-10/22)

BUNDLE 50%
56%
Q6 3%
Quality indicator

4%
Q5 75%
78%
Q4 75%
75%
Q3 42%
50%
Q2 52%
63%
Q1 55%
68%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90%
Q1 Q2 Q3 Q4 Q5 Q6 BUNDLE
NON INF 55% 52% 42% 75% 75% 3% 50%
Hospital 68% 63% 50% 75% 78% 4% 56%

Adherence to guidelines (%)

Table 1 – Evaluation data after 1st intervention (3 months) – all SAB patients.
Before After
Process and outcome QI Intervention Intervention p value
Q1 68 % 59 % 0,4695
Q2 63 % 63 % >0,9999
Q3 50 % 63 % 0,3483
Q4 75 % 100 % >0,9999
Q5 75 % 78 % >0,9999
Q6 4% 48 % <0,0001

57
 Bundle 56 % 69 % 0,0375
30day mortality 41 % 41 % >0,9999

Table 2 – Evaluation data after 1st intervention (3 months) – Target group SAB patients.
Before After
Process and outcome QI Intervention Intervention p value
Q1 52 % 59 % >0,9999
Q2 55 % 59 % >0,9999
Q3 42 % 65 % 0,2321
Q4 75 % N N
Q5 75 % 73 % >0,9999
Q6 3% 53 % <0,0001
Bundle 50 % 62 % 0,0124
30-day mortality 51 % 47 % >0,9999

Table 3 – Evaluation data after all interventions (9 months) – all SAB patients.

Process and outcome QI Before Intervention After Intervention Goal Achieved p value
Q1 68% 66% 83% ⮽ >0,9999
Q2 63% 68% 78% ⮽ 0,6879
Q3 50% 64% 65% ⮽ 0,1761
Q4 75% 100% to remain ☑ >0,9999
Q5 75% 76% 90% ⮽ >0,9999
Q6 4% 74% 54% ☑ <0,0001
Bundle 56% 75% 76% ⮽ <0,0001
30-day mortality 41% 28% 26% ⮽ 0,2275

Table 4 - Evaluation data after all interventions (9 months) – Target group SAB patients.

Process and outcome QI Before Intervention After Intervention Goal Achieved p value
Q1 52% 61% 67% ⮽ 0,6396
Q2 55% 63% 70% ⮽ 0,3487
Q3 42% 56% 57% ⮽ 0,3497
Q4 75% 100% remain ☑ >0,9999
Q5 75% 77% 90% ⮽ >0,9999
Q6 3% 68% 53% ☑ <0,0001
Bundle 50% 71% 70% ☑ <0,0001
30-day mortality 51% 32% 36% ☑ 0,1001

58
 Table 5 – Barriers of the project.
Clarity of the
1. Guidelines Recommendati recommendatio
factors on n Lack of national or local guidelines for SAB
Recommended Accesibility of Difficult accesibility for appropriate antibiotics
clinical the clinical from 8pm till 6 am as the hospital pharmacy is
intervention intervention closed
2. Factors relating
to the individual
health care Knowledge and Existing domain Doctors out of ID clinic doesnt have
professional skills knowledge knowledge about recomendation towards SAB
Awareness and
familiarity with
the Doctors out of ID clinic doesnt know the
recomendation appropriate antibitotics for SAB
Knowledge
about own
practise relative
to the No feedback on the SAB managment
recommendatio anywhere in the hospital, no data available
n about the managment
As there are no local guidelines for any ATB
treatemnt in the hospital the doctors are not
Proffesional Self-monitoring supported or motivated any how to use ATB
behaviour and feedback optimized way
4. Professional Commucation No communication about appropriate ATB use
interactions and influence happening in the hospital
Lack of interdisciplinary communication to
recommend ATB use consistently. Doctors
complains that they get another
Team recommedation from microbiologist and from
proccesses ID.
Referral
processes
between different Lack of communication on appropriate ATB
levels o care use
5. Incentives and Information Big limitations of the information systems - no
resources system alerts for appropriate usage of ATB
Continuing
education No motivation for doctors for education in ATB
system use
Assistence for No appropriate control of ATB use which is in
clinicans the hands of microbiologist
The doctors are not controled almost anytime
6. Capacity for Mandate, which ATB to use, the antibiotic centre
organisational authority, belonging to microbiology dont play their role
change accountibility for lack of personal and time
Nobody (management, microbiology or ID)
Capable rise the topic of appropriate use of ATB and
leadership presented as crucial, abscence of ATB team
Relative strengh According to the doctors in my hospital the
of supporters hierarchy of senior and junior doctor is very
and opponents strong and for older doctors is difficult to

59
 accept any change in their practices, the
young doctors have to obey
Lack of any data to show to the prescribers,
Monitoring and no data about compsuption of ATB, about
feedback their own prescription behavior

Table 6 – Interventions created depending on the identified barriers.

Barriers Subtype Subsubtype Intervention Implemented

Clarity of the
1. Guidelines Recommendatio recommendatio New local guidelines for SAB were
factors n n officially issued on January 30, 2023 30.01.2023
2. Factors
relating to the Education: Guidelines presented at a
individual hospital-wide seminar on January 9,
health care Knowledge and Existing domain 2023 (filmed and available to everyone
professional skills knowledge on the intranet). 09.01.2023
Awareness and Unsolicited infectious disease
familiarity with consultations for SAB patients were
the implemented as a standard practice in
recomendation May 2023 01.05.2023
Local seminars on the existence and
application of these guidelines were
performed, with infectious disease May to
4. Proffesional Commucation support, in each department of the September
interactions and influence hospital from May to September 2023. 2023
Dialogue was initiated in January 2023
with the microbiology, clinical
pharmacy, and epidemiology
departments to discuss the need for
cooperation in the management of
Team SAB, not just limited to SAB
proccesses management. January 2023
A new communication channel
between microbiology and infectious
diseases was created in February 2023
5. Incentives Information to report new cases of SAB in the
and resources system hospital. February 2023
An AMS team, with the backing of
hospital management, was introduced
to the department heads and director
6. Capacity for for the first time in May 2023 and was
organisational Capable officially integrated into the hospital
change leadership structure on March, 2024 28.05.2023

60
Annex No 1 – First page of the local guidelines

61
ESCMID AMS Certificate Program - My Project Report

Name
Marco Ripa
Institution
IRCCS San Raffaele Scientific Institute
Country
Italy
Tutor
Jesús Rodríguez-Baño
Title of Project
Antimicrobial de-escalation in patients with Enterococcus faecalis monomicrobial bacteremia
Date of submission
March 11th 2024

Abstract* (max 300 words)

Background and Aim
The use of inappropriate broad-spectrum antimicrobial treatment in patients with documented etiology is a
relevant issue for Antimicrobial Stewardship (AMS) Programs. Aim of this study was to examine the impact of an
AMS intervention in the management of Enterococcus faecalis (EF) monomicrobial bacteremia, focusing on
antimicrobial prescribing practices and the implementation of de-escalation strategies.
Materials and Methods
Patients with monomicrobial ampicillin-susceptible EF bacteremia (excluding patients with reported/confirmed
penicillin allergy) admitted to IRCCS San Raffaele Hospital (Milan, Italy) between January-December 2019 (pre-
intervention) and March-September 2023 (intervention) were analyzed. A quality indicator (“[Number of
patients who were de-escalated from empirical broad-spectrum antimicrobials to ampicillin or
amoxicillin/number of patients who received broad spectrum antimicrobials] X 100”) was used to assess de-
escalation practices, and two surveys among ID physicians were conducted to identify barriers and facilitators to
de-escalation. The intervention focused on these barriers and involved the following key strategies: education
and modelling, persuasion, environmental restructuring, and enablement. Results are presented as median
(interquartile range) and absolute number (percent) as appropriate. The groups were compared using Chi-square
and Wilcoxon rank-sum tests.
Results
Participation in surveys was 58% and 63%. In the pre-intervention period, 14/35 (40.0%) patients with EF
bacteremia received definitive treatment with ampicillin/amoxicillin, while 21/35 (60.0%) were treated with
broad-spectrum antimicrobials. In the intervention period, 13/38 (34.2%) patients were treated with broad
spectrum antimicrobials, while 25/38 (65.8%) received ampicillin/amoxicillin as definitive therapy (P-value 0.027
compared to pre-intervention). No significant differences in patients’ characteristics or outcomes were observed
between patients receiving ampicillin/amoxicillin and patients receiving broad spectrum antimicrobials, and
between patients in the pre-intervention and intervention period.
Conclusions
Despite similar patient outcomes, de-escalation practices for EF monomicrobial bacteremia were inconsistent.
Targeted interventions focused on identified barriers helped increasing the proportion of patients undergoing
de-escalation in the intervention period.

Stage 1 Introduction: description of the setting and the AMS program in place and identification of the problem
I work as an ID physician in a tertiary-care hospital with about 1.200 beds. The hospital has several internal
medicine units, a hematology ward where bone-marrow transplantation is performed, a pediatrics ward with a
neonatal intensive care unit, three adult intensive care units (general/post-surgical, neurosurgical,
cardiothoracic), several surgical wards, and a 24-bed ID department. ID consultants are available on site
weekdays from 9am to 5pm, and ca be reached on-call 24/7 for remote consultation. There is no official
antimicrobial stewardship program. Nevertheless, ID physicians routinely evaluate all patients with positive
blood cultures and perform consultations at the request of treating physicians from different specialties. ID
physicians are aware of all positive blood cultures in the hospital and proactively evaluate all bacteremic patients.
Most patients with positive blood cultures are evaluated at least once (at the time of blood cultures positivity),

62
while follow-up for streamlining is not performed consistently. One of the most commonly prescribed
antimicrobials for empiric and definitive treatment is piperacillin-tazobactam, which usually considered an
adequate definitive treatment even for patients with monomicrobial Enterococcus faecalis bacteremia.
Nevertheless, no data were available to describe the magnitude of the problem.

Stage 2 Measurement
First, I evaluated the antimicrobial treatment received by patients with Enterococcus faecalis bacteremia. I
calculated the number of patients with monomicrobial ampicillin-susceptible Enterococcus faecalis bacteremia
who received broad-spectrum antimicrobial treatment (i.e., piperacillin-tazobactam) and were not de-escalated
to ampicillin/amoxicillin-based treatment. The time frame selected was January-December 2019 (to exclude
potential confounders related to the COVID-19 pandemic, which compromised the routine activity of ID
consultants in the hospital). The list of patients with Enterococcus faecalis bacteremia was provided by the
Microbiology Department. Clinical, microbiological and treatment data were extracted from the electronic
clinical charts. Patients with reported/confirmed penicillin allergy were excluded from the analysis. As a quality
indicator, I used “(Number of patients who were de-escalated from empirical broad spectrum antimicrobials to
ampicillin or amoxicillin/number of patients who received broad spectrum antimicrobials) X 100”.
Between January and December 2019, 122 patients with Enterococcus faecalis bacteremia were identified.
Among them, 41 (33.9%) had monomicrobial bacteremia. Six patients had reported/confirmed penicillin allergy.
Thirty-five patients were therefore included in the analysis. Among 35 patients, 14 (40.0%) received definitive
treatment with ampicillin/amoxicillin, while 21 (60.0%) were treated with broad-spectrum antimicrobials, mainly
piperacillin-tazobactam (81.0%). Patients’ characteristics are shown in table 1. No significant differences were
found in terms of age, comorbidities, ward of admission, source of infection, and presentation with septic shock.
Treatment duration was significantly longer in patients treated with ampicillin/amoxicillin (10 [8-12] vs. 13 [10-
43], P-value 0.044), but this difference was not evident when excluding patients with infective endocarditis (10
[8-12] vs. 10 [9-17], P-value 0.466). Clinical outcomes (clinical cure at 7 days and at end-of-treatment, intra-
hospital mortality, 6-month recurrence, subsequent infection, C. difficile infection) were similar between groups.
Taken together, these data show that among patients with monomicrobial Enterococcus faecalis bacteremia, de-
escalation to a narrow-spectrum antimicrobial therapy such as ampicillin/amoxicillin was not consistently
performed. No significant differences were found between patients who underwent de-escalation and patients
who maintained broad-spectrum treatment, showing that this choice was probably not entirely influenced by
patients’ characteristics and clinical presentation. Nevertheless, patients’ outcomes were not significantly
different between the two groups.

Table 1 – Characteristics of patients in the pre-intervention period (Jan-Dec 2019)

Characteristics No ampicillin/amoxicillin use Ampicillin/amoxicillin use P-value
(N=21) (N=14)
Age, years 67 (62-77) 72 (56-79) 0.999
Charlson comorbidity index 5 (3-8) 6 (3-8) 0.907
Ward 0.166
Cardiac surgery 0 (0.0%) 2 (14.3%)
Cardiology 3 (14.3%) 1 (7.1%)
ER 1 (4.8%) 3 (21.4%)
Hematology 2 (9.5%) 0 (0.0%)
Medicine 3 (14.3%) 2 (14.3%)
Nephrology 0 (0.0%) 2 (14.3%)
Neurology 1 (4.8%) 1 (7.1%)
Abdominal surgery 1 (4.8%) 0 (0.0%)
Rehab 3 (14.3%) 0 (0.0%)
Urology 7 (33.3%) 3 (21.4%)
Source 0.107
Abdominal 3 (14.3%) 0 (0.0%)
Vascular catheter 3 (14.3%) 2 (14.3%)
Infective endocarditis 0 (0.0%) 4 (28.6%)
Soft tissue 1 (4.8%) 1 (7.1%)
Urinary 12 (57.1%) 5 (35.7%)
Unknown 2 (9.5%) 2 (14.3%)

63
 Source control performed 8 (38.1%) 7 (50.0%) 0.486
Septic shock 1 (4.8%) 1 (7.1%) 0.766
Positive follow-up blood cultures 3/15 (20.0%) 2/9 (22.2%) 0.897
Empirical treatment duration, days 3 (1-5) 5 (2-7) 0.570
Definitive treatment duration, days 10 (7-11) 12 (7-40) 0.127
Total treatment duration, days 10 (8-12) 13 (10-43) 0.044
Definitive treatment <0.001
Ampicillin/amoxicillin 2 (14.3%)
Ampicillin+ceftriaxone 7 (50.0%)
Amoxicillin-clavulanate 5 (35.7%)
Piperacillin-tazobactam 17 (81.0%)
Vancomycin 3 (14.3%)
Linezolid 1 (4.8%)
Time to ampicillin/amoxicillin, days - 0 (0-3)
Clinical cure at 7 days 18 (85.7%) 9 (64.3%) 0.139
Clinical cure at end-of-treatment 18 (85.7%) 14 (100.0%) 0.139
Intra-hospital mortality 3 (14.3%) 0 (0.0%) 0.139
6-month recurrence 2 (9.5%) 1 (7.1%) 0.805
Subsequent infection during 3 (14.3%) 1 (7.1%) 0.515
hospitalization
C. difficile infection 1 (4.8%) 1 (7.1%) 0.766

Stage 3 Barriers and facilitators

In order to better assess potential barriers and facilitators, I conducted a survey among ID colleagues in which
several clinical scenarios were presented, giving the physicians the opportunity to choose among different
treatments, and a final part regarding the knowledge of the current practice in the hospital. The results of the
survey were discussed during a dedicated meeting (see below).
The potential barriers and facilitators were identified using the Flottorp framework as a reference.
Major barriers:
- Guideline factors
- Absence of specific guidelines regarding Enterococcus faecalis bacteremia
- Guidance for management derived primarily from available literature, expert opinion, and individual
experience
- Factors relating to the individual healthcare professional
- Lack of knowledge regarding the usual antimicrobial management of Enterococcus faecalis bacteremia
in my hospital
- Lack of intention and motivation to change pre-existing behaviour/custom
- Professional interactions
- Lack of communication within the ID team to engage in a specific behaviour (e.g. de-escalation)
Possible facilitators:
- Factors relating to the individual healthcare professional
- Presence of appropriate skills to manage patients with Enterococcus faecalis bacteremia
- All patients with bacteremia are managed by experienced ID physicians
- Intention and motivation to reduce use of broad-spectrum antimicrobials
- Professional interactions
- Good collaboration with attending physicians in different wards
- Incentive and resources
- Availability of necessary resources
- Informatic tools to identify pts with Enterococcus faecalis bacteremia
- Financial incentives
- Reduction of broad-spectrum antimicrobials use
- Capacity for organisational change
- Mandate from hospital administration to evaluate all patients with bloodstream infections

Stage 4 Development and execution of the stewardship intervention

In order to address the identified barriers, I proceeded as follows:

64
- Guideline factors: In February 2023, I reviewed and presented the available literature regarding the
management of Enterococcus faecalis bacteremia, with a focus on the possibility of using narrow-spectrum
antimicrobials when polymicrobial infection was not documented or suspected (education, modelling).
- Factors relating to the individual healthcare professional: In February 2023, I presented the data collected
regarding the management and outcomes of Enterococcus faecalis monomicrobial bacteremia in my hospital in
2019 (education); between March and September 2023, I stimulated colleagues to consider de-escalation
whenever feasible, in light of the positive impact of streamlining and the available literature suggesting favorable
outcomes with the use of narrow-spectrum therapy (persuasion), also in order to change the social context by
making de-escalation more frequent to stimulate the adoption of de-escalation by other colleagues and change
current habits (environmental restructuring).
- Professional interactions: between March and September 2023, I focused on updating an electronic database
of patients with bacteremia, to be used as a clinical tool to know how patients with bloodstream infections were
being treated, in order to evaluate patient deemed eligible for de-escalation (enablement).
Data was analyzed in January-February 2024.

Stage 5 Evaluation of the stewardship intervention

After the original survey, I sent out a similar one in April 2023 to re-evaluate and emphasize the messages from
the previous meeting. In the original survey, 58% of the colleagues participated, while 63% of colleagues
participated in the second survey. All ID specialists participated to the educational session which provided the
data obtained from the survey and highlighted the data regarding the management of Enterococcus faecalis
bacteremia in our Institution, along with the review of the available literature.
To assess the impact of the intervention, data regarding patients with Enterococcus faecalis bacteremia between
March and September 2023 were analyzed. Among 72 patients with Enterococcus faecalis bacteremia, 38
(52.8%) had monomicrobial bacteremia. No patients had reported/confirmed allergy. Among 38 patients, 13
(34.2%) were treated with broad spectrum antimicrobials, while 25 (65.8%) received ampicillin/amoxicillin.
Patients’ characteristics are described in table 2. Again, no significant differences were found in terms of age,
comorbidities, ward of admission, source of infection, and presentation with septic shock. Treatment duration
was significantly longer in patients treated with ampicillin/amoxicillin (10 [7-11] vs. 14 [10-18], P-value 0.001),
even after excluding patients with infective endocarditis. Clinical outcomes (clinical cure at 7 days and at end-
of-treatment, intra-hospital mortality, 6-month recurrence, subsequent infection, C. difficile infection) were
similar between groups.
The proportion of patients who received ampicillin/amoxicillin for monomicrobial Enterococcus faecalis
bacteremia was significantly higher in the intervention period compared to the pre-intervention phase (40.0%
vs. 65.8%, P-value 0.27). Patients’ characteristics were similar between the two periods (table 4), except for the
proportion of patients who underwent source control, which was higher in the pre-intervention period (42.9%
vs. 21.1%, P-value 0.045). Outcomes were also not significantly different between groups.

Table 2 – Characteristics of patients in the intervention period (Mar-Sep 2023)

Characteristics No ampicillin/amoxicillin use Ampicillin/amoxicillin use P-value
(N=13) (N=25)
Age, years 66 (56-76) 77 (56-84) 0.141
Charlson comorbidity index 5 (3-7) 5 (3-7) 0.785
Ward 0.369
Cardiac surgery 1 (7.7%) 0 (0.0%)
Cardiology 0 (0.0%) 1 (4.0%)
ER 4 (30.8%) 7 (28.0%)
Hematology 1 (7.7%) 0 (0.0%)
Infectious diseases 0 (0.0%) 1 (4.0%)
Medicine 0 (0.0%) 3 (12.0%)
Ob-Gyn 1 (7.7%) 0 (0.0%)
Abdominal surgery 2 (15.4%) 2 (8.0%)
Rehab 1 (7.7%) 1 (4.0%)
Urology 3 (23.1%) 10 (40.0%)
Source 0.135
Abdominal 5 (38.5%) 3 (12.0%)
Vascular catheter 1 (7.7%) 0 (0.0%)

65
 Infective endocarditis 0 (0.0%) 3 (12.0%)
Urinary 4 (30.8%) 12 (48.0%)
Unknown 3 (23.1%) 7 (28.0%)
Source control performed 2 (15.4%) 6 (24.0%) 0.537
Septic shock 0 (0.0%) 1 (4.0%) 0.465
Positive follow-up blood cultures 0/12 (0.0%) 5/23 (21.7%) 0.081
Empirical treatment duration, days 1 (0-1) 4 (1-6) 0.118
Definitive treatment duration, days 10 (7-11) 10 (6-18) 0.314
Total treatment duration, days 10 (7-11) 14 (10-18) 0.001
Definitive treatment <0.001
Ampicillin/amoxicillin 12 (48.0%)
Ampicillin+ceftriaxone 7 (28.0%)
Amoxicillin-clavulanate 4 (16.0%)
Piperacillin-tazobactam 12 (92.3%)
Imipenem 1 (7.7%)
Time to ampicillin/amoxicillin, days 0 (1-5)
Clinical cure at 7 days 9 (69.2%) 19 (76.0%) 0.653
Clinical cure at end-of-treatment 10 (76.9%) 20 (80.0%) 0.825
Intra-hospital mortality 1 (7.7%) 2 (8.0%) 0.973
6-month recurrence 0 (0.0%) 1 (4.0%) 0.465
Subsequent infection during 3 (23.1%) 6 (24.0%) 0.949
hospitalization
C. difficile infection 0 (0.0%) 0 (0.0%) 0.999

Table 3 – Characteristics of patients in pre-intervention and intervention period

Characteristics Pre-intervention Intervention P-value
(N=35) (N=38)
Age, years 70 (59-78) 75 (56-82) 0.600
Charlson comorbidity index 6 (3-8) 5 (3-7) 0.276
Ward 0.205
Cardiac surgery 2 (5.7%) 1 (2.6%)
Cardiology 4 (11.4%) 1 (2.6%)
ER 4 (11.4%) 11 (28.9%)
Hematology 2 (5.7%) 1 (2.6%)
Infectious Disesase 0 (0.0%) 1 (2.6%)
Medicine 5 (14.3%) 3 (7.9%)
Nephrology 2 (5.7%) 0 (0.0%)
Neurology 2 (5.7%) 0 (0.0%)
Ob-Gyn 0 (0.0%) 1 (2.6%)
Abdominal surgery 1 (2.9%) 4 (10.5%)
Rehab 3 (8.6%) 2 (5.3%)
Urology 10 (28.6%) 13 (34.2%)
Source 0.088
Abdominal 3 (8.6%) 8 (21.1%)
Vascular catheter 5 (14.3%) 1 (2.6%)
Infective endocarditis 4 (11.4%) 3 (7.9%)
Soft tissue 2 (5.7%) 0 (0.0%)
Urinary 17 (48.6%) 16 (42.1%)
Unknown 4 (11.4%) 10 (26.3%)
Source control performed 15 (42.9%) 8 (21.1%) 0.045
Septic shock 2 (5.7%) 1 (2.6%) 0.507
Positive follow-up blood cultures 5/24 (20.8%) 5/35 (14.3%) 0.510
Empirical treatment duration, days 4 (1-6) 3 (1-5) 0.687
Definitive treatment duration, days 10 (7-13) 10 (6-13) 0.466
Total treatment duration, days 10 (9-14) 11 (9-17) 0.751
Definitive treatment 0.053

66
 Ampicillin/amoxicillin 2 (5.7%) 12 (31.6%)
Ampicillin+ceftriaxone 7 (20.0%) 7 (18.4%)
Amoxicillin-clavulanate 5 (14.3%) 4 (10.5%)
Piperacillin-tazobactam 17 (48.6%) 14 (36.8%)
Imipenem 0 (0.0%) 1 (2.6%)
Vancomycin 3 (8.6%) 0 (0.0%)
Linezolid 1 (2.9%) 0 (0.0%)
Time to ampicillin/amoxicillin, days 0 (0-3) 1 (0-5) 0.361
Clinical cure at 7 days 27 (77.1%) 28 (73.7%) 0.732
Clinical cure at end-of-treatment 32 (91.4%) 30 (78.9%) 0.136
Intra-hospital mortality 3 (8.6%) 3 (7.9%) 0.916
6-month recurrence 3 (8.6%) 1 (2.6%) 0.265
Subsequent infection during 4 (11.4%) 9 (23.7%) 0.172
hospitalization
C. difficile infection 2 (5.7%) 0 (0.0%) 0.135

Conclusions
My project addressed a very specific problem (Enterococcus faecalis monomicrobial bacteremia), and was
distinctive in that it primarily involved ID colleagues rather than physicians from other specialty. However,
although I would have expected high engagement in antimicrobial stewardship programs especially among ID
physicians, participation to the survey was low, with about 40% of colleagues not participating in either survey.
Furthermore, while the intervention appears to have actually increased the number of patients who de-
escalated, there was still 35% of patients who could potentially have been shifted to narrow-spectrum
antimicrobials who did not de-escalate. As there were no differences between patients who de-escalated and
those who did not, it is possible to argue that other unmeasured factors (including physicians’ attitude and
habits), may have played a crucial role in this choice. Given that knowledge regarding the topic was adequate
among ID colleagues, other interventions could be expected to have an impact on antimicrobial prescribing,
especially environmental restructuring (i.e., changing the context by making de-escalation more frequent).
Indeed, after an initial period during which persuasion (personal stimulation to other colleagues to consider de-
escalation) may have played an important role, subsequently the practice of de-escalation became more
consolidated even without external stimuli.
Before participating in the AMS Certificate Program, I was naively convinced that the most difficult task of an
antimicrobial stewardship program would be to educate and update physicians on current appropriate clinical
practices, as if the main barrier to appropriate antimicrobial use could be attributed mostly to lack of knowledge.
Nevertheless, trough the AMS Certificate Program and during my project, I realized that the problem is very
complex and multifaceted, and that addressing issues regarding the appropriate use of antimicrobials requires a
multidisciplinary approach. At the beginning of the AMS Certificate Program, my hospital did not have a
structured antimicrobial stewardship program, but I am know trying to use the knowledge acquired during the
Program to help create one. I was involved in several meetings with Hospital Administration, Microbiology, and
Pharmacy, and we are on the way to institutionalizing a group specifically dedicated to Antimicrobial
Stewardship. Initially, we will start to focus on specific problems, with the hope of extending our activities in the
future (also by obtaining more dedicated personnel). As an example, since the end of 2023, we transitioned from
paper to electronic prescribing, and I am now focusing on creating a real-time evaluation of patients receiving
inappropriate broad-spectrum antibiotics (in particular, carbapenems), to assess the extent of the problem and
try to design appropriate interventions. In addition, we conducted a retrospective evaluation of the
appropriateness of pre-operative antimicrobial prophylaxis, and we are analyzing the results to identify areas of
improvement. Several other projects are underway, and I hope the experience acquired during the course will
be helpful in further developing an AMS program at my hospital.
Overall, I feel that the AMS Certificate Program gave me the theoretical tools to help implement an Antimicrobial
Stewardship Program in my hospital, and I believe that my project, although small, has helped me to understand
in practice what difficulties might be encountered during such process. I would also like to emphasize the
importance of the interactions with the other members of the AMS Certificate Program, which were really helpful
in understanding the issues in very different contexts, and stimulated very productive and insightful discussions
regarding the implementation of AMS programs.

67
Title: Appropriate de-escalation in carbapenem susceptible E. coli and K. pneumoniae bloodstream
infections
Authors: Matteo Piccica1, Letizia Attala1, Dario Mannini1, Dario Nigi2, Marco Fognani2, Elisabetta Innocenti3,
Jesús Rodríguez-Baño4, Massimo Antonio Di Pietro1
Affiliation
1 Unit of Infectious Diseases, Department of Medical Specialties, Azienda USL Toscana Centro (AUTC), Bagno a Ripoli,
Florence, Italy
2 Department of Experimental and Clinical Medicine, University of Florence - Florence (Italy)
3 Department of Pharmacy, Azienda USL Toscana Centro (AUTC), Bagno a Ripoli, Florence, Italy
4 Unidad Clínica de Enfermedades Infecciosas y Microbiología, Instituto de Biomedicina de Sevilla (IBiS), Hospital
Universitario Virgen Macarena, Departamento de Medicina, Universidad de Sevilla, CSIC, Seville, Spain;

Abstract
Background
Infections caused by multi-drug resistant pathogens pose a significant threat, prompting broad-spectrum
antibiotic therapy. Timely transition to targeted antibiotics based on antibiograms is vital, yet often neglected,
especially in smaller hospitals lacking antimicrobial stewardship programs.
Methods
Initially, a retrospective audit examined electronic charts of K. pneumoniae and E. coli bacteremia cases from May
2022 to October 2022, focusing on the proportion of de-escalated patients. Then, an interventional phase ran
from April 2023 to September 2023, involving daily review of positive blood cultures. For each patient with
bacteremia, proactive bedside consultations discussed de-escalation opportunities. A plenary event highlighted
the project's importance three months later. After project completion, the de-escalation ratio was recalculated
using DOOR MAT scores. Bacteria were categorized into five susceptibility classes, with antibiotic desirability
ranked into six levels.
Results
Cohorts included 42 patients in the observational phase (Group A) and 43 in the intervention phase (Group B),
showing comparability in pre-existing conditions and illness severity (Table 1). In-hospital mortality rates
remained consistent between groups. In Group A, 62% were eligible for de-escalation, with 52% successfully de-
escalated, while in Group B, 81% were eligible, with 95% effectively de-escalated. Group B exhibited a
significantly higher overall de-escalation rate (p<0.001) and optimal de-escalation rate (p<0.001). DOOR-MAT
analysis provided a more detailed insight into the improvement. The S-S-S-S-S and R-S-S-S-S strains received
ideal treatment in 13% and 0% of cases, respectively, during the observational phase, compared to 67% and 81%
of cases, respectively, during the interventional phase (Figure 1).
Conclusions
The structured AMS project consistently improved antibiotic appropriateness and likely played a role in raising
awareness about the importance of prompt targeted therapy. DOOR MAT serves as a valuable tool for
evaluating the intricacies of de-escalation and can be an effective metric in similar interventions.

68
Introduction
Infections caused by multi-drug resistant pathogens are widespread, progressively resulting in a significant
number of deaths1. Consequently, physicians treating patients with serious infections often fear the risk of
resistance and initiate broad-spectrum antibiotic therapy as a precautionary measure2.
One of the most crucial measures to mitigate the emergence of those bacteria is promptly transitioning from
broad-spectrum antibiotic empirical therapy to a more targeted approach based on antibiograms3. However,
proactive de-escalation is not consistently pursued, particularly in small peripheral hospitals lacking a structured
antimicrobial stewardship program (AMS).
Bloodstream infections caused by Escherichia coli and Klebsiella pneumoniae provide an ideal scenario to study the
appropriateness and rate of de-escalation. In Italy empirical therapy for Gram-negative bacteremia typically
involves β-lactam antibiotics with activity against extended spectrum β-lactamase producers (ESBL) and multi-
drug resistant (MDR) P. aeruginosa4.
To address this issue, we conducted an experimental AMS intervention aimed at improving the de-escalation rate
in the most common Enterobacterales bacteremia.
Methods
Study population
The study comprised two phases: observational and interventional. During the observational phase, we
retrospectively collected clinical data of all patients admitted to Santa Maria Annunziata Hospital (OSMA),
Florence, Italy, with cases of K. pneumoniae and E. coli bacteremia from May 2022 to October 2022 (Group A). In
the interventional phase, we retrospectively collected clinical data, after implementing a proactive de-escalation
program, for all patients admitted from April 2023 to September 2023 with the same bacteremia (Group B).
Eligibility criteria for enrollment were: (i) age ≥18 yeas; (ii) positive blood culture for carbapenem susceptible K.
pneumoniae or E. coli.
Exclusion criteria were: (i) death in the first 48 hours (ii) presence of carbapenem resistance.
Data collection and variables
Demographic, clinical, laboratory, microbiological, treatment and outcome information were retrospectively
collected by revising electronical medical records. The data were recorded in a secure electronic sheet. Baseline
patient conditions included the most important comorbidities, allergies, and the Charlson Comorbidity Index
(CCI)5. Disease severity at the time of blood culture collection was assessed using the APACHE-II score6.
Depending on the source of bacteremia we differentiated between urinary tract infections (UTI), low respiratory
tract infections (LRTI), intrabdominal infections (IAI) or primary bloodstream infections. The presence of
coinfection was recorded, specifying the bacterium species responsible for it. The date of empiric antibiotic
therapy and any subsequent switch to targeted therapy were recorded along with the exact date of availability of
antibiograms.
Intervention and outcome
In the observational phase, we calculated the proportion of de-escalation and the time gap between the
availability of antibiograms and de-escalation. The types of de-escalation addressed were: (i) withdrawal of Gram-
positive therapy; (ii) switching from combination to monotherapy; (iii) transitioning to narrower spectrum
therapy. We defined "optimal de-escalation" as the best possible therapy based on antibiograms,
pharmacokinetic/pharmacodynamic (PK/PD) parameters, and allergies. We defined "days without de-
escalation" as the number of days from when the susceptibility test was available until a lower spectrum drug was
administered or until the end of therapy (if de-escalation never occurred). Inappropriate escalation was defined as
the scenario in which, despite antibiograms indicating the possibility of transitioning to a narrower antibiotic,
there was a decision to escalate to a broader spectrum without any clear rationale, suggesting a misunderstanding
of the antibiogram.

69
Subsequently, an interventional phase ran from April 2023 to September 2023, entailing daily reviews of positive
blood cultures. For each patient with a positive blood culture for K. pneumoniae or E. coli, proactive bedside
consultations were conducted. During these consultations, the infectious disease (ID) specialists involved in the
project discussed de-escalation opportunities with colleagues on the receiving ward. Furthermore, the
importance and significance of de-escalation were emphasized.
Three months later, a plenary event highlighted the importance of the project and facilitated the sharing of
preliminary data, eliciting positive feedback.
After the completion of the project, the de-escalation ratio, the number of "optimal de-escalations," and the
"days without de-escalation" were recalculated.
To better understand how the structured intervention of our AMS program has changed the approach to de-
escalation, we have applied the DOOR-MAT methodology7. Bacteria were classified into five susceptibility
classes based on their response to increasingly broader antibiotics. Consequently, we identified five patterns of
increased antibiotic resistance: S-S-S-S-S, R-S-S-S-S, R-R-S-S-S, R-R-R-S-S, R-R-R-R-S. From the combination
of administered therapy and susceptibility pattern, we constructed the DOOR MAT grid, delineating six ranked
categories (as displayed in Figure 1): ideal treatment (most desirable), slight overtreatment, moderate
overtreatment, heavy overtreatment, severe overtreatment, undertreatment (less desiderable). Ideal treatment was
characterized by targeting therapy matching the narrowest available antibiotic, with each subsequent step
representing overtreatment (e.g., administering ceftriaxone to a patient with amoxi-clav susceptibility was
considered slight overtreatment), while undertreatment involved employing non-active molecules in vitro.
Numerical scores were assigned to each category and analyzed as continuous variables. In cases where patients
were treated with different antibiotic categories over several days, an average value was computed to determine
the final score.
A pharmacoeconomic analysis was conducted with the assistance of hospital pharmacists. The actual expenses
were compared to the optimal expenses, defined as the cost of therapy if every patient had been de-escalated to
the narrowest option upon availability of antibiograms.
Statistical analysis
Continuous variables were expressed as medians and interquartile ranges (IQR); categorical variables were
expressed as percentages of the group to which they belong. The Mann-Whitney U test was used to compare
non-normally distributed continuous variables; The Kruskal-Wallis test was used in case of comparison of ≥3
groups. Categorical variables were evaluated by the two-tailed Fisher’s exact test. The multivariate Cox model
was performed on the weighted population to compare the outcome between the observational and
interventional groups and the hazard ratio (HR) with 95%CI was reported.
Results
During the observational phase, there were 58 positive blood cultures for E. coli and K. pneumoniae, but only 42
were included in the analysis (29 for E. coli and 13 for K. pneumoniae). In the interventional phase, there were 63
positive blood cultures for E. coli and K. pneumoniae, yet only 43 were included in the analysis (30 for E. coli and 13
for K. pneumoniae).
The median age of the population was 76 (IQR 67-84), with 48 being male (56.5%). Concerning pre-existing
comorbidities, the median CCI score was 5 (IQR 4-7), and the median APACHE-II score was 12 (IQR 9-17).
The in-hospital all-cause mortality rate was 14.1%, and the mean hospital stay was 13 days (IQR 8-17). UTI
accounted for 53 cases (62.4%), IAI for 14 cases (16.5%), LRTI for 7 cases (8.2%), and primary bloodstream
infection for 11 cases (12.9%). Population of observational and interventional phase were fully comparable,
complete demographic and clinical information is summarized in Table 1.
In 53 patients (62.3%), empiric therapy consisted of piperacillin/tazobactam, with vancomycin added in 10 cases
(11.8%). During the observational phase, 26 patients (61.9%) could be de-escalated, while in the interventional
phase, this number rose to 35 (81.4%). The percentage of de-escalations carried out among those deemed
possible was 52% during the observational phase and 94% during the interventional phase (p<0.001). The

70
proportion of optimal de-escalations increased from 22% during the observational phase to 94% during the
interventional phase (p<0.001). The percentage of cases receiving de-escalation on the same day as antibiogram
availability rose from 32% during the observational phase to 56% during the interventional phase (p<0.05).
Similarly, the proportion of cases receiving optimal de-escalation on the same day as antibiogram availability
increased from 0% during the observational phase to 56% during the interventional phase (p<0.001). The rate of
inappropriate escalation decreased from 26% to 5% (p<0.05).
The proactive consultation had its greatest impact in optimizing the transition from broad to narrow molecules.
Conversely, the rate of discontinuing Gram+ coverage and shifting from combination to monotherapy was
already satisfactory during the observational phase, as depicted in Figure 2.
DOOR MAT was applied to both the observational and interventional groups. Neither undertreatment nor
severe overtreatment occurred in either patient group. In group A, there were 6 cases (14.3%) of heavy
overtreatment, compared to none in group B. Moderate overtreatment decreased from 11 cases (26.2%) to 1
(2.3%), and slight overtreatment decreased from 17 cases (40.5%) to 9 (20.1%). Ideal treatment increased from 8
cases (19%) in group A to 33 (76.7%) in group B. The complete DOOR MAT is displayed in Figure 3.
The total cumulative days from the availability of susceptibility test results until a de-escalation was carried out or
until the end of therapy (if de-escalation never occurred) were 148 in the observational phase and 24 in the
interventional phase (P<0.001). Similarly, the total cumulative days from the availability of susceptibility test
results until optimal de-escalation was executed or until the end of therapy (if de-escalation never occurred) were
222 in the observational phase and 34 in the interventional phase (p<0.001).
During the observational phase, the effective expenses were 38% higher than the potential cost of optimal
therapy. In the interventional phase, the effective expenses were only 3.5% higher than the potential cost of
optimal therapy.
In the multivariate Cox model (Table 2), a higher rate of ideal therapy wasn't associated with increased mortality;
the hazard ratio (HR) ranged from 3.75 (95% CI 0.1-127.4, p=0.5) for less desirable therapy to HR 2.7 (95% CI
0.2-44.1, P=0.5) for optimal therapy. Conversely, although this trend wasn't statistically significant, it suggested a
potential reduction in mortality.
Moreover, higher rates of ideal therapy seemed to correlate with a trend towards shorter hospital stays, although
this association didn't reach statistical significance (Table 3). In fact, the HR for in-hospital stays longer than 14
days ranged from 4.8 (95% CI 0.9-25.6, p=0.06) for less desirable therapy to HR 1.4 (95% CI 0.4-5.1, p=0.6) for
ideal therapy (Table 3).
Discussion
We examined the appropriateness of target antibiotic therapy both before and after implementing a structured
AMS program. Without proactive consultation, the degree of de-escalation was found to be suboptimal. The
prevalence of overtreatment and the lack of de-escalation in a considerable number of cases underscore the
necessity to enhance the quality of care for patients with bloodstream infections caused by E. coli and K.
pneumoniae. Discussing this behavior with colleagues appears to be linked to a lack of awareness regarding the
importance of promptly adjusting antibiotic regimens. Colleagues appear to maintain the perspective that once a
patient begins treatment with a broad-spectrum antibiotic, switching to a narrower spectrum antibiotic, especially
if it requires three to four days to complete the antibiotic course, may be considered ineffective. Indeed, when
examining the "days without de-escalation," it becomes evident how easily a significant number of days of
inappropriately broad antibiotic therapy can accumulate among a few patients. By emphasizing this concept, we
believe it is possible to raise awareness among colleagues regarding the importance of addressing unnecessary
ecological pressure on resistance development, adverse drug reactions, and costs, while simultaneously
maintaining treatment efficacy8. Furthermore, prompt de-escalation could facilitate earlier patient discharge or
transitioning to a full oral therapy.
The lower impact on improving "in-time de-escalation" could be attributed to the limited human resources fully
dedicated to this task. The chronic shortage of personnel assigned to AMS is a well-known cause of its inefficacy.
In the literature the the staff to be allocated to an AMS program ranges from 3 to 6 FTE (full-time equivalents)

71
per 1000 hospital beds9,10. In our experience, this project was pursued every day, in addition to routine activities,
solely based on personal self-motivation. We believe that a mandate from stakeholders, committing to dedicate
specific work hours to the AMS project, is crucial. Considering also the fact that AMS projects usually lead to a
reduction in direct and indirect costs related to antibiotic therapy, investing in human resources capable of
managing these activities is definitely worthwhile.
The satisfactory performance in discontinuing Gram+ coverage and transitioning from combination to
monotherapy during the observational phase could be attributed to a pre-existing greater awareness of the
differences between Gram+ and Gram- target therapy. Conversely, the lower rate of switching from broad to
narrow molecules underscores a lack of attention to reducing the antibiotic spectrum according to the
antibiogram.
The DOOR MAT appears to be a useful tool for assessing the appropriateness of antibiotic treatment based on
bacterial susceptibility. Furthermore, it provides a quantitative and qualitative measure of the improvement
attained through an intervention. Our analysis revealed a more favorable definitive antibiotic treatment due to
increased and expedited de-escalation. Integrating DOOR MAT results with outcome measures enables us to
demonstrate the concurrent maintenance of efficacy in broader spectrum therapies without raising mortality rates
or prolonging hospitalization durations.
Our study has several limitations. The cohort size is small, thereby limiting the statistical analysis's power.
Additionally, the population exhibits a high average CCI and APACHE-II score, potentially affecting confidence
in de-escalation without the consultation of an infectious disease specialist. Furthermore, there is no precise
definition of de-escalation available in the literature.11
Conclusion
In our experience, the rate and speed of de-escalation in a cohort of patients with bloodstream infections caused
by E. coli or K. pneumoniae consistently increase with a proactive AMS program, without adversely affecting key
outcomes such as mortality and length of hospital stay. DOOR MAT serves as an attractive metric for assessing
the effectiveness of such interventions as it integrates the appropriateness of antibiotic spectrum and efficacy.
Future perspectives may involve repeating the analysis immediately after the proactive intervention to assess
whether colleague awareness continues to increase.

References
1. Antimicrobial Resistance Collaborators. Global burden of bacterial antimicrobial resistance in 2019: a
systematic analysis. Lancet. 2022;399(10325):629-655. doi:10.1016/S0140-6736(21)02724-0
2. Evans L, Rhodes A, Alhazzani W, et al. Surviving Sepsis Campaign: International Guidelines for
Management of Sepsis and Septic Shock 2021. Crit Care Med. 2021;49(11):e1063-e1143.
doi:10.1097/CCM.0000000000005337
3. Karam G, Chastre J, Wilcox MH, Vincent JL. Antibiotic strategies in the era of multidrug resistance. Crit
Care. 2016;20(1):136. doi:10.1186/s13054-016-1320-7
4. Harris PNA, Tambyah PA, Lye DC, et al. Effect of Piperacillin-Tazobactam vs Meropenem on 30-Day
Mortality for Patients With E coli or Klebsiella pneumoniae Bloodstream Infection and Ceftriaxone
Resistance. JAMA. 2018;320(10):984. doi:10.1001/jama.2018.12163
5. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity
in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373-383.
6. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification
system. Crit Care Med. 1985;13(10):818-829.

72
7. Wilson BM, Jiang Y, Jump RLP, et al. Desirability of Outcome Ranking for the Management of
Antimicrobial Therapy (DOOR MAT): A Framework for Assessing Antibiotic Selection Strategies in the
Presence of Drug Resistance. Clin Infect Dis. 2021;73(2):344-350. doi:10.1093/cid/ciaa1769
8. Paul M, Dickstein Y, Raz-Pasteur A. Antibiotic de-escalation for bloodstream infections and pneumonia:
systematic review and meta-analysis. Clin Microbiol Infect. 2016;22(12):960-967.
doi:10.1016/j.cmi.2016.05.023
9. ten Oever J, Harmsen M, Schouten J, et al. Human resources required for antimicrobial stewardship
teams: a Dutch consensus report. Clinical Microbiology and Infection. 2018;24(12):1273-1279.
doi:10.1016/j.cmi.2018.07.005
10. Doernberg SB, Abbo LM, Burdette SD, et al. Essential Resources and Strategies for Antibiotic
Stewardship Programs in the Acute Care Setting. Clinical Infectious Diseases. 2018;67(8):1168-1174.
doi:10.1093/cid/ciy255
11. Huttner B, Pulcini C, Schouten J. De-constructing de-escalation. Clin Microbiol Infect. 2016;22(12):958-959.
doi:10.1016/j.cmi.2016.09.024

73
Figure 1: Bacteria were classified into classes based on their susceptibility to increasingly broader antibiotics,
while therapy desirability was ranked into six levels.

R-R-R-R-S
R-R-R-S-S
R-R-S-S-S
R-S-S-S-S
S-S-S-S-S
Treatment/resistance
Narrow - Amoxi/clav., ampi/sulbactam
Intermediate I - ceftriaxone
Intermediate II - pip./tazo., cefepime, ceftazdime
Broad I - carbapenemes
Last resort - Last resort - new BL/BLI, cefiderocol

Ranked combination Score assignment

according to antibiotic
chosen
Ideal (most desirable) 100
Slight overtreatment 80
Moderate overtreatment 60
Heavy overtreatment 40
Severe overtreatment 20
Undertreatment (less desiderable) 0

74
Figure 2: The number of de-escalations made, categorized by type, in relation to those possible

Type of de-escalation
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Stop Gram + From comb to Narrower Stop Gram + and From Stop Gram + and All types
mono spectrum narrow combination to monotherapy
narrow

Observational Interventional

75
Table 1: Characteristics of observational and interventional cohort populations

Variable Observational (n=42) Interventional (n=43) p

Ward chategory, n (%)
Medical 35 (83%) 33 (77%)
Intensive care 5 (12%) 7 (16%) 0.78
Surgical 2 (5%) 3 (7%)
Age, median (IQR) 77 (67-84) 75 (66-85) 0.83
Males (%) 23 (55%) 25 (58%) 0.83
Smoke (%) n=30 n=23
Yes 9 (20%) 4 (17%)
Former smoker 15 (50%) 10 (44%) 0.76
No 9 (30%) 9 (39%)
Obesity (%) 11 (28%) 9 (22%) 0.61
Charlson comorbidity index, n (IQR) 6 (4-8) 5 (4-7) 0.52
Penicillin allergy (%) 1 (2%) 2 (5%) 1
Type of infection (%)
Bacteremia 5 (12%) 6 (14%)
UTI 28 (67%) 25 (58%) 0.63
LRTI 2 (5%) 5 (12%)
IAI 7 (17%) 7 /16%)
Bacteria (%)
E. coli 29 (69%) 30 (70%) 1
K. pneumoniae 13 (31%) 13 (30%)
Coinfection (%) 9 (21%) 3 (7%) 0.07
APACHE-II score, n (IQR) 11 (8-15) 13 (10-18) 0.05
Deceased (%) 5 (12%) 7 (16%) 0.76

76
Table 2: Multivariate Cox model for 30 day mortality in hospitalized patients with positive blood cultures for E.
coli or K. pneumoniae

Variables HR 95% CI P value

DOOR MAT
Score 2 Ref Ref Ref
Score 3 3.75 0.1-127.4 0.5
Score 4 3.98 0.2-81.6 0.4
Score 5 2.7 0.2-44.1 0.5
Males 1.22 0.3-4.4 0.5
CCI>5 1.6 0.5-5.8 0.5
APACHE-II >10 2.2 0.4-11.3 0.4
Coinfection 2.0 0.3-16.7 0.5
Type of infection
Bloodstream infection Ref Ref Ref
UTI 0.11 0.01-0.7 <0.05
IAI 0.3 0.04-1.7 0.2
LRTI 1.5 0.2-11.1 0.7
Type of bacterium
E. coli Ref Ref Ref
K. pneumoniae 0.25 0.1-1.2 0.08

77
Table 3: Multivariate Cox model for in hospital stay longer than 14 days in patients with positive blood cultures
for E. coli or K. pneumoniae

Variables HR 95% CI P value

DOOR MAT
Score 2 Ref Ref Ref
Score 3 4.8 0.9-25.6 0.06
Score 4 2.3 0.6-9.7 0.4
Score 5 1.4 0.4-5.1 0.6
Males 0.93 0.6-3.7 0.9
CCI>5 1.5 0.5-5.8 0.3
APACHE-II >10 1.7 0.6-4.6 0.3
Coinfection 0.7 0.1-4.1 0.7
Type of infection
Bloodstream infection Ref Ref Ref
UTI 1.4 0.3-7.7 0.7
IAI 0.7 0.2-2.5 0.5
LRTI 0.9 0.07-11.2 0.9
Type of bacterium
E. coli Ref Ref Ref
K. pneumoniae 1.29 0.3-6.1 0.7

78
 ESCMID AMS Certificate Program
My Project Report

Name
Mona Mustafa Hellou
Institution
Rambam Healthcare Campus
Country
Israel
Tutor
Paul van der Linden
Title of Project
Optimizing blood culture collection practice
Date of submission
March 16, 2024

Abstract:
Background: blood cultures are essential diagnostic tool for blood stream infections. When indicated, it should be
collected according to the best practice standards and in the right setting. Inappropriate blood culture collection may
influence negatively the results, leading to high contamination rate when overused. Moreover, it could lead to
underdiagnoses when not obtained prior to antibiotic treatment, and difficulty in de-escalation of antibiotic therapy, thus
overuse of broad-spectrum antibiotics which increases antibiotic resistance development. As an antimicrobial
stewardship (AMS) project, we aimed to improve the blood culture collection process in the emergency department (ED).
Methods: we did a retrospective audit from January 2023 till March 2023, measuring quality metrics of blood culture
collection process from electronic database. The results were reported to a focus committee that was established for the
target of this project. A bundle of interventions was implemented and actively monitored and reported monthly to the
committee for a 6-months period. Later on, rethinking and optimization of the interventions has been made. Another
period of monitoring is expected to start in April 2024.
Findings: our pre-interventional results showed that the compliance to the best practice standards of blood culture
collection was low. During the intervention and monitoring period there has been an increase in the rate of obtaining at
least two blood culture sets from 6.5% in the pre-interventional period to 15%. Another increase has been shown from
41% to 83%, 69% to 80% and 78% to 93% in the rate of collection date & time documentation, rate of collection site
documentation and the rate of incubation within 3 hours from the collection, respectively.
Conclusion: our AMS interventional project has had an overall positive impact on performance of blood culture collection
process in the ED, yet more should be done to strive for optimization.

Stage 1: Introduction.
I work as an Infectious disease (ID) physician at Rambam Healthcare Campus. It is a tertiary academic hospital with 1030
beds, that serve a population of 2 million people in north of Israel. It has an ED with 400 daily visits, six Internal medicine
departments and subspecialties (nephrology and cardiology departments), hematology and HSCT unit, one big surgery
department consisting of 5 units, neurosurgery with an ICU, OBGYN department, heart and chest surgical department
with an ICU, two orthopedic departments, two general ICU departments, two oncology departments and a new ED for
oncologic patients, and a separate children’s hospital. We also have ambulatory clinics. At our hospital an AMS program
was available partially; with preauthorization approach, 24-hours ID consultation availability, treatment protocols and
educational material, but with no defined AMS team.
As part of a national surveillance program for adequacy of blood culture management, it was found that in our hospital,
compliance to the best practice standards of blood culture collection was low; mainly with obtaining small blood volumes
and low rates of obtaining at least two sets of cultures per patient. This, despite the program of the IPC unit for instruction
of new staff members on how to take blood cultures properly.
I chose to start a project on diagnostic stewardship with an aim of optimizing blood culture collection practice, because
of the importance of cultures as a diagnostic tool which guides physicians to better treatment and hence better patients’
outcomes, and on the other hand allowing antibiotic stewardship by de-escalation of antibiotic therapy and even
stopping it when unnecessary.

79
Stage 2: Measurement.
The national center for infection prevention & control and for antibiotic resistance in Israel monitors quarterly the blood
culture performance among hospitals. Preanalytical, analytical and postanalytical 14 quality metrics are monitored and
reported to the hospital’s director, head of IPC unit and director of ID unit.
After receiving the results of our hospital, we performed an internal audit by a three months retrospective monitoring
from 1.1.2023- 31.3.2023, using our local electronic program for data extraction which retrieves the data from our
microbiology lab. Because of the complexity of implementing an AMS intervention considering all the 14 metrics and all
the departments in the hospital, we have decided to focus on the ED of adults specifically the nursing staff, as our study
group of interest, since they are the responsible for blood culture drawing, choosing preanalytical quantitative metrics
which are related mostly to human performance. Our metrics included:
1. Mean blood volume (ml) per bottle.
2. Rate of obtaining at least two blood culture sets (within one calendric day from the first set) for a patient.
3. Contamination rate (per patient). For contamination assessment we took the blood cultures with a growth of coagulase
negative staphylococci in only one or two blood culture bottles for a patient.
4. Rate of collection date & time documentation.
5. Rate of collection site documentation.
6. Rate of blood cultures incubated within 3 hours from the sampling.
7. Appropriate performance rate; which includes drawing at least 2 sets for a patient, with time, date and venipuncture
site documentation and incubation within 3 hours of sampling. It is a metric that is not reported by the Israeli national
center for infection prevention & control and for antibiotic resistance. We have decided to add it because we think it
reflects the optimal practice of blood culture collection.
A total of 4140 blood culture bottles were obtained during the retrospective study period from 1.1.2023 till 31.3.2023.
Our overall performance in blood culture collection in the ED is low as it shown in table 1. The mean blood volume was
4.6 ml when the recommended volume according to previous studies is 8-10 ml of blood per bottle. The percentage of
drawing 2 sets is 6.5% which is very low, as it is recommended to obtain always at least 2 sets when blood cultures are
indicated. The contamination rate is 4% which is very close to our target, but we aspire to decrease it even more, as some
facilities succeeded to reach below 1% of contamination rate. Only 41% of cultures with documentation of date and time
of collection, and 69% with collection site documentation when the target is 100%. As the sensitivity of blood cultures is
crucially influenced by the transmission time of the bottles from collection to incubation, we thought that a rationale
transmission time would be up to 3 hours. According to the results, 78% of the bottles were incubated within 3 hours
which is relatively reasonable compared to the performance on other metric, but we think there is a chance for
improvement reaching 100%. Moreover, the appropriate performance rate was 1% which is disappointingly very low.

Quantitative metric Definition Performance Target

Mean blood volume (ml) Extracted from the company’s data 4.6 ml 8 ml

Rate of obtaining at least two Number of patients with at least 2 sets per episode/ 6.5% 50%
blood culture sets per patient number of patients for whom blood cultures were
drawn
Contamination rate Number of patients with blood culture with CONS/ 4% < 3%
number of patients for whom blood cultures were
drawn
Rate of collection date & Number of bottles with documentation/ total 41% 100%
time documentation number of blood cultures bottles
Rate of collection site Number of bottles with documentation/ total 69% 100%
documentation number of blood cultures bottles
Rate of incubation within 3 Number of bottles incubated within 3 hours from 78% 100%
hours from the collection collection/ total number of blood cultures bottles
Appropriate performance Number of bottles with 2 sets, date & time & 1% 50%
rate collection site documentation and incubation within
3 hours/ total number of blood culture bottles
Table 1. Description of the quality metrics, its definitions, results from the retrospective audit and our achievement long
term targets.

80
Stage 3: Barriers and facilitators.
At first, we have formed a focus group composed of two ID physicians (director of ID unit at our hospital and myself), the
director of the ED, and the head of the nursing staff in the ED. I presented the data to this focus group and later to the
nursing staff, since they are our group of interest being the responsible for obtaining blood cultures in the ED. Using the
Cabana framework and the COM-B system, we identified the following barriers:
- Lack of knowledge. Most of the nursing staff involved in drawing blood cultures didn’t know about the need for
at least two sets of blood cultures before starting antibiotic treatment.
- Lack of awareness. This was especially prominent in the contamination issue.
- Lack of self-efficacy. There was some member of the staff who believed that they cannot perform the project.
- External barriers:
• lack of time and resources. The lack of a dedicated team- for example, phlebotomist team, for the task
of blood culture collection in the ED. The project could produce a work overload on the nursing staff
reducing their capability of optimal performance especially regarding the collection of at least two sets
of blood culture for each patient. Later on, during the implementation period, this barrier has led to
the recognition of the lack of uniform definitions and a working algorithm as another barrier.
• lack of reimbursement. This affects negatively the motivation.
• Cumbersome process. Since we still don’t have a computerized program for entering patients’ data,
we have to fill out the data in a written clumsy form which is time consuming; then we have to tie the
paper forms to each bottle and then to send it to the lab

Regarding facilitators:
- The support of the ED management (directors of medical and nursing staff).
- The accountability and commitment of the focus team for engagement in the project.
- The feasibility of extracting data from the electronic database.

Stage 4: Development and execution of the stewardship intervention.

The interventions and work plan are presented in a timeline frame in figure 1.
Since we had multiple barriers from different domains of the framework we had to prioritize the interventions addressing
these barriers according to their feasibility, and short- and long-term durability. The following interventions were made:

- To address lack of knowledge and awareness we intervened with Education. Interactive lectures were given with
information about the importance of blood cultures as a diagnostic tool for identifying the pathogen responsible
for a patient’s infection, thus improving patients’ outcomes by prompt diagnosis and appropriate treatment.
We also emphasized on the high contamination rate and its possible consequences including excessive
unnecessary work for the microbiology lab and other diagnostic procedures, and overuse of antibiotics. We also
presented the bundle of actions for adequate and optimal blood culture collection procedure, backed up with
recommendations from the literature, with an access to them.
- To address lack of self-efficacy we did a monthly follow up and feedback on the performance of blood culture
collection by monitoring the defined metrics mentioned above. The report was sent to our focus group, and the
head of the nursing staff disseminate it to all the staff. Using their own performance data is a kind of mixed
interventions of Persuasion- by inducing positive feelings when there is improvement, and negative feelings
when there is regression-, and Modelling; by being their SELF models.
- To address lack of time and resources. To address this barrier, we chose Enablement by asking the hospital
management for a dedicated phlebotomist team, but unfortunately our request was refused because of limited
financial resources. We realized at this point that we have to think of other solutions to engage the nursing staff
for better performance. So, we used Persuasion to induce positive feeling and stimulate action, by presenting
the performance of other hospitals and comparing it to ours.
- Regarding the lack of a working algorithm, we developed a simplified practical flowchart (figure 2) describing
the indications for blood culture collection. It was discussed and then disseminated to the staff.
- Regarding lack of reimbursement. Unfortunately, this barrier has not been addressed yet, because there is no
financial support to incentives, and we didn’t want to use Coercion because it may yield negative outcomes
specifically in this project.
- To address the Cumbersome process, we chose Environmental restructuring. We tried to design a simple form
only for blood cultures (not for the whole tests performed at the microbiology test) to enter the patient’s data,
but we were informed that a new computerized program will be launched soon, with no need for written forms
anymore.

81
Stage 5: Evaluation of the stewardship intervention.
As mentioned previously, we did a monthly follow-up measuring the metrics presented in table 1 from June 2023 till
November 2023. The results are presented in figure 3.
Ninety percent of the nursing staff was present at the first meeting where we had discussed the results of the
retrospective report, presented the project, discussed the barriers and presented a lecture on blood culture collection
procedure and its importance in diagnostic stewardship. Educational material and the monthly monitoring results during
the implementation period were disseminated to all staff members.
The monthly mean number of blood cultures was 1600 bottles (aerobic and anaerobic). Regarding the mean blood
volume, unfortunately we had technical problems so we couldn’t extract the data, but it was fixed recently so it will be
monitored during the next 6 months period of measurement.
The percentage of obtaining at least two blood culture sets in the pre-interventional period was 6.5%. During the
intervention phase it has raised to around 15%. Although we are far from the recommended target which is 100%
according to guidelines, still there is an increase, and to be more realistic we had determined a target of 50%. To note,
during October-November there was a significant decrease. We couldn’t determine the exact reason, but it could be
related to that time has elapse since our intensive intervention, or related to the outbreak of the war in Israel. One
explanation that came up later was the lack of uniform definitions and a working flowchart- a barrier that wasn’t
recognized at the beginning but only at this timepoint of measurement. We had addressed this barrier with developing
a working flowchart (figure 2) which was disseminated to the staff at the second week of March 2024, so its impact would
be pronounced in the next measurement phase. The contamination rate has decrease to 3%, but later on it has increased
again to 5% and remained stable. The rate of blood cultures with documented time & date of collection was 41% prior to
our intervention. After the intervention, it has increased so that in November the rate was 83%. A similar but more
moderate increase in the rate of collection site documentation was seen from 69% to 80%. With the introduction of the
computerized forms we strive a 100% achievement. There was another encouraging increase in the rate of incubated
blood cultures within 3 hours of collection from 78% in the pre-intervention to 93% in the end of the first measurement
phase (November 2023). The appropriate performance rate is extremely low, but an inspiring increase has been seen,
from 1% prior to intervention to 8% in the end of measurement phase.

Conclusion:
Although we haven’t come to the closing point of the project, but at this timepoint, our project has had an overall positive
impact on performance of blood culture collection process in the ED. Although the improvement was modest in some
parameters such as the rate of obtaining two sets, or absent in others as in the case of contamination rate, there seems
to be inspiring collaborative multidisciplinary engagement with promising achievements. Yet, more should be done to
strive for optimization.
At the beginning of the project I have faced some challenges, mainly trying to figure out the best way to outreach to
stakeholders and to engage them in the project. This is my first project as a leader of the AMS team in the hospital. During
the project’s time, I have earned experience in different fields; from dealing with technical and practical details that
appear during intervention implementation, to dealing with the psychological barriers affecting the success of the
project. Participating in the AMS Certificate program have enriched my knowledge and experience in lot of terms,
especially in the behavioral change psychology. At the hospital level, it gave me tools to develop our AMS program and
expand our activities. Over the last two years, we have performed interventional projects in different fields, including
technology intervention to decrease antibiotic use in orthopedic departments, updating and distributing the antibiotic
prophylaxis protocol in regard to the cross allergy of penicillin and cefamezin, designing electronic interactive antibiotic
consumption dashboard and others.

Lastly, I would like to sincerely thank my director Prof. Mical Paul for her support and guidance throughout the project
period, and for encouraging me to participate in the AMS certificate program.

82
 April 2023, 1st week
January-March 2023
• 1st measurement phase
• Retrospective audit
• Measuring the quality metrics

April 2023, 2nd week

Meeting with the focus group
• Presenting the results
• Planning the project

April 2023, 3rd week

Meeting with the nursing staff
• Presenting & discussing the results
• Presenting the project & discussing
barriers
• Educational session on appropriate
blood culture collection procedure

June-November 2023
• Measuring the quality metrics
monthly
• Reporting the results monthly
to the focus group and the
nursing staff
February 2024
Meeting with the focus group
• Presenting the results of the 6 months
in continuous graphs
• Analyzing the results & evaluating
current intervention
• Discussing existing & new barriers
• New barrier- lack of uniform definitions
and a working algorithm
• Discussing possible future interventions
March 2024
• Providing a simplified
working algorithm

Future plans…
April-August:
• 2nd measurement phase
• Computerized forms (summer 2024)
• Meeting with the focus group

Figure 1. Project Timeline

83
Figure 2. Working flowchart

A B
16%
15%
14% 6%

5% 5% 5% 5%
11%
10% 4%
8%
3%
6%

JAN-MARCH JUNE JULY AUGUST SEPTEMBER OCTOBER NOVEMBER JAN-MARCH JUNE JULY AUGUST SEPTEMBER OCTOBER NOVEMBER

C 91%
93% D
89% 8%
85% 85%
83%
80%
78% 78% 77%
75%

69% 70%
66% 67%
62% 63%
60% Date & Time
57% 3% 3% 3%
Collection Site
52% 2%
Incubation 3h
1% 1%
41%
JAN-MARCH JUNE JULY AUGUST SEPTEMBER OCTOBER NOVEMBER JAN-MARCH JUNE JULY AUGUST SEPTEMBER OCTOBER NOVEMBER

Figure 3. A- Rate of obtaining at least two blood culture sets. B- Contamination rate. C- Rate of collection date & time documentation,
Rate of collection site documentation, Rate of incubation within 3 hours from the collection. D- Appropriate performance rate.

84
ESCMID AMS Certificate Program
My Project Report

Name
Oana Joean
Institution
Hannover Medical School
Country
Germany
Tutor
Prof. Sören Becker
Title of Project
Culture matters: a multimodal behaviour change intervention to improve the quality of blood culture collection
in an academic emergency medicine department
Date of submission
17.03.2024

Abstract* (max 300 words)

Background:
Blood cultures are among clinical microbiology laboratories' most important diagnostic tests. Suboptimal
quality (contamination, few samples/patient, late testing timepoint, low blood volume) increases diagnostic
costs and can impact patient outcomes negatively. Here, we describe a behavioural change intervention aiming
to increase the quality of blood culture collection.
Methods:
The study was performed in an emergency medicine (ER) department at a large German academic hospital. It
included three phases: 1) problem evaluation, 2) behavioural diagnostic and intervention design and 3)
implementation. Semi-structured interviews following the COM-B (Capability, Opportunity, Motivation –
Behaviour) model were used to identify critical influences for the target behaviour.. A multimodal intervention
consisting of educational sessions for residents, education and persuasion targeting nurses and residents in the
ER, monthly reminders to the ER physicians and environmental restructuring of the ER was performed.
Results:
In the pre-intervention phase, the contamination rate in the internal medicine department was 6.5%. This led
to additional diagnostic costs of up to 2.500 Euro/month and unnecessary antibiotic treatments for 10 [10;11]
(median[IQR]) patients/month. Almost 70% of the contaminated blood cultures had been drawn in the ER. At
the end of the implementation phase, sank from 6.5% to 2.7%. Only 27% of the contaminated blood cultures
came from the ER, compared to 70% pre-intervention. Furthermore, in median 3 less inappropriate
treatments/month occurred due to contaminated blood cultures.
Conclusion:
Our intervention was tailored to the setting and significantly reduced the contamination rate. Successful
interventions are only possible in a work culture open to interdisciplinary communication and where patient
safety improvement is a priority.

Stage 1 Introduction: description of the setting and the AMS program in place and identification of the problem

The hospital where I work as an infectious disease physician is a 1500-bed maximum care academic hospital
with a nationwide catchment area. It is also the largest transplant centre in the country.

An AMS team was already established at the start of the program. It consisted of a microbiologist, an infectious
disease physician (myself), and a clinical pharmacist with infectious disease training. Local treatment guidelines
were available and each department received feedback on prescription behaviour once yearly. Also, weekly
ward rounds occurred in some departments.
AMS activities were not a priority for hospital leadership, and our authority was limited. Therefore, the
adherence varied significantly between physicians and between departments; some departments were reticent
to comply with the recommendations or to become involved in the AMS rounds. Additionally, there was no
collaboration between the AMS team and the infection prevention and control team. Nurses were not involved

85
in any AMS actions. Due to time-constrains and high physician- and patient- turnover, not AMS activities had
been implemented in the emergency department (ER). Furthermore, the fact that most medication orders
were still paper-based was a high barrier to performing audits or point-prevalence studies.

A problem that was evident very frequently during ward rounds was the fact that microbiological diagnosis was
scarcely performed before the start of antimicrobial treatment. When performed, it was often incorrect and
led to inappropriate conclusions, most often to unnecessary antibiotic treatments. In particular, blood cultures
were often contaminated, there was only one pair of blood cultures drawn, they were not sent immediately to
the laboratory (especially from the emergency room) and they were not drawn before antibiotic treatment.
This is why I decided to focus on blood cultures quality improvement.

Stage 2 Measurement

I performed a 6-month retrospective audit of the contamination rate of blood cultures (quality indicator) drawn
in the department on Internal Medicine (Jan.-June 2019). Data were extracted from microbiological reports and
patient files (clinical correlate, antibiotic prescription due to contaminated blood cultures). The year 2019 was
chosen because: 1) it was the last year before the pandemic (to avoid confounders related to changes in patient
management structures associated with COVID19 and 2) the paper-based patient files would have been already
scanned (no electronic records available for general wards).

Contaminated blood cultures were considered:

a) Single positive detection of coagulase-negative Staphylococci, Corynebacterium spp., Bacillus spp.,
Micrococcus spp. and / or Cutibacterium acnes and without any clinical correlation according to patient
documentation
b) detection of several species from the same patient without correlation to the patient presentation (e.g. 3
types of coagulase negative Staphylococcus spp. shortly after a resuscitation for myocardial infarction)
evidence

The following quality indicators were assessed:

1) Structure indicator: the availability of a standard operating procedure/guideline for correct blood-culture
sampling
2) Process indicator: the blood culture contamination rate
The contamination rate was calculated as: number of contaminated blood culture sets/total number of blood
culture sets collected.
3) Outcome indicators:
- Patients treated due to contaminated blood cultures/month median (IQR)
- Additional diagnostics costs due to contamination: all contaminated blood cultures have the
consequence of performing an antibiogram, which increases the cost of diagnostic from approximately 14 Euro
to approximately 103 Euro/bottle (a net increase of 89 Euro/bottle). Costs due to unnecessary antibiotics were
not assessed.

Results of audit:
1) Structure indicator: no standard operating procedure on correct blood culture sampling was available
2) Process indicator: the blood culture contamination rate in the aforementioned timeframe was 6,5%; More
than 70% of contaminated blood cultures had been sampled in the ER.
3) Outcome indicators:
- 10 [10;11]/month (median [IQR]) patients had been treated due to contaminated blood culture
results
- approx. 2.500 euros / month extra expenses due to additional diagnostic costs

Conclusion/interpretation: the contamination rate was higher than the CLSI recommendations (<3%) and led to
inappropriate antibiotic therapies and additional diagnostic costs. Most contaminations seemed to happen in
the ER, which is why the following steps focused on this department.

86
Stage 3 Barriers and facilitators

I performed semi-structured interviews with physicians and nurses from the ER and observed how blood
cultures were being sampled in different shifts and from different persons/professional groups.
The following barriers were identified using the COM-B model and the Theoretical Domain Framework:

1) Knowledge and skills (capability) on how many blood culture pairs should be sampled, on whether sampling
should take place at different intervals/sites, on adequate hygiene.
NurseA: ‘I always thought 2 blood cultures meant 2 bottles, not 2 pairs’
PhysicianA: ‘I often have difficulties drawing blood cultures at different sites because our patients have difficult
vascular access. I try a lot and possibly contaminate the skin/sample or I drop it and I only send one pair of
blood cultures.’
PhysicianB: ‘I don’t remember having heard that drawing at least 2 sets of blood cultures was important. I
reckoned 1 set should be enough. Our patients are usually anaemic and I didn’t want to harm them’

2) Beliefs about consequences (intrinsic motivation)

PhysicianA: ‘Yeah…..I rarely get to see what happens with the patients I admit. I never really thought that a
false positive result might harm them…’
PhysicianC: ‘I always thought that culture contaminations would be ignored’
PhysicianD: ‘If the ward receives a weird blood culture result, they can always sample again... Why should I
invest so much energy in this thing (i.e., sterile blood culture sampling)?’
NurseA: ‘another thing to burden our work-schedule…why is this more important than the other 20 things I
have to do?
ER attending: ‘wow, this costs us quite a bit over a year’

3) Social influences (extrinsic motivation) affecting one’s behaviour (assumed peer pressure to perform in a
certain way, social habits)
NurseA: ‘we’ve always sampled like this, everybody does it like this’
Nurse B: ‘this is the way I way taught from my supervisors, they do it like this, so it ought to be OK’

4) Environmental context and resources (opportunity)

Nurse A and Nurse B: ‘we don’t have enough space in the storing room for more blood culture bottles’; ‘we
would need to order more blood culture bottles than before’

Facilitating factors were the fear of harming the patients (Emotions; both from the nurses and the physicians)
and the support from the ER attendings and ER head nurses. Along the process, a colleague from the infection
prevention team got interested in the project and offered support with educational activities.

Stage 4 Development and execution of the stewardship intervention

To address the limitations in knowledge and skills, a local standard operating procedure defining the correct
sampling method was developed by the AMS team and the infection, prevention, and control team. When
uploaded to the clinic’s internal network of the clinic, it was available to both physicians and nursing staff (both
professional groups involved in blood sampling).
Furthermore, educational sessions were organised for both physicians and nurses: these were mandatory for
physicians before rotating to the ER, nurses had the opportunity to attend such a session every 6-8 weeks
(nursing staff shifts were taken into account). During these sessions, information about correct sampling and
the standard operating procedure was offered. Blood culture sampling was also demonstrated (Education,
Training, Enablement). In addition, the immediate consequences for the patients were discussed and support
from senior physicians was reiterated (Persuasion, Enablement).

To address the limitations in space, the storage room in the ER was reorganised by the head nurses
(Enablement, Environmental Restructuring).
Once monthly, ER physicians were reminded about the importance of correct sampling by email (Persuasion).

87
 Figure 1: Intervention design using the COM-B model

Figure 2: Timeframe of development and implementation of intervention

Stage 5 Evaluation of the stewardship intervention

To evaluate the intervention, I performed an audit of the months Feb-May 2022, considering the same
indicators as in stage 2.
1) Structure indicator: a standard operating procedure (SOP) was now available
2) Process indicator: the blood culture contamination rate was 2.7 % (a decrease from 6.5%); 27% of the
contaminated blood cultures had been sampled in the ER (drop from 70%)
3) Outcome indicators:
- 7 [7;8]/month (median [IQR]) patients had been treated due to contaminated blood culture results
(in median 3 less patients than in the previous audit)

88
 - costs not available at the time of evaluation

Process evaluation:
Ten out of 12 physicians attended the educational session, which reached all at once. In contrast the
educational sessions did not reach all nurses at once and were attended by 18 out of 25 nurses. All received
email reminders.
I did not perform another set of semi-structured interviews, nor did I ask for structured feedback from the ER
personnel. Most physicians rotate every 6 months; to ensure continuity, an education and session training for
doctors will be performed before every rotation. The implication of the attendings and head nurses was
probably the most important facilitator.
The lack of electronic records was a high barrier to timely audits and feedback. I didn’t perform an interrupted
time series analysis, so some of the change I observed could have been related to a general trend over the past
years.

Conclusions

The most important thing I learnt during this project was that the limitations I had assumed were not
necessarily the most important limitations. In other words, I learnt that I need to engage in a dialogue with all
the stakeholders. I also learnt that nurses and physicians might have very different problems and ways to deal
with problems. Because nurses are heavily involved in many diagnostic steps, I think they should be involved
more in AMS programs. While spending time in the ER, I discovered other targets for future diagnostic
stewardship programs (urine dip, urine culture, wound swaps, etc).

I realised that correct diagnostic sampling was not part of the formal training of medical students at my
university (feed-back from colleagues, who had studied here). I started a ‘Diagnostic Stewardship’ seminar for
5th year medical students. Thus far, it is optional but it was attended by 127 students in 2023 (approx. 25% of
the cohort).
Together with ID and microbiology colleagues from other German medical centres, I performed a nationwide
web-based survey among healthcare workers (both physicians and nurses) and diagnostic laboratory
personnel. Its main aim was to point out problems at the different stages of the diagnostic pathway and to
identify targets for future diagnostic stewardship improvement interventions. It explored domains like:
‘interprofessional cooperation’, ‘knowledge level’ and ‘perceived need for improvement’. Preliminary results
will be presented at this year’s ECCMID (posters P2641 and P2642) in the session ‘Antimicrobial stewardship
and prescribing’.

89
ESCMID AMS Certificate Program
My Project Report

Introduction

Name
Osman Sezer Cirit
Institution
Gaziantep City Hospital
Country
Türkiye
Tutor
Sophie Schneitler
Title of Project
Blood culture diagnostic intervention in time of earthquake catastroph to improve diagnostics and therapy
Date of submission
18.03.2024

Abstract* (max 300 words)

Background
Diagnosis of bloodstream infections remains one of the most critical functions in hospitals and there are known
gold standard in diagnostics. The aim of the study was to evaluate and create better blood culture (BC)
diagnostics (BCD).

Methods
After identification of deficits in BCD. A case-control-intervention study was carried out on two intensive care
units at the Gaziantep Dr. Ersin Arslan Training and Research Hospital in Türkiye. The study comprised four
different phases: May to June (measurement), July (Diagnostic Stewardship intervention including staff survey,
training, and, continued measurement), from August to October (measurement), and from November to
December 2023 (measurement). Interactive BC-training was conducted on the case ward, accompanied by a
pre- and posttest-survey consisting of questions on demographics and 10 topic related multiple-choice-
questions. We assessed contamination rate in BCs. Data was extracted from the LIS. We also compare the pre-
and posttest-results.

Results
A total of 59 (100%, females 37/ 68%) nurses received the training. Participants was in 80% between the ages
of 20-29 an 46% had 2-5 years of job experience. Only 15% of participants reported receiving formal training in
BC sampling, most answered “by observing” and “with the help of senior colleagues”. 19 nurses (32.2 %)
completed the post-test. The intervention led to a considerable increase in the rate of obtaining two or more
sets. Reduced contamination rates could be detected at case ward.

Conclusions
The study was conducted in the earthquake catastrophe and the hospital was directly affected. However, we
got a lot of support for BCD beside the environmental circumstances. We introduced new and faced barriers
(e.g. not to know which staff will work further after the earthquake). However we saw in time of catastrophes
it is also possible to implement new and necessary measurement is this is the best diagnostics for patients.

90
Stage 1 Introduction: description of the setting and the AMS program in place and identification of the problem

The hospital I work as a Clinical Microbiologist is a training and research hospital with 880 beds. According to
the 2021 statistics total number of patient days was 299913, total number of patient admission was 1176192.
In my hospital, inappropriate sampling of blood cultures takes place, especially in the intensive care unit. Blood
culture samples from most patients were collected and sent in a single bottle, and there was no habit of taking
them as a set or sending them in an anaerobic bottle. We aimed to compare two separate intensive care units,
which we designed as control and case. We trained the nurses working in the intensive care unit, which we
determined as cases, on blood culture collection. We conducted a pretest survey consisting of 13 questions to
the participants before the training (Supplement 1).

Study design
A case-control-intervention study was carried out on two intensive care units at Dr. Ersin Arslan Training and
Research Hospital in Gaziantep, Türkiye, selected based on their bed capacity, number of staff, and previous
number of blood culture request numbers. The study comprised four different phases: From May to June
(period 1; measurement), in July (interim period; Diagnostic Stewardship intervention including staff survey,
training, and, continued measurement), from August to October (period 2; measurement), and from November
to December 2023 (follow-up period; measurement).

Staff questioning and training

A total of 59 nurses in the intensive care unit, who were identified as cases, were given 20-25 minutes of
interactive training to raise awareness about correct blood culture collection accompanied by a pre- and
posttest survey, consisting of 3 questions regarding demographic information and 10 multiple choice questions.
At the end of the training, a 7-8 minute video on correct blood culture collection was shown.
After the second period, the staff survey was used again with the same questions as a final posttest.

Stage 2 Measurement

To evaluate proper blood culture sampling, we assessed a 2 month retrospective and 4 month prospective of
blood culture quality indicators. We assessed:
(1) contamination rate in blood cultures (Fig.1)
(2) positive result rate in blood culture and
(3) the rate of obtaining two or more sets of blood cultures using Quality Indicators (Fig.2).
Data was extracted from the laboratory and hospital information system.

Stage 3 Barriers and facilitators

The following barriers to guidelines are:

(1) lack of awareness of local guidelines.
Regarding intensive care nurses:
(2) Lack of awareness on the subject.
Regarding blood culture bottles:
(3) Lack of awareness of the set concept and anaerobic bottle and the need of standard procedure by taken the
blood.

Stage 4 Development and execution of the stewardship intervention

To address:
(1) the lack of confidence in the guideline, a local guideline updated according to the national guidelines. To
increase
(2) nurses awareness, training programs were organized on this issue, and pretests and questionnaire were
conducted before these trainings.

91
Stage 5 Evaluation of the stewardship intervention
A total of 59 (100%, females 37/ 68%) nurses received the training and completed the test. Participants was in
80% between the ages of 20-29 an 46% had 2-5 years of job experience. Only 15% of participants reported
receiving formal training in BC sampling, most answered “by observing” and “with the help of senior
colleagues. 19 nurses (32.2 %) completed the post-test. The intervention led to a considerable increase in the
rate of obtaining two or more sets. Reduced contamination rates could be detected at case ward.

Conclusions

What have you learned?

I observed that providing face-to-face training on proper blood culture sample collection increased the
motivation of both me and the staff to implement the best diagnostics for the patients.
What has the project brought to your daily practice?
When I evaluated the pretest answers before the training, I became aware of which topics I should focus more
on in the next training. When I saw the visible improvement in the quality indicators that I used to evaluate
whether the training result was effective, I believed that I should extend this training to all personnel involved
in microbiology sample collection.
Describe the impact of the program on the development of an Antimicrobial Stewardship Program in your
own institution: what have you been able to achieve over the 2-3 years on the ESCMID AMS Certificate
Program?
First of all, I established the antimicrobial stewardship and prevention and control of healthcare-associated
infections working group at the Clinical Microbiology Specialty Society (KLIMUD), where I am a board member.
We organized a webinar and an onsite course in 2023 to raise awareness about this field. In our 2024 planning,
we planned one onsite course, one onsite regional meeting and five webinars to strengthen this activity in our
society. Our working group does not only consist of microbiology experts; we also invited infectious disease
specialists, infection control nurses and clinical pharmacists as invited members to emphasize the importance
of teamwork and interdisciplinarity.
Although I currently serve as an active member of the infection control committee at the hospital where I work,
we are in the process of establishing the antimicrobial stewardship committee, which is little known in our
country, together with the infectious disease specialist and clinical pharmacist we are on the pathway to
develop further standards in AMS. My project results are now helpful to show up the need for our hospital.
Which (un)expected developments influenced the process (positively or negatively)
At the beginning of the study, it was aimed to compare the intensive care units in two different buildings of the
hospital. However, we had to make changes in the project design as one of the buildings was affected and
collapsed in the earthquake that occurred in our region in February. Two different intensive care units located
in the same building were included in the study. In July, a pretest was administered to nurses in the case group
and training on correct blood culture collection was completed. While the number of personnel planned to be
post-tested during the follow up period was 59, only 19 personnel could be post-tested due to some of the
personnel being transferred to the newly built hospital and it was not possible to reach them any longer.
However, beside of this there were a lot of changes over weeks in staff plans, ward plans etc., before at the
end it was possible to start the study.

92
SUPPLEMENT 1

Questionaire
(This is a short questionaire related to blood culture diagnostics. The recommended time you need to answer
this test is 10 minutes. Each question has only one answer. Thank you very much for your participation.)
1) How many years of job experience do you have?
a) <6 mo
b) 6-12 mo
c) >1-2 y
d) 2-5 y
e) >5 y

2) How old are you?

a) <20 y
b) 20-29 y
c) 30-39 y
d) 40-49 y
e) 50-59 y
f) over 60 y

3) What is your gender?

a) Male
b) Female

4) Have you ever been previously officially trained in blood culture sampling?
a) Yes
b) No
c) I’m not sure
d) No comment

5) How were you been previously trained in blood culture sampling technique? (more than one answer
possible)
a) By observing and helping seniors
b) Explained orally by colleagues
c) Audio-visual aids
d) Other method (from HIS)
e) I am not trained before

6) How many blood culture collections you have done before (approximately)?
a) Daily
b) Weekly
c) 1-2x/month
d) quarterly
e) I’m not sure

7) Is it necessary to do hand-disinfection before blood culture sampling?

a) Yes
b) No
c) I’m not sure
d) It is depending on the situation
e) No comment

8) What is the correct disinfection technique of preparation of skin site?

a) From centre to periphery
b) From periphery to center
c) Zigzag technique

93
 d) None of the above
e) Technique is not important

9) What is the optimal amount of blood volume needed for one bottle blood culture in adult patients?
a) Less than 5-7 ml
b) Between 8-10 ml
c) Between 11-19 ml
d) Between 20-30 ml
e) More than 30 ml

10) How much is the duration of Contact time (CT) for povidone iodine?
a) 30 second
b) 1-2 minute
c) 3-4 minute
d) I’m not sure
e) No comment

11) Does disinfection of the rubber stopper of the blood culture bottle before filling reduce potential
contamination ?
a) Yes
b) No
c) I’m not sure
d) It is depending on the situation
e) No comment

12) What blood culture bottles should be taken as standart for blood cultures ?
a) From an aerobic and fungal bottle
b) From an anaerobic and fungal bottle
c) From an aerobic and an anaerobic bottle
d) From an aerobic and mycobacterial bottles
e) From an anaerobic and mycobacterial bottles

13) Which of the following is FALSE?

a. The period in which the concentration of microorganisms in the blood is highest is the period of 30-60
minutes before the onset of fever.
b. Since it is not always possible to predict the period of fever, blood cultures should be taken as soon as
possible after the fever starts to rise.
c. Blood cultures should be taken after starting antibiotic therapy.
d. Blood cultures should not be taken for surveillance purposes.
e. Blood culture bottles should never be placed in the refrigerator before or after sampling.

Fig.1 The rate of contamination rate in blood cultures

94
 Contamination Rate
90
80
70
Number of BC

60
50
40
30
20
10
0

Case Case total Control Control total

Fig.2 The rate of obtaining two or more sets

The rate of obtaining two or more sets

100 100
100 92,3 100 100
100 100
80 75 80 80
66,6 75
60 50
60
33,3 37,5
40 33,3
20 28,5 50
20 18,1 33,3
0 0 0
0 0
0 0 0 0 0 0 0
0 0 0 0 0
0 0

Case Control

95
ESCMID AMS Certificate Program
My Project Report
Appropriateness of antibiotic use at day 3-4 d of admission
Name
Shakil A. Shakar
Institution
Regional Hospital North Denmark, Hjoerring
Country
Denmark
Tutor
Jaap ten Oever
Title of Project
Appropriateness of antibiotic use at day 3-4 d of admission
Date of submission
15 March 2024

Abstract* (max 300 words) ( copy of accepted abstract to ECCMID 2024 no: #02591 (350 words), accepted for
poster P2596 : appendix 1 )

Appropriateness of antibiotic use: An Antimicrobial stewardship implementation project in a regional

hospital in North Denmark
Background
An Antimicrobial stewardship (AMS) implementation project is an essential part of the ESCMID AMS Certificate
program. Before initiation the perceived problem at our 200-bed hospital was that the prescribed antibiotics
are not timely switched from intravenous to oral solutions, further not changed to a reduced spectrum
antibiotics when applicable, and not timely discontinued. Lastly adherence to local antimicrobial guidelines at
admission seemed lacking. The aim of our project was to measure the above-mentioned parameters to disclose
areas for improvement.
Methods
We performed a retrospective point prevalence survey to measure the appropriateness of antibiotic use 3-4
days after admission. Defined population was all adult patients receiving intravenous antibiotics, initiated at
the department of Emergency medicine, and admitted at the department of Internal
medicine/gastroenterology and department of Geriatrics at day 3-4. We extracted a total of 200 patients
(chosen randomly, 100 from each department) from 1st of March 22 to 31st of Aug. 22 to assess the charts for
appropriate antibiotic choice and use.
Results
At the department of Internal medicine/gastroenterology 13 % were receiving antibiotics inappropriately at
day 3-4 (87% appropriateness). Further 17% of chosen antibiotics at admission were non-adherent to local
antimicrobial guidelines. At the department of Geriatrics only 5 % were found to receive antibiotics
inappropriately at day 3-4 (95% appropriateness). Here 18% were initiated on antibiotics non-adherent to local
guidelines. Rates for switch, change and discontinuation/stop of intravenous antibiotics was Switch=67%,
Change=14% and stop=14% at the Dep. Of Internal medicine/gastroenterology and Switch=61%, Change=5%
and stop=31% at the Dep. Of Geriatrics.
Conclusions
Our choice for improvement was in the department of Emergency medicine as the measured departments
didn’t leave much to be improved upon. To elevate their adherence to local antimicrobial guidelines, we
performed an open audit to disclose facilitators and barriers using the Flottorp framework as a reference. After
analysing barriers and facilitators their choice was to adapt the regional guidelines to their local settings. We
will present the post-interventional data 2 months after implementation after measuring their performance.

96
Stage 1 Introduction: description of the setting and the AMS program in place and identification of the
problem

I work in a regional hospital with approximately 200 beds. The hospital has a general internal medicine
combined with medical gastrointestinal, an abdominal surgery, an orthopedic surgery, and urology, a
gynecology, and a pediatric department. All these in conjunction with an intensive care unit with 6 beds. At the
start of the program, there was no AMS program available in our hospital. Treatment guidelines were available
with a possibility of remote supervision from clinical microbiologist or ID physicians on demand and by me, an
ID physician as well, if needed. But no one has actively been advising on appropriate use of antimicrobials.
Prior to the pandemic, the hospital went through a quality control program from 2017 to 2019. The aim was to
elevate the adherence to the current local antibiotic guidelines and to promote a timelier iv to oral switch of
antimicrobials. Base performance of the medical departments concerning iv-oral switch was a poor 55%, but an
excellent 90% at the end of the project period.

Thus, measurements and results of an improvement project concerning antibiotic use, must take this into
account when interpreting the results.

The perceived problem at the beginning of my project was that the prescribed antibiotics are not changed to a
reduced spectrum antibiotic, after positive microbiology has arrived. Further antibiotics are not timely
switched from iv to oral solutions, even when clinical and paraclinical result suggest otherwise. So not a timely
iv to oral switch. Next an In-appropriate length of antibiotic use, meaning antibiotics are not timely
discontinued, stopped or switched.
Finally, a large portion of the admitted patients are not started in antibiotics after the current updated local
antimicrobial guidelines. So, adherence to local antimicrobial guidelines seamed lacking too.

To design an improvement project, we used local data that has shown; 27% of the blood cultures that show
growth are positive after 1 day, 98% positive after 3 days incubation. Logistics will further demand ½ to 1 day
for transportation of blood-cultures without heat, to the regional microbiological lab, that provide us with this
service.
And finally, only 2% of blood-cultures becomes positive after day 3-7.

Stage 2 Measurement

By defining my perceived problem of in-appropriate antibiotic use at my hospital, we decided to perform a

retrospective point prevalence survey to measure the Appropriateness of antibiotic use from admission until
day 3-4.
For this we choose my department: Department of Internal medicine/gastroenterology.
This was presented to the Team of Quality to make my project an official quality improvement project.
By their support we then presented the project to the hospital leaders and head of departments for a final
approval by signing an official contract.

We then decided to extract a total of 100 adult patients (chosen randomly) from 1st of March 22 to 31st of Aug
22.
These were further defined as:
All adult patients receiving intravenous antibiotics, initiated at the department of emergency medicine, and
admitted at the dep. Of Internal medicine/gastroenterology at day 3-4. Then assess the charts from admission
until day 3-4 for appropriateness of antibiotic choice and use.

Collaboration was initiated with the hospital pharmacist for data collection abaid after the official approval of
the project (It took a month to finalize).

So, the aim was to review the charts by me as objectively as possible with this in mind, and measure
appropriateness of antibiotic use from admission until day 3-4 of hospital stay.

The number of patients receiving appropriate antibiotics 3-4 days after startup(nominator)

97
 in comparison to
The total number of patients initiated or still receiving iv. AB at 3-4 days after startup (denominator)

Definition of an appropriate antibiotics: (Just 1 of these criteria’s need to be fulfilled, if applicable)

1 Has there been a timely iv to oral switch?
2 Is the antibiotics timely been changed to a reduced spectrum AB after positive microbiology
3 has the antibiotics been timely discontinued?

We also decided on collecting data on:

Adherence to local antimicrobial guidelines
Is the antibiotics chosen at admission in accordance with local/regional antimicrobial guidelines?
Distribution on types of infections
SSTI, RTI, UTI, Sepsis, Other
Distribution on types of antibiotic classes chosen
Penicillin’s (J01C), Cephalosporins and other beta lactams (J01D), Macrolides (J01F)
Quinolones (J01M)

After collecting all these data, we decided to extract another 100 patients from the same time-period and
criteria from the department og Geriatrics.
This was because of a concern that by my presence in the department as a supervisor would have had
influenced the prescribers to a more appropriate antibiotic use. Thus, the total number of patients in this
retrospective survey became 200. And hopefully by this we have reduced my presence in the Dep. Of Internal
medicine & Gastroenterology in the measurements.

Dep. Of Internal medicine & Gastroenterology:

100 patients extracted randomly out of total hit of 265.
13 % inappropriate AB handling at day 3-4 (87% appropriateness)
17% non-adherent to local guidelines. 70% of these got wrongly started on Piperacillin Tazobactam
Switch: 67%, Change : 14%, Stop : 14%
SSTI: 23%, RTI: 23%, UTI: 21%, Sepsis: 28%, Other: 5%
Penicillin’s (J01C) 91, Cephalosporins and other beta lactams (J01D) 9, Macrolides (J01F) 5,
Quinolones (J01M) 13

Switched = from iv to oral solutions

Changed = to a reduced spectrum iv antibiotic
Discontinued/Stop

Dep of Geriatrics:
100 patients extracted randomly. Out of a total hit of 287
5 % inappropriate AB handling at day 3-4 (95% appropriateness)
18% non-adherent to local guidelines. 66% of these got wrongly started on PipTazo.
Switch: 61% Change: 5% Stop: 31%
SSTI: 4%, RTI: 34%, UTI: 30%, Sepsis: 29%, Other: 3%
Penicillin’s (J01C) 87, Cephalosporins and other beta lactams (J01D) 11, Macrolides (J01F) 5
Quinolones (J01M) 8

98
actrim ha il ha ar

actrim ha il ha ar

99
actrim ha il ha ar

actrim ha il ha ar

100
 actrim ha il ha ar

To my big surprise the appropriateness was quite high in both department and dept. Of geriatrics seams to
perform even better than my own Figure 1. Both departments displayed a high rate of timely iv to oral switch
which is essential for a short duration of hospital stay and in reducing overall antimicrobial resistance
development Figure 2. A bit surprising but maybe not after further analyses, shows that the dept. Of Geriatrics
have a high rate of discontinuation of antibiotics. Almost 1/3 of the patients had their treatment incl antibiotics
discontinued, as they were assessed to be at the end of their life.
A difference between the departments according to the infection site was noted. We see a higher rate of SSTI
in my department, whereas not surprising a higher rate of RTI in the Dep of Geriatrics Figure 3.
When it came to the choice of antimicrobial classes, both departments showed a similarly high rate of penicillin
use which is in accordance with our regional antimicrobial guidelines Figure 4.

Lastly, we had chosen to assess the adherence to the regional antimicrobial guidelines at admission Figure 5.
As it is shown 17% and 18% of the reviewed charts, in total 35/200 patients or 17,5% of the patients did not get
antibiotics as the local guidelines suggested.
A deep dive in these charts revealed that 70% of these with a non-adherent choice of antimicrobial, were
initiated on Piperacillin Tazobactam at admission out of a clinical assessment that the patient was in septic
shock. But in fact, the registered parameters did not suggest this. At least they did not fulfill the criteria for
infection without focus and possible organ failure with a blood pressure that is still low after relevant iv. fluid
supplements for at least an hour. They tend to define a septic shock wrongly at admission prior to relevant
stabilization and other treatments including fluid supplementation. This could be due to the guideline, so a
potential structural indicator of interest.

So finally, the measurements of appropriate antibiotic use at the 2 medical department did not leave a lot of
room for improvements. But it seems our culprit might be in the emergency department, and the improvement
should be focused on the dept. of Emergency Medicine. Although their adherence to regional antimicrobial
guideline is already excellent and high (>80%).
The aim was not to assess whether the treatment should at all be initiated, but the impression was that a fair
amount of the initiated antimicrobial treatment should be awaited until further assessment was done. An
opportunity for a future improvement project.

101
Stage 3 Barriers and facilitators

We decided to focus our aim and improvement project in the dept. of emergency medicine (ED).
This needed preparation as we had to get a new approvement from the leaders, which we got.
Using the Flottorp framework as a reference we aimed to disclose barriers and facilitators regarding adherence
to local guideline for choice of antibiotics upon admission.
To facilitate this, we invited the medical doctors from Emergency medicine to an open discussion. Further also
to choose an intervention strategy/area. After reviewing the above measurements they very positively wanted
to focus on the sepsis guideline and also prepare an Antibiotic man as a poster with the all first-line treatments
visible.

Barriers that got disclosed was as follows Figure 6:

Regarding guidelines (1) They had a lack of confidence in the local guideline, and found it non-applicable to the
situation they needed it in. Especially difficulty in deciding when a patient was defined as in septic shock and
when not. This because the respective antibiotic recommendations are different.
Regarding individual healthcare provider factors (2) Fear that the chosen antibiotic is of a too narrow
spectrum leading the patient’s condition to further worsen. Or the choice might harm the patient (
recommended first line treatment for sepsis without a focus : Iv Ampicillin + Gentamycin )
Regarding interactions of healthcare providers: (3) Upon consulting a patient with senior professionals the
suggested antibiotic was not always in adherence to local guideline.

Regarding individual
healthcare provider
factors
Fear of wrong choice
of antibiotics, and
harming patient Regarding interactions
of healthcare providers
Regarding guidelines:
Adwise from senior
lack of confidence in
professionals not
the local guideline
according to local
guidelines

Barriers for
adhereance
to local
guidelines

102
 Facilitators for
adhereance to
local AB
guidelines

They wants to
improve their image Highly motivatewd
towards other to do their best
departments

Want to contribute
in reducing AMR by
following local Ab
guidelines

Figure 6

Stage 4 Development and execution of the stewardship intervention

Our choice was to address:

(1) The lack of confidence in the guideline.

For this a local guideline committee was convocated by me and 2 clinicians from ED and developed an
adaptation of the local guideline for sepsis. We decided on developing our own adaption in form of a flowchart,
that could help and aid in choosing the correct antibiotic treatment regime by better helping the prescribers in
decision making, especially how and when to declare the patient in a septic shock.
The Team of Quality had another abandoned project as an Antibiotic man as a poster, where you could quickly
read the first-line antimicrobial treatments for pneumonia, UVI etc... This project was from 2019 where the
hospital wanted to elevate the adherence to the regional antimicrobial guideline. As it fits our aim, we
incorporated it in our implementation strategy. See Figure 8 and 9, where both are in Danish and the sepsis
flowchart is ready for print. We also developed a poster Figure 10.

(2) The fear of wrong choice of antibiotics.

We deliberately choose clinicians from ED to join the committee to promote a sense of ownership in the
intervention and present the finished flowchart for their colleagues for feedback, before final approvement.
Further we defined that the fear was embedded in the definition of weather the patient was in a septic shock
or not. This got specifically defined in the flowchart they developed and an emphasis on when and how to use
gentamycin. Especially a reduced kidney function as an EGFR < 30 ml/min. or a dialysis patient promoted fear
and drives a choice toward Piperacillin Tazobactam as a first choice for sepsis, instead of what the guideline
suggests as first line therapy. And the regional guideline changed as we went to the guideline group with these
concerns as others also had done. They updated the guidelines concerning gentamycin when and when not to
use. Until then there was no restriction on initiating gentamycin. This was done I collaboration with out
nephrologists.
In this process I got invited by the regional antimicrobial guideline group as an ID physician, to become a
permanent member of the guideline group, representing my hospital.

(3) Advice from senior professionals was not according to local guidelines:
We will make the local guideline in form of a flowchart visible in Ed as posters and as a printed pocket-guide.
This would address any clinician variability as consulting seniors sometime comes from bureaus were their daily
job is in a different region of Denmark, where the local guidelines are slightly different regarding the choice of
antibiotics in sepsis without shock.

103
As a supplement we also did a Com-B component analysis Figure 7.
The target behavior was Clinicians at ED should start Ab according to the local guidelines.
Our choice was to look at
1: Psychological capability to understand and interpret the local guidelines, where the guidelines at hand does
not always fit the specific situation the clinician experience in sepsis.
2: An Social opportunity where ED should be a stakeholder and should produce and adapt materials to their
own setting with support. This will promote a sense of ownership of the change. And lastly
3: An automatic (and reflective) motivation with a consistent interpretation of the adapted guidelines that are
accepted and approved by all clinicians working in ED, to a continues acceptance of the adapted/changed
guidelines in ED. As we wanted the senior to be the promoters of the adapted guidelines, we also wanted to
assess the reflective motivation, and promote the seniors to continuously support the adapted guidelines.
Further we promoted the idea that they should elect an antibiotic champion/ambassador amongst the seniors
whom I will work closely with in the future.

Target behaviour: Clinicians at ED should start Ab for sepsis according to the local guidelines
COM-B Component What needs to happen for this Is there a need for change? Yes/No
target behavior to occur?

Capability: Physical Be able to access guideline by Yes: posters and remake of guideline design
computers, Guideline posters, a
senior to confer with
Capability: Psychological Understanding and interpreting the Yes, guideline does not always fit the
local guideline situation the clinician experience in sepsis.

Opportunity: Physical Locally adapted flowcharts, AB Yes, these does not exist.
posters
Opportunity: Social ED should be a stakeholder and YES. When clinicians from ED involve
should produce such materials themselves in adapting the guideline, they
themselves with support will feel a sense of ownership of the change.
Motivation: Reflective Clinicians believing that following the Yes: As it is a behavioral change a continues
adapted local guideline is not harmful support in this decision making is needed
for the patient and will not reflect and should be promoted by seniors at ED
badly on the quality of work they
deliver.
Motivation: Automatic A consistent interpretation of the YES. A general continues acceptance of a
adapted guideline that is accepted locally adapted guideline.
and approved by all clinicians
working in ED
What COM-B components Psychological capability, social opportunity and automatic motivation need to change
should you focus on in your in order for the target behaviour to occur.
intervention?
Figure 7

104
 alg a n bio s ehandling
eningi s
se Penicillinallergi
P instru s e p vir ede pa enter o ithrom cin mg
mo icillin mg p.o dg. p.o

ipro o acin mg p.o

P vir et ebrile pa enter
ller ipro o acin mg mo icillin lavulans re Penicillinallergi
Penicillinallergi . mg mg o i o acin mg
ithrom cin lterna vt, ithrom cin
m P. mg i et,
dg ved re se anamnese l en lpenicillin io , g Penicillinallergi larithrom cin
sien eller eller Pheno meth lpenicillin mg eller
ri a io , g p.o o ithrom cin mg p.o

oderat sv r, Penicillinallergi e uro im

en lpenicillin io , g , g og
Penicillinallergi og larithrom cin mg
e ota im g og en lpenicillin io , g larithrom cin mg ller
etronida ol mg og entamicin mg g o i o acin mg p.o
og evt. entamicin mg g ma d gn ndocardi s
ma d gn og etronida ol mg se v r,
P instru s involvering a ere lungelapper
Penicillinallergi
eller sat
Pip Ta o g , g og o i o acin mg
larithrom cin mg
Penicillinallergi mpicillin g
ipro o acin mg p.o g entamicin mg g t pis pneumoni
og entamicin mg g ma d gn ppig hospitals onta t .e s.
larithrom cin mg
ma d gn eller dial se
o ithrom cin mg e uro im , g
Penicillinallergi mpicillin g ller mo icillin lavulans re
p.o
e uro im , g og g entamicin mg g mg mg p.o
entamicin mg g ma d gn g mo icillin mg p.o
ma d gn
spira onspneumoni
Pip Ta o g , g eller
Penicillinallergi mpicillin g mo icillin lavulans re
e uro im , g og g entamicin mg g mg mg p.o
entamicin mg g ma d gn en lpenicillin io , g og g mo icillin mg p.o
ma d gn evt.
lo acillin g suppleres ved ed penicillinallergi
alment p vir ede pa enter e uro im , g

Figure 8

105
Figure 9

PosterSepsisflowchar
tShakilShakarAMS.pdf
Figure 10

Stage 5 Evaluation of the stewardship intervention

Post-implementation data cannot be presented as implementation is yet to start as we have finalized the sepsis
flowcharts and posters, and send them for printing. We will re-measure the adherence two months after
implementation, where I want to include whether starting antibiotics was at all indicated, and by that initiate a
new improvement project. This will be in conjunction with an updated regional UTI guidelines that includes an
decision making algorithm to help with when to initiate antibiotics, or rather when not to.
Motivation to participate in an improvement intervention in the dept. of ED was at first a bit lacking. But luckily
my own wife is a senior physician in ED and head of education. She has also become my future AMS
ambassador and promoted the implementation project as highly important to reduce and control the overall
development of antimicrobial resistance. Many of the seniors were at first very sceptic and had a fear of being
disclosed as a bad performing department. But the young colleagues were highly motivated as they could air
their difficulty in using antibiotics even when specific guidelines are present. All the meetings held in ED was
with most of their staff present. They are continuously understaffed and are dependent on substitute personal
including medical doctors that are often from outside the region and hence not all staff members can be
reached concerning an implementation. So as overall goal is to elevate their adherence, a higher focus on the
adherence in the dail wor where the senior’s discus and comment on the choice and use o antibiotic is
already a big win. They have asked for continues education in the latest updated antimicrobial guidelines
where I will continuously update my wife so she can continuously facilitate this by promoting young doctors to
take responsibility for this.

106
Conclusions

Overall measurements were surprisingly good in all departments including the department of ED. And the
interest in our project have been immense as other departments have been asking to be measured similarly on
their performance. s we don’t have an D consultanc et, the pro ect has promoted awareness on antibiotic
use and now much more than before getting calls for ID consult, concerning patients often involving decision
concerning antibiotic treatment. This has now influenced my leaders to accept a future recruitment to
strengthen ID speciality in our hospital.

Before initiating the ESCMID AMS Certificate my leaders had a wish to establish a functioning OPAT unit.
So soon after initiating the certificate in summer 2022, we started the process of defining an OPAT unit in our
setting. This would show to be a very healthy collaboration with the administration, other OPAT units in the
region, Team for Quality and other senior staff-members. OPAT got initiated on sept 1st 2023 after nearly 1 year
of preparations. This have been a good investment of time and effort, as the same preparation must happen
with establishing of an antimicrobial stewardship program.
Since 2019, no antibiotic audit has been performed in our hospital. In collaboration with team for quality I was
kindly asked to facilitate this and be a part of a yearly audit in the future. Presenting the audit report connects
me with colleagues that normall don’t wor with and can acilitate an overall appropriate use o antibiotics
My leaders have accepted to initiate the preparations for establishing of an AMS program, starting March 2024.

On a negative note, some colleagues are afraid that I am educating myself to become an hospital antibiotic
police officer and will remotely review their patients charts secretly and send them correction notes daily. Or
enforce them to endure a session of how to adhere to the local guidelines.
I have tried to promote the idea that it is important to hinder the overall development of AMR, but some
colleagues, especially senior colleagues, takes it personal when advised otherwise in their choice of antibiotics,
and especially when advised to discontinue treatment. I think good feedback with understanding will remove
this scepticism over time. And luckily our national board of health have initiated an AMS guideline group to
promote AMS programs in all hospital in Denmark.

Understanding behaviour changing science as a tool and be able to use it in many different aspects by just
simply asking what facilitate an inappropriate behaviour and what are the barriers to change it. This becomes a
very essential tool that hopefully will elevate the success rate for any given implementation. And often it’s very
important to just sit and listen and not aim for immediate solutions as I tended to do prior to the AMS
Certificate.
And by the way, on a personal note, my wife is equally happy with this behaviour in me when it comes to
private matters. Although we have decided to skip the baseline data collection.

The overall impact by participating in the ESCMID AMS Certificate program has been huge for me and our little
200 bed hospital by my opinion:

It has promoted awareness on AMR on many levels. Including national level as we the Danish participants have
formed an AMS work group in our national infectious disease society where we are trying to influence
policymakers on a national level.
Helped initiate an OPAT unit to free up patient beds but also to be a part of a future AMS program.
Initiated a continues collaboration with the Team for Quality concerning antibiotic use in the hospital.

107
Restarted a hospital wide yearly Antibiotic audit to foresee in-appropriate antibiotic use and design new
improvement projects.
Become a fixed member of the regional antibiotic guideline group.

And most important goal reached with the ESCMID AMS Certificate was to acquire cutting edge knowledge to
prepare, initiated and run an AMS program, which is now in progress.

Appendix 1:

A0_Poster_ECCMIDSh
akilShakarP2596.pdf

108
ESCMID AMS Certificate Program
My Project Report

Name
Susana Ramos Oliveira
Institution
Hospital Pedro Hispano – Local Health Unit of Matosinhos
Country
Portugal
Tutor
Pilar Retamar Gentil
Title of Project
Adequacy of de-escalation therapy of bloodstream infections of a tertiary hospital: the impact of an
Antimicrobial Stewardship Intervention
Date of submission
17th March 2024

Abstract* (max 300 words)

Introduction: De-escalation from broad-spectrum to narrow-spectrum antibiotics is recognized as a significant

strategy to diminish the selective influence of antibiotics, however the adherence to this recommendation is
reportedly low. We report an antimicrobial stewardship intervention that occurred over a period of two years in
a tertiary hospital, focusing on the de-escalation therapy of bloodstream infections (BSI). It was developed by a
multidisciplinary team composed with Infectious Diseases, Clinical Microbiology and Internal Medicine physicians
and a Clinical Pharmacist.
Measurement: We retrospectively audit the episodes of BSI caused by Staphylococcus spp., Streptococcus spp.,
Enterococcus spp., Pseudomonas aeruginosa and Enterobacterales that occurred between September 2022 and
February 2023 in the Medical Department. The electronic patient files of all patients were revised manually.
Results: The overall appropriateness of drug choice for the microorganism and the source of infection was
superior to 95%, however the appropriateness of spectrum considering the antimicrobial susceptibility testing
(AST) (and co-infections) was 54%. Only 50% of patients with BSI by Staphylococcus aureus methicillin-sensible
received the best therapy indicated. The lowest appropriateness of spectrum was found to be in treatment of
Pseudomonas aeruginosa, where only 14% of patients received the shorter-spectrum considering the AST.
Intervention: An intervention was developed addressing the several barriers found using the Flottorp framework
as reference. Regarding factors relating to the individual healthcare professional, and professional interactions,
several educational sessions were performed. These involved clinical case discussions and some theorical key
learning points, and were positively received by the prescribing clinicians. Addressing the guideline factors (due
to the absence of a clear guideline of BSI management), a diagram of the targeted treatment of BSI of P.
aeruginosa was developed.
Conclusion: The evaluation of the stewardship intervention is planned to occur between September 2024-
February 2025.

Stage 1 Introduction: description of the setting and the AMS program in place and identification of the problem

At the start of the program – 2022 – I was a 3rd year resident of Infectious Diseases (ID) at a tertiary hospital with
380 beds.
The hospital has four major departments for in-patients: Medical (includes four 30-bed wards of Internal
Medicine, a 15-25-bed ward for COVID-19 patients, a 13-bed ward of Intermediate Care Unit), Surgical [includes
two 30-bed wards of general surgery, and three 30-bed wards of orthopedics (1), urology (1), ophthalmology and
otorhinolaryngology (1)], Emergency [includes the Emergency Department and Intensive Care Unit (12-24 beds)]
and Department of Pediatrics and Women Health (includes a 30-bed ward of Gynecology and Obstetrics and 20-
bed ward of Pediatrics).
Although the hospital team includes professionals in ID and other specialties (Pneumology, Cardiology,
Endocrinology, Dermatology, Neurology, Gastroenterology, Nephrology, Gastroenterology, Oncology, and

109
Hematology), these do not have a specific ward, and work, mainly, as consultants for the main departments
previously described. Each ward has a specific ID specialist consultant.
In my hospital, Selective Reporting of Antibiotic Susceptibility Testing (AST) Results was implemented in 2011 by
the microbiology department, together with one ID specialist. In 2014, it was obligatory to implement a program
to validate/adequate the use of some antibiotics of restrained use (carbapenems, ceftazidime/avibactam,
quinolones, and intravenous doxycycline), in the first 48h of prescription. This latter program is currently
performed by the ID specialists (7), but not by residents (including me), and it is the main AMS intervention
established in my hospital.
There are no current local treatment guidelines (empirical or targeted), however, the physicians are used to
sharing the decision of the (empirical) antimicrobial therapy with the ID physicians in the most severe cases.

The implementation problem I chose was the de-escalation therapy of bloodstream infections (BSI), in the
Medical Department. Most of the time, the targeted therapy is decided solely by the Internal Medicine specialist
(who is, ultimately, responsible for the patient) once the patient is stabilized. Also, because the short-spectrum
antibiotics are not validated by the ID team, the targeted treatment of multi-susceptible microorganisms is not
often discussed, leading to the use of a broader spectrum than needed.
The only data available on the adequacy of targeted therapy in my hospital was from a retrospective study
conducted by me, another ID physician, and some Internal Medicine doctors, referring to the infections managed
in OPAT/Hospitalization at home during a period of 17 months (not only BSI). The results showed that 44% of
patients received targeted therapy of a wider spectrum than indicated by the AST, due to the absence of the
antibiotic in the OPAT setting in 66% of cases. After this study was performed, (some) Internal Medicine doctors
reflected on the results and showed receptivity to discussing targeted therapy of multi-susceptible
microorganisms and understanding better the principles of antimicrobial use and stewardship.

This project was submitted and approved by the ethics committee.

Stage 2 Measurement

The microbiological agents selected for this project were Staphylococcus spp., Streptococcus spp., Enterococcus
spp., Pseudomonas aeruginosa and Enterobacterales. We decided to include several agents to increase the
sample size as my hospital has a small number of beds (~130-150 beds in the Medical Department).
Before deciding the period of analysis, a period of three months (December 2020 to February of 2021) was
assessed. In this period, 222 episodes of BSI by Staphylococcus spp., Streptococcus spp., Enterococcus spp.,
Pseudomonas aeruginosa and Enterobacterales occurred in my hospital, 145 of which in the Medical
Department. As some patients (not quantified) did not receive targeted treatment (due to early death or decision
not to treat), we decided to extend the period of analysis to six months, to guarantee a good sample size.

To evaluate the adequacy of targeted treatment of BSI, we performed a six-month retrospective audit of all the
episodes of BSI caused by Staphylococcus spp., Streptococcus spp., Enterococcus spp., Pseudomonas aeruginosa
and Enterobacterales (between September 2022 and February 2023). Microbiological data was provided by the
Head of the Department of Clinical Microbiology, and the electronic patient files of all patients were revised
manually. Only the patients who (1) had a microbiologically confirmed BSI by the agents previously described,
(2) with AST available (timely), (3) who received targeted treatment for the BSI (intention to treat population),
and (4) admitted in the Medical Department when targeted therapy was decided were included.
For every episode of BSI, we assessed: (1) the source of infection (catheter-related, respiratory, abdominal,
urinary, endovascular/endocarditis, skin and soft tissue, central nervous system, osteoarticular, and
indeterminate), (2) the empirical and (3) targeted therapy, (4) presence of other infection microbiologically
documented (that may alter the therapy), (5) intention to cover other agents empirically* (that may alter the
therapy), (6) reasons for not prescribing a certain antibiotic (if explicit), (7) duration of therapy, (8) recurrence of
bacteriemia at 30-days, (9) incidence of Clostridioides difficile colitis at 30-days, and (10) mortality at 30-days.

* - expressed in clinical notes. Example: “BSI by Escherichia coli sensible to piperacillin tazobactam. We maintain
therapy with ertapenem to guarantee ESBL coverage in a patient with previous infections by ESBL-producing
microorganisms.”

110
The quality indicators for BSI by all microorganisms included (a) appropriateness of drug choice for the
microorganism (efficacy), (b) appropriateness of drug choice for the source of infection (efficacy), and (c)
appropriateness of spectrum considering the AST, and other agents (if explicit).

(a) = x 100 %

(b) = x 100 %

( )
(c) = x 100 %

Results (depicted in table 1) show that the overall appropriateness of drug choice for the microorganism and the
source of infection was superior to 95%, however the appropriateness of spectrum considering the AST, and
other agents (if explicit) was 54%.
Only 50% (7/14) of patients with BSI by MSSA received the best therapy indicated: cefazolin (n=0) or flucloxacillin
(n=7). The remaining received treatment with amoxicillin/clavulanate acid (n=2), ceftriaxone (n=4) and
piperacillin/tazobactam (n=1), in some cases due to mixed infections.
The lowest appropriateness of spectrum was found to be in treatment of Pseudomonas aeruginosa, where only
14% (1/7) patients received the shorter-spectrum considering the AST.

The Defined Daily Dose (DDD)/100 bed-days of penicillin, amoxicillin/ampicillin, flucloxacillin,

amoxicillin/clavulanate acid, ceftriaxone, piperacillin/tazobactam, meropenem, imipenem-cilastatin,
ceftazidime/avibactam, vancomycin and daptomycin were assessed during the measurement period, in the
Medical Department. These results are depicted in Table 2. Expectably, these indicators quantify the antibiotic
use for all the infections in the Medical Department (not specifically the BSI).

Stage 3 Barriers and facilitators

Several (unformal) individual and group interviews with colleagues of ID, and Internal Medicine specialists and
residents were performed, before and during the intervention (see below). The majority of the interviews were
with an Internal Medicine specialist (part of the team, see below) and involved discussing the management of
some patients (real patients, part of the measurement population) and the motivations behind some prescription
patterns.

The following barriers were found using the Flottorp framework as a reference:
- Guideline factors:
. Quality of the evidence supporting the recommendation - According to the professionals in my hospital, the
quality of evidence that supports the recommendation on appropriate antibiotic use is not clear (p.e There are
no local guidelines. Also, there is a lack of clear evidence on the reason why is more appropriate to prescribe two
antibiotics with combined shorter-spectrum over one broad-spectrum antibiotic - p.e ertapenem and ampicillin
for coverage of an ESBL E.coli and E. faecalis (instead of imipenem), especially if this combination involves two
beta-lactams (concern for increased adverse effects); p.e ceftriaxone plus metronidazole over
piperacillin/tazobactam for the treatment of bacteriemia by E.coli from an abdominal source).
. Strength of the recommendation - The strength of the recommendation on appropriate antibiotic use is not
clear (p.e conflicting data on the efficacy of de-escalation strategies as a tool to prevent the development of
antimicrobial resistance).
. Feasibility of the clinical intervention - According to the professionals in my hospital, appropriate use of a certain
antibiotic is not practical/ perceived as not being practical (p.e using Penicillin for the treatment of Invasive
Pneumococcal Disease; using flucloxacillin for the treatment of Staphylococcus aureus BSI, due to several
administrations per day, while ceftriaxone is only one, without renal adjustment).
. Observability of the recommended behavior - According to the professionals in my hospital, they have doubts
about the benefits of using antibiotics appropriately (p.e using penicillin over ceftriaxone for the treatment of
Streptococcus spp. may not reach a desirable effect, and carries more perceived risk, as manipulation of the
catheter several times a day).

111
 - Factors relating to the individual healthcare professional:
. Existing domain knowledge - According to the professionals in my hospital, they do not have enough pre-existing
knowledge concerning appropriate antibiotic use (p.e believing piperacillin/tazobactam has a wider spectrum
than ertapenem due to Pseudomonas coverage; believing that the coverage of Enterococcus faecalis is similar
within the carbapenems).
. Awareness and familiarity with the recommendation - not knowing the species of high risk of induction of AmpC
with risk of clinical failure with the use of certain antibiotics. Not knowing where to consult (reliably) the renal
adjustment for the common antibiotics prescribed.
. Agreement with the recommendation and Expected outcome – Some professionals in my hospital do not agree
with some recommendations, as they do not believe in desired outcomes (p.e In my hospital, at Hospitalization
at Home is possible to administer ceftazidime/avibactam, but not ceftazidime; I remembered a particular case,
where the prescribing doctors believed that transferring the patient to Hospitalization at Home (receiving
ceftazidime/avibactam) would be better than staying at the hospital receiving ceftazidime – in this case, the
individual advantages of being discharged from the hospital would, in their point of view, outweigh the
advantage of applying antimicrobial stewardship principals). Also, some physicians manifested lack of confidence
on my recommendations because I was a resident, and their ID consultant never discussed some of these themes
with them.
. Intention and motivation – According to the professionals in my hospital, there is a lack of motivation because
short spectrum antibiotics are not more effective than wider spectrum (and the ultimate goal is to reach clinical
cure, that is equal in both).
. Emotions – Fear of using a short spectrum antibiotic in immunocompromised patients, or with severe
comorbidities, or severely ill.

- Patient factors:
. Patient preferences/needs - According to the professionals in my hospital, patient’s preferences (real or
perceived) include a once-daily endovenous treatment rather than several times a day, because it may have
fewer adverse effects and facilitate the transit to OPAT.
. Patient motivation - Patients may not be motivated to adhere to several daily doses of oral treatment after
discharge.

- Professional interactions:
. Communication and influence - Opinions and communication among professionals may hinder adherence
(belief that Internal Medicine physicians should be able to prescribe antibiotics without the support of an ID
consultant; reluctancy to discuss some “simple” patients, as it can be perceived as lack of knowledge; reluctancy
to change the therapy previously decided by an older/influential college; peer pressure to not be so thorough).

Considering the setting of my hospital, the following facilitators for this project were found:
. Interest from Internal Medicine physicians to provide the best therapy to the patient: it is a daily habit (Monday
to Friday) to revise the pending results of microbiological cultures and AST, as well as the patient’s
pharmacological prescription - there is a daily opportunity to reflect on the adequacy of therapy.
. Easy communication between Internal Medicine and ID and Clinical Microbiology physicians. Every clinician
(even residents) has a small cellphone, and we can call directly each other by dialing 4 numbers; the list of
contacts is easily available. Both ID and Clinical Microbiology physicians are very open to discuss antibiotic
therapy. Additionally, each Internal Medicine ward has an ID consultant, with whom discuss frequently the in-
patients (antibiotic therapy, differential diagnosis or management of patients), and most wards have a close
relationship with the respective ID consultant.
. Selective Reporting of AST Results – it is well documented in the literature (and also within our institution:
Pereira M, Poças G, Alves V, Impact of urinary selective antibiogram in primary care, J Bras Patol Med Lab. 2021;
57: 1-5) that selective reporting can influence the prescribing practice, in favor to shorter spectrum antibiotics.
In my institution, clinicians tend to confirm with ID or Clinical Microbiology the susceptibility of an antibiotic not
reported in the antibiogram (and that they were using/ wish to use). This creates an opportunity for discussion
of the targeted treatment, and usually Internal Medicine doctors accept the suggestions of ID or Clinical
Microbiology.
. Pharmacovigilance: pharmacists can contact the prescribing doctor directly (on the small cellphone) and discuss
the need to alter the dosage.

112
. Support from the ID and Clinical Microbiology Department for this project leaded my me. The Head of
Department of Clinical Microbiology provided the microbiological results. All ID physicians of my department
were willing to help in the intervention, and facilitated it (see below).
. Special area of interest – as residents are not a part of the AMS team in my hospital, and are not used to discuss
antibiotic therapy choices with other clinicians in the absence of an ID specialist, it could be difficult to justify
why am I trying to discuss it. However, the fact that I have done some studies in my institution concerning
antibiotic use (and some physicians are aware), and I am currently doing this AMS certificate justifies this special
area of interest of mine, and renders (some) clinicians more open to discussion.
. National Health Service – absence of economic restrains in the choice of antibiotics (For example, in India,
patients frequently received meropenem instead of ertapenem, because it is cheaper for the patient).

Stage 4 Development and execution of the stewardship intervention

The intervention was planned to occur between March and August 2023.

1. Deciding the team: Together with my tutor, we have decided that, besides including an ID and Clinical
Microbiologist (included both Heads of Department), our Stewardship team should include a Clinical Pharmacist
and an Internal Medicine physician. The Clinical Pharmacist included was part of the Infection Prevention and
Control group, and has a special interest in Antimicrobial Stewardship. The Internal Medicine physician was
thought to be essential to understand better the doubts some clinicians may have about antibiotic prescription
and how to properly address them (very helpful in understanding better some barriers). Also, we believed that
having a representative of the Medical Department in the intervention would help us to avoid the distance that
might happen in these interventions “from us to them”. The Internal Medicine physician chosen worked
previously in the Intermediate Care Unit (a somewhat closed setting) and was currently working as a chief of
ward of Internal Medicine. Besides having a very good relationship with the vast majority of the physicians
(specialists and residents), her recommendations are looked up to.

2. Understanding (more) barriers (previously explained) - the main understanding of the barriers came from the
discussions with the Internal Medicine member.

3. Decide the type of interventions that would be useful and, most importantly, well received by the Internal
Medicine clinicians - March, April and May 2023
As the interviews (to understand the barriers) did involve clinical case discussions, and, in the Internal Medicine
member’s opinion, were very useful to understand common mistakes and expose areas where clinicians might
benefit from additional support and education, we decided it would be useful to use the same model and present
it to the clinicians in Educational Sessions.
Also, we agreed that the lack of specific knowledge was an important barrier that would highly benefit from a
direct interview by developing (completed) Guidelines that could support the clinician’s decision making.
Importantly, the possibility to consult this document without a direct contact with the ID team would, perhaps,
be more accepted, i.e, would not show vulnerability by evidencing the need to consult.

4. Perform the intervention – Educational sessions - between May and August 2023
These educational sessions addressed, mainly, the (1) factors relating to the individual healthcare professional,
and (2) professional interactions.

Specific clinical cases were presented has they provided room for specific questions/discussions and to endorse
some key learning points – some examples:
- Different coverage of Enterococci by carbapenems, as one patient received ertapenem for the treatment of an
E.faecalis BSI (due to other co-infections), and most clinicians did not know it was not receiving adequate therapy.
- Difference between the antimicrobial susceptibility of E.faecalis and E.faecium and which treatment options
would be adequate before the AST report (ex. Its not adequate to initiate Vancomycin or Daptomycin after an
identification of E.faecalis; its not adequate to maintain an amino/ureidopenicillin after the identification of an
E.faecium – guarantee the best treatment timely).
- Different coverage of Gram-negative bacilli by carbapenems (ex. Its not adequate to maintain meropenem for
the treatment of a monobacterial BSI by a Serratia marcescens) as several patients received meropenem or
imipenem for the treatment of agents non-Pseudomonas, instead of ertapenem.

113
- Possibility of using piperacillin/tazobactam for the treatment of ESBL-producing organisms. In Portugal, the
most common ESBL is CTX-M, which maintains susceptibility to piperacillin/tazobactam, and it is informed in the
AST reports in my hospital. We discussed the clinical cases where it could be adequate to use
piperacillin/tazobactam, and where we should opt for carbapenems.

Each of these educational sessions was presented to small groups (specialists and residents of a specific ward,
with a maximum of 10 people), in order to increase the possibility/openness for discussion. To improve the level
of trust of the clinicians in (my) proposed discussion, I presented the clinical cases in the presence and with
endorsement of the ID consultant of each Internal Medicine ward (in whom they had a close relationship).
In general, both the ID consultants and the Internal Medicine team were very enthusiastic about this type of
initiative, however, unfortunately, one chief of ward refused the presentation of these sessions in his ward.

5. Understanding (more) barriers: during and after these educational sessions, some clinicians (mostly residents
that felt very open to expressing their doubts to me, but also some specialists with whom I already had a close
relationship) explained to me more barriers to the adherence of de-escalation techniques.

6. Decide the type of interventions that would be useful and, most importantly, well received by the Internal
Medicine clinicians: in some specific Educational Sessions, some Internal Medicine doctors expressed their
curiosity on receiving feedback for their prescribing adequacy, so the team agreed it would be useful to perform
an audit and feedback intervention.

7. Perform the intervention – development of local guideline – between September 2023 and February 2024
To address the (3) guideline factors, a diagram of the targeted treatment of BSI of Pseudomonas aeruginosa was
developed by the whole team (see Supplementary Material). It consists on a 5-page document for consultation
– page 1 describes, schematically, the spectrum of antibiotics with Pseudomonas coverage, and which to use,
preferentially, in case of sensibility (ex. ceftazidime over piperacillin/tazobactam), and why. The second page
describes the dosage and renal adjustment of each antibiotic with Pseudomonas coverage, and was endorsed by
the Clinical Pharmacist. The third page illustrates the efficacy of each antibiotic according to the source of
infection (penetration, and the pertinence of associating other antibiotics, such as dual therapy in endocarditis,
or anaerobic coverage in case of abdominal source), and the duration of treatment according to the type of
infection. The fourth page explains the evidence regarding the short/long regimens and the possibility of
switching antibiotic therapy from endovenous to oral. This document is currently under evaluation by the Quality
Department of the Hospital, so it can be published as an official document.

------
Although the intervention was planned to occur during 6 months, it was not possible to conclude it during this
period, due to a high workload, and because I was not physically working in the ID department. Between January
and June, although working in my hospital, I was assigned to the intensive care unit (distinct Department), with
a high workload, and strict schedule to perform the educational lectures. I managed to do some of them in my
free days. In July I had an opportunity to do an Observership in a ESCMID Collaborative Centre for three weeks.
Since September, I have been completing my scheduled internships abroad, as defined at the beginning of this
program (India between September and November 2023, and Spain between January and April 2024).

More interventions planned:

8. Perform the intervention – more Educational Sessions
9. Perform the intervention – development and presentation of the guidelines of the targeted treatment for
other agents
10. Perform the intervention – audit and feedback session

Stage 5 Evaluation of the stewardship intervention

The evaluation of the intervention was planned to occur between September 2023 and February 2024 (the same
months to avoid seasonal confounders), using the same measures.
We decided not to perform an evaluation of the intervention as this latter was extended, and there could be
confounders of evaluation during the intervention, and not after (the initial design of this project). Also, it is
possible that the strength of the intervention could be improved with more time/availability, and more

114
educational sessions (that some clinicians expressed it would be useful). Unfortunately, this is only possible after
May 2024 (when I return physically to my hospital).
We plan to perform an evaluation between September 2024 and February 2025.

Conclusions

I conclude this report with a bittersweet taste, as I would have liked to have finished the stage 5, but, importantly,
proud of the work developed by the team.

I understood that the stewardship problems that we might think exist, probably do exist, but we need to
concretize them with a correct measurement. In my case, I confess I thought that the prescribing adequacy would
be, sometimes, ever less than what was evidenced. In the end, although there is clear room for improvement
and for a stewardship intervention, I was positively surprised while revising some electronic files by the adequacy
of some prescriptions.
While reflecting on the barriers, I felt more connected to the prescribers. I might have, unfairly, attributed some
prescribing patterns as a lack of care, but I now recognize that these problems often steam from complex
underlying factors that may not be immediately apparent. I believe an important part of this project was realizing
this, and explain my new point of view to the Head Chief of Department of ID and Clinical Microbiology, who
somewhat shared this misconception.
I can liken this learning process to the Dunning-Kruger graph. At times, doubts can arise from limited knowledge,
confidence from an incomplete understanding, and numerous questions from the complexity of the subject. One
of the most important aspects that I have learned is that we all may traverse the various phases of this graph at
some point, even on the same subject. I believe that is both humbling and empowering, and discovering that in
a collaborative teamwork has a special significance.

During the intervention phase, I had very different thoughts and emotions. As I started – and finish – this Program
as a resident, I cannot help reflecting on the impact of the AMS certificate on my personal and professional
growth. Although I have learned technical and theorical knowledge, I have learned a bit more than expected into
interpersonal relationships. The (growing) recognition among some colleagues, while striving to conquer trust
with some others was a challenging duality. If you do not have by your side the right team – or mentor – you can
get easily stuck, being harder than imagined to advance in times of uncertainty. In fact, being a resident is an
ongoing process of self-discovery, where we develop our foundation principals for critical thinking (and
prescribing). However, it is also dominated, in my opinion, by a feeling of being overwhelmed. This was, perhaps,
one of the most significant challenges of this project (that influenced negatively): besides feeling overwhelmed
by workload, and struggling to maintain motivation, I sensed a persistent guilt for not achieving more. Together
with doubts about the ability to successfully implement an intervention project, I ended up prolonging the
implementation phase. However, it was on my tutor that I found the drive to continue - while it may sound cliché,
the idea that the journey matters more than the destination felt very fitting in these moments.

Indeed, although not measured, I believe that we managed to have some impact in my institution. I could share
a very impacting moment for me: the head chief of Intermediate Care Unit was very surprised by the clinical case
that I presented in one of the Educational Sessions, and by the different opinions of their staff on the best
approach for the patient. While I was presenting, let’s say a little bit of discussion came up, and I had the feeling
it would be difficult to get to those clinicians. Later that day, while I was working on the Emergency Room, we
met again to evaluate an unstable patient. After the patient was stabilized she said “Oh, and I didn’t say anything
before, but I think we all learned a lot today. Thank you. Can you come more often?”. Also, one resident thanked
me “I understood more of antibiotics with you in 15 minutes, than I did in the whole University”. Some also said
“I’m sorry I missed the session, I’ve heard my colleagues commenting it was very good”. I believe these comments
illustrate the usefulness of the intervention and keep us motivated to continue performing more, and, of course,
evaluate its impact and share our finding with the prescribing doctors.

Finally, a notable positive aspect of this process was the opportunity to expand my horizons, and getting to know
people you look up to. When we are little, we look up to many people, and as we get older, such admiration
tends to diminish. Sometimes we still hold admiration for others, but it feels distant and intangible, as if we could
never be part of their success. However, through this Program, I have met many inspiring people that I truly look
up to, and with whom I felt very comfortable walking alongside. I could never thank them enough.

115
 Appropriateness of
Microorganism Number of episodes Appropriateness of drug Appropriateness of drug spectrum considering the Main antibiotic used
choice for the choice for the source of AST (and other agents if
microorganism (efficacy) infection (efficacy) explicit)
All 127 97% 95% 54% ceftriaxone (46/127)
Enterobacterales 58 98% 97% 53%
Escherichia coli 35 100% 97% 54% ceftriaxone (21/35)
Enterobacter cloacae 2 100% 100% 50% ertapenem (2/2)
Klebsiella oxytoca 1 100% 100% 100% ceftriaxone (1/1)
Klebsiella pneumoniae 13 100% 100% 53% ceftriaxone (5/13)
Klebsiella aerogenes 1 0% 0% 0% piperacillin/tazobactam (1/1)
Morganella morganii 1 100% 100% 100% ertapenem (1/1)
Proteus mirabilis 3 100% 100% 33% ceftriaxone (2/3)
Serratia marcenscens 2 100% 100% 50% ceftriaxone; piperacillin/tazobactam
Enterococcus spp. 14 93% 86% 71%
Enterococcus faecalis 10 90% 90% 70% ampicillin (4/10)
Enterococcus faecium 4 100% 75% 75% vancomycin (3/4)
Pseudomonas aeruginosa 7 100% 100% 14% piperacillin/tazobactam (6/7)
Staphylococcus spp. 34 100% 97% 68%
Staphylococcus aureus meticillin-resistant 2 100% 100% 100% vancomycin (2/2)
Staphylococcus aureus meticillin-sensible 14 100% 100% 50% flucloxacillin (7/14)
Staphylococcus epidermidis 8 100% 88% 63% vancomycin (4/8)
Staphylococcus hominis 10 100% 100% 80% vancomycin (4/10)
Streptococcus spp. 13 92% 85% 23%
Streptococcus agalactiae 1 100% 100% 100% amoxicillin (1/1)
Streptococcus bovis 1 100% 100% 0% ceftriaxone (1/1)
Streptococcus anginosus group 2 100% 50% 0% ceftriaxone (2/2)
Streptococcus mitis group 1 100% 100% 100% piperacillin/tazobactam (1/1)
Streptococcus pneumoniae 8 88% 88% 13% ceftriaxone (4/8)

Table 1: Appropriateness of drug choice for the microorganism (efficacy), appropriateness of drug choice for the source of infection (efficacy), and appropriateness of spectrum
considering the antimicrobial susceptibility testing, and other agents (if explicit) – results from the retrospective audit between September 2022 and February 2023.

116
 Antibiotic DDD/1000 days
Sodic/Potassium Benzilpenicillin 4
Amoxicillin (oral) 1
Ampicillin (ev) 8
Amoxicillin/clavulanic acid (ev) 91
Amoxicillin/clavulanic acid (oral) 15
Flucloxacilina (ev) 16
Ceftriaxone (ev) 69
Piperacillin/tazobactam (ev) 92
Meropenem (ev) 21
Imipenem/cilastatin (ev) 4
Ceftazidime/avibactam (ev) 8
Vancomycin (ev) 20
Daptomycin (ev) 4

Table 2: Defined Daily Dose (DDD)/100 bed-days between September 2022 and February 2023

117
 Tratamento dirigido de bacteriemias por Pseudomonas aeruginosa
Atualizado de acordo com a literatura disponível em fevereiro de 2024 [1-4]

1º linha: Ceftazidima ou Cefepime Porque preferimos as cefalosporinas de 3ª e 4ª

geração a piperacilina/tazobactam?
Se resistente
As cefalosporinas apresentam menor cobertura
Piperacilina/tazobactam de anaeróbios (não necessária em infeções
monobacterianas por P.aeruginosa) e a
Se resistente piperacilina/tazobactam é uma opção terapêutica
Como estratégia poupadora eficaz em algumas infeções por agentes
de carbapenemos. produtores de ESBL (que hidrolisam as
Ciprofloxacina ou Levofloxacina cefalosporinas).
Se resistente

Meropenem ou Imipenem
Se alergia à penicilina: Utilizar preferencialmente meropenem
Se resistente em vez de imipenem (exceto se
- A reação cruzada com cefalosporinas infeção polimicrobiana com
de 3ª e 4ª geração, aztreonam e Ceftolozane/tazobactam necessidade de cobertura de
carbapenemos é extremamente baixa, Enterococcus spp., em que o único
pelo que são opções terapêuticas Se resistente
carbapenemo validado é o imipenem).
geralmente, seguras.
Ceftazidima/avibactam
- Em caso de dúvida, ou reação grave,
utilizar antibióticos não beta-lactâmicos Se resistente Apenas disponível no nosso
como quinolonas ou aminoglicosídeos hospital, de momento, por
(amicacina – apenas em monoterapia Cefiderocol AUE.
para ITU, mesmo com bacteriémia, sem ou
choque sético, para durações até 7 dias). Ceftazidima/avibactam + Aztreonam (sinergismo)

Se ausência de outras opções, considerar:

 Combinações com, pelo menos, dois fármacos ativos in vitro: colistina, amicacina, fosfomicina.
118 por AUE.
 Não existem dados clínicos sobre a utilização de imipenem/relebactam ou meropenem/vaborbactam, apesar de haver evidência crescente de eficácia in vitro. Disponíveis no nosso hospital apenas
Tabela 1. Posologia endovenosa e oral das opções terapêuticas para o tratamento de bacteriemia por P.aeruginosa [5-6]

Clearance >50 ml/min Clearance 10-50 ml/min Clearance <10 Hemodiálise Notas
ml/min
Ceftazidima (ev)
Cefepime (ev) 2g q8h 30-50: 2g q12h 1g q24h 1g q24h (após diálise)
10-29: 1g q12h
Meropenem (ev) Confirmar sensibilidade com
microbiologia se meningite (MICs
diferentes).[2]
Piperacilina/ 4,5g 6/6h 20-40: 3.375g q6h ou 4,5g <20: 2,25g q 6h ou 2,25g q6h (+ 0,75g após
tazobactam (ev) q8h 4,5g q 12h diálise)
Imipenem (ev) 0,5-1g 6/6h 40-50: 0,5g 6/6h 250-500mg q12h 250-500mg q12h (após
21-40: 350-500mg q8h diálise)
Ceftolozane/ 3g q8h 30-50: 1500mg q8h 1500mg dose única + 3000 mg dose única + 300mg
tazobactam (ev) 15-29: 750mg q8h 300mg 18h q8h
Ou 1,5g 3x/semana, após a
diálise
Ceftazidima/ 2.5g q8h 31-50: 1,25g q8h 1,25g q24h 1,25g q24h
avibactam (ev) 16-30: 1,25g q12h
Cefiderocol (ev) 2g q8h 30-59: 1,5 g q8h <15: 0,75g q12h 0,75g q12h (após diálise)
15-29: 1g q8h
Aztreonam (ev) 2g q6h 10-30: 2g q12h 2g q24h 2g q24h (após diálise) Evitar doses elevadas se
insuficiência hepática
Ciprofloxacina (ev) 400mg q8h 10-30: 200mg q8h <10: 400mg q 24h 400mg q24h (após diálise)
Ciprofloxacina (oral) 750mg q12h 10-30: 500mg q12h <10: 500mg q24h 500mg q24h (após diálise)
Levofloxacina (ev) <20: 750mg e 750mg dose única,
750 mg 24h 20-49: 750mg q48h posteriormente posteriormente 500mg q48h
Levofloxacina (oral)
500mg q48h ou 250 q24h

119
 Tabela 2. Eficácia da terapêutica antimicrobiana de acordo com o foco de infeção (ponto de partida/focalização da bacteriemia) por Pseudomonas aeruginosa. Tempo sugerido de
tratamento e possibilidade de tratamento oral. Legenda: verde – recomendado; amarelo: utilizar com precaução; vermelho – evitar.

Penetração / foco
Urinário Abdominal Respiratório CVC Pele/tecidos moles Indeterminado SNC[7] Endocardite[8] Osteoarticular [9-11]
Antibiótico

Ceftazidima | Cefepime +metronidazol a) Tratamento de

Piperacilina/tazobactam b) combinação:
preferencialmente
Meropenem um beta-lactâmico
e aminoglicosídeo
Imipenem c) (eventualmente
Ceftolozane/tazobactam adicionar
Obrigatório
quinolona).
Ceftazidima/avibactam +metronidazol remover CVC! a)
Cefiderocol Discutir com
equipa de
Aztreonam +metronidazol a) Doenças
Infeciosas e
Ciprofloxacina +metronidazol
Microbiologia
Levofloxacina +metronidazol Clínica.

Amicacina d) b)
Apenas em combinação
Fosfomicina
Colistina Apenas em combinação b)

Duração do tratamentof) 10-14 dias (ver sumário de evidência a favorecer durações mais curtas) e) 6 semanas
Switch endovenoso oralg)

a) Ceftazidima e aztreonam apresentam boa penetração na pele e tecidos moles, mas não cobrem cocos Gram positivo frequentemente implicados (incluindo Staphylococcus spp.).
b) Preferir outros fármacos com melhor penetração no SNC.
c) Evitar imipenem no tratamento de infeções do SNC pelo risco de neurotoxicidade (mas apresenta boa penetração).
d) Pouca penetração sistémica no pulmão (necessita de níveis sérios altos que são frequentemente tóxicos).
e) Pouca evidência para meningite – discutir com especialidade (a maioria dos autores sugerem no mínimo 3 semanas). Abcesso cerebral 6-8 semanas. [12-13]

120
f) Resumo da evidência a sugerir 8-10 dias de tratamento nas bacteriemias por P.aeruginosa, em vez de 14:
Existe um único estudo randomizado controlado sobre o tempo de tratamento de bacteriemias (7 vs 14 dias) que incluiu infeções por P.aeruginosa[15], porém a amostra
de doentes com este patogénio foi reduzida, e não foi possível generalizar os resultados para esta infeção. Atualmente, está a decorrer um estudo randomizado
controlado[14] para esclarecer qual a duração ideal do tempo de tratamento de bacteriemias por P.aeruginosa.
A evidência baseada em estudos retrospetivos sugere que durações mais curtas (8-10[16], 7-11[17], ou 6-10 dias[18]) não apresentam diferença estatisticamente
significativa na mortalidade ou recidiva aos 30 dias que durações superiores (14-17[16], 12-21[17], ou 11-15 dias[18], respetivamente). Estes estudos apenas foram
realizados em doentes com bacteriemia não complicada, com controlo de foco realizado, tendo sido excluídas bacteriemias complicadas por infeção do sistema
nervoso central, osteoarticular ou endovascular/endocardite. De notar que incluíram doentes imunodeprimidos (sob tratamento imunossupressor, infeção VIH estadio
SIDA, neoplasia ativa com ou sem quimioterapia recente, neutropenia, transplantados de órgão sólido e de células hematopiéticas). No entanto, nesta última
população, a percentagem de doentes incluídos foi escassa e, outro estudo demonstrou taxas de recidiva de infeção respiratória e/ou bacteriemia superiores em
durações de tratamento inferiores[19].
Assim, em populações especiais, no doente com bacteriemia complicada ou persistente, o clínico deve considerar a realização de esquemas terapêuticos mais
prolongados, até que exista evidência científica que esclareça a duração ideal do tratamento nesta tipologia de doentes.

g) Resumo da evidência a favorecer o switch endovenoso oral [20-21]:

Ainda não existem dados clínicos que suportem o tratamento oral na endocardite ou infeções do SNC por P.aeruginosa, pelo que não fazemos essa recomendação.
Nos restantes casos, uma meta-análise recente (inclui P.aeruginosa e outros agentes, e inúmeros focos, incluindo bacteriemia, endocardite e osteomielite), a
probabilidade de sucesso a longo prazo é +7% (intervalo de confiança 95%, -1 a 15%), a favorecer o tratamento oral (em detrimento do endovenoso). Em nenhum
estudo randomizado controlado o tratamento oral foi inferior ao tratamento endovenoso exclusivo. Assim, recomendamos que o clínico considere o tratamento oral
sempre que:
 Não haja bacteriémia persistente
 Tenha sido realizado controlo de foco adequado
 O doente esteja estável clínica e hemodinamicamente, não apresentando compromisso da absorção oral
 Haja disponibilidade de fármaco com elevada biodisponibilidade oral (inclui quinolonas), com sensibilidade ao agente demonstrada.

Para além do fármaco adequado, é importante otimizar a duração do tratamento e a forma de administração! Porquê?

É importante considerar durações de tratamento menores (quando existe evidência que suporte igual eficácia) porque estas estão associadas a durações inferiores de
permanência hospitalar e menores taxas de efeitos adversos associados a antibioterapia sistémica. Para além disso, a exposição prolongada a antimicrobianos é um dos
principais promotores de aquisição de resistência antimicrobiana, predispondo o doente a adquirir infeções por agentes multirresistentes/difíceis de tratar.

As vantagens do tratamento oral (vs endovenoso) incluem menor estadia hospitalar e menor incidência de efeitos adversos associados aos acessos endovenosos.121
Referências bibliográficas:
1- Paul M. et al. European Society of Clinical Microbiology and Infectious Diseases (ESCMID) guidelines for the treatment of infections caused by multidrug-resistant Gram-negative bacilli (endorsed
by European society of intensive care medicine). Clin Microbiol Infect. 2022 Apr;28(4):521-547. doi: 10.1016/j.cmi.2021.11.025.
2- The European Committee on Antimicrobial Susceptibility Testing. Breakpoint tables for interpretation of MICs and zone diameters. Version 13.1, 2023. http://www.eucast.org.
3- Tamma PD. Et al. Infectious Diseases Society of America Guidance on the Treatment of AmpC β-Lactamase-Producing Enterobacterales, Carbapenem-Resistant Acinetobacter baumannii, and
Stenotrophomonas maltophilia Infections. Clin Infect Dis. 2022 Jul 6;74(12):2089-2114. doi: 10.1093/cid/ciab1013.
4- Babich T. et al. Ceftazidime, Carbapenems, or Piperacillin-tazobactam as Single Definitive Therapy for Pseudomonas aeruginosa Bloodstream Infection: A Multisite Retrospective Study. Clin Infect
Dis. 2020 May 23;70(11):2270-2280. doi: 10.1093/cid/ciz668.
5 – Gilbert DN, Chamber HF. The Sanford Guide to Antimicrobial Therapy 2024, 54th Edition; 2024
6 – Mensa J., Soriano A. Guía de Terapéutica Antimicrobiana – Mensa; 2024.
7- Haddad N, Carr M, Balian S, Lannin J, Kim Y, Toth C, Jarvis J. The Blood-Brain Barrier and Pharmacokinetic/Pharmacodynamic Optimization of Antibiotics for the Treatment of Central Nervous
System Infections in Adults. Antibiotics (Basel). 2022 Dec 19;11(12):1843. doi: 10.3390/antibiotics11121843.
8- Delgado V. et al. ESC Scientific Document Group, 2023 ESC Guidelines for the management of endocarditis: Developed by the task force on the management of endocarditis of the European Society
of Cardiology (ESC) Endorsed by the European Association for Cardio-Thoracic Surgery (EACTS) and the European Association of Nuclear Medicine (EANM), European Heart Journal, 2023;,
ehad193, https://doi.org/10.1093/eurheartj/ehad193
9- Spellberg B, Lipsky BA. Systemic antibiotic therapy for chronic osteomyelitis in adults. Clin Infect Dis. 2012 Feb 1;54(3):393-407. doi: 10.1093/cid/cir842. Epub 2011 Dec 12. PMID: 22157324; PMCID:
PMC3491855
10- Thabit AK, Fatani DF, Bamakhrama MS, Barnawi OA, Basudan LO, Alhejaili SF. Antibiotic penetration into bone and joints: An updated review. Int J Infect Dis. 2019 Apr;81:128-136. doi:
10.1016/j.ijid.2019.02.005. Epub 2019 Feb 14. PMID: 30772469
11- Rempenault C. et al. Treatment of bone and joint infections by ceftazidime/avibactam and ceftolozane/tazobactam: a cohort study. J Glob Antimicrob Resist. 2021 Jun;25:282-286. doi:
10.1016/j.jgar.2021.04.003. Epub 2021 Apr 30.
12- Tunkel AR. et al. 2017 Infectious Diseases Society of America's Clinical Practice Guidelines for Healthcare-Associated Ventriculitis and Meningitis. Clin Infect Dis. 2017 Mar 15;64(6):e34-e65. doi:
10.1093/cid/ciw861.
13- Bodilsen J. et al. ESCMID Study Group for Infections of the Brain (ESGIB). European society of Clinical Microbiology and Infectious Diseases guidelines on diagnosis and treatment of brain abscess
in children and adults. Clin Microbiol Infect. 2023 Aug 29:S1198-743X(23)00399-3. doi: 10.1016/j.cmi.2023.08.016. Epub ahead of print. PMID: 37648062.
14- Molina J. et al. SHORTEN-2 trial team. Study protocol for a randomized clinical trial to assess 7 versus 14-days of treatment for Pseudomonas aeruginosa bloodstream infections (SHORTEN-2
trial). PLoS One. 2022 Dec 22;17(12):e0277333. doi: 10.1371/journal.pone.0277333.
15- Yahav D. et al. Seven versus fourteen Days of Antibiotic Therapy for uncomplicated Gram-negative Bacteremia: a Non-inferiority Randomized Controlled Trial. Clin Infect Dis 2019 69:1091-8
16- Fabre V. Amoah J, Cosgrove SE, Tamma PD. Antibiotic Therapy for Pseudomonas aeruginosa Bloodstream Infections: How Long Is Long Enough? Clin Infect Dis. 2019 Nov 13;69(11):2011-2014. doi:
10.1093/cid/ciz223.
17- Bae M. et al. Short versus prolonged courses of antimicrobial therapy for patients with uncomplicated Pseudomonas aeruginosa bloodstream infection: a retrospective study. J Antimicrob
Chemother. 2021. https://doi.org/10.1093/jac/dkab358
18- Babich T. et al. Duration of Treatment for Pseudomonas aeruginosa Bacteremia: a Retrospective Study. Infect Dis Ther. 2022 Aug;11(4):1505-1519. doi: 10.1007/s40121-022-00657-1. Epub 2022 May
25.
19- Olearo F. et al. Optimal Treatment Duration of Pseudomonas aeruginosa Infections in Allogeneic Hematopoietic Cell Transplant Recipients. Open Forum Infect Dis. 2020 Jun 23;7(7):ofaa246. doi:
10.1093/ofid/ofaa246.
20- Wald-Dickler N. et al. Oral Is the New IV. Challenging Decades of Blood and Bone Infection Dogma: A Systematic Review. Am J Med. 2022 Mar;135(3):369-379.e1. doi: 10.1016/j.amjmed.2021.10.007.
Epub 2021 Oct 27.
21- Landersdorfer CB, Gwee A, Nation RL. Clinical pharmacological considerations in an early intravenous to oral antibiotic switch: are barriers real or simply perceived? Clin Microbiol Infect. 2023
Sep;29(9):1120-1125. doi: 10.1016/j.cmi.2023.04.009. Epub 2023 Apr 12.

Proposta terapêutica realizada por Susana Ramos Oliveira, Raquel Calisto, Valquíria Alves, Ana Durães, Isabel Neves, Pilar Retamar-Gentil, a fevereiro 2024

122
ESCMID AMS Certificate Program
My Project Report

Name
Susanna Deutch
Institution
Department of Infectious Diseases, Aarhus University Hospital
Country
Denmark
Tutor
Nathan Peiffer
Title of Project
Improved usage of standard treatment packages for community-acquired pneumonia in the
emergency department
Date of submission
11th March 2024

Abstract* (max 300 words)

Background
In Denmark, there is an increasing overall usage of Piperacillin/Tazobactam. Patients with community acquired
pneumonia often recieve Piperacillin/Tazobactam, despite recommendations of Penicillin for mild and moderate
cases. In the Central Denmark Region, regional guidelines are available for the empiric treatment of community
aquired pneumonia and standardized treatment packages (STPs) are available in the electronic medical record.
These are linked to the regional guidelines and precribed antibiotics for the complete treatment course, thereby
supporting the prudent use of antibiotics. However, the usage of STPs needed help for implementation in the
clinic.

Methods:
In the Emergency Department (ED) of Aarhus University Hospital an Antimicrobial Stewardship (AMS) project
was initiated. Barriers and fascilitators were identified. An intervention programme with educational sessions,
technical support and spoken and written feedback was initiated.
A point prevalence survey was done prior to intervention and every 6 months during intervention in the first
year. For patients in the ED with medical conditions, we audited a total of 1116 (435 prior to intervention, 338
after 6 months and 343 after 12 months, respectively) medical records. These data were supported by estimates
from an digital survellance data base. As process indicator the usage of STPs were assesed. The main aim was a
50% adherence to STPs for specific diagnoses. The nominator was all patients with a STP and the denominator
was all patients where empiric therapy was precribed in the case of infection and, in particular, pneumonia.

Results:
The usage of STPs increased overall from 41 % to 67 % and for pneumonia 36% to 53%. A clear reduction of
Piperacillin/Tazobactam was not seen.

Conclusion:
The usage of STPs is a useful quality improvement tool for succesful antibiotic stewardship, but cannot stand
alone. Ongoing feedback, education and advice is crucial for prudent antibiotic use in a busy clinical setting.

Stage 1 Introduction
In Denmark, there is an increasing overall usage of Piperacillin/Tazobactam and selected Gram-negative bacteria
show critical levels of resistance to Piperacillin/tazobactam. Piperacillin/Tazobactam is often prescribed as
emperiric therapy for patients with suspected infection where focus is unknown (sepsis). Furthermore, patients
with suspected community-acquired pneumonia often recieve Piperacillin/Tazobactam, despite
recommendations of Penicillin for the empiric treatment of mild and moderate cases. In the Central Denmark
Region, regional guidelines are available for the empiric treatment of community aquired pneumonia and

123
standardized treatment packages (STPs) are available in the electronic medical record. These are linked to the
regional guidelines and precribes antibiotics for the complete treatment course, thereby supporting the prudent
use of antibiotics. However, the implementation of the usage of STPs may be improved. Table 1 shows selected
diagnosis and their standard treatment packages available in the electronical medical record.

Table 1. Regional guidelines for selected diagnosis of communiy-acquired pneumonia

Tentative diagnosis Standard treatment package
Mild pneumonia 2 days of Benzylpenicillin and 3 days of Phenoxymethylpenicillin
Moderate pneumonia 2 days of Benzylpenicillin and Clarithromycin (i.v. or p.o.) and 5 days of
Phenoxymethylpenicillin
Severe pneumonia 7 days of Piperacillin/tazobactam and Clarithromycin (i.v. or p.o.)
COPD and mild 2 days Benzylpenicillin and 5 days of oral Amoxicillin
pneumonia
Aspiration pneumonia 7 days of Piperacillin/tazobactam
Sepsis Piperacillin/tazobactam

I work as an ID physician in Aarhus University Hospital with app. 800 beds. The Emergency Department have
patients admitted 24/7 and these patients are allocated into different specialties (ortopedics, general surgery,
geriatry and internal medicine). Approximately 150 different doctors from 7 different internal medicine
specialities inclusive Emrgency Medicine treat patients who are allocated to internal medicine. A majority of
these doctors are on duty in the ED seldomly and less than twicw per month. After the initial diagnostic work-up
and treatment, the patients are tranferred to other wards or discarged. This project was located in the Emergency
Department (ED) in collaboration with an ED physician and initiated during the last part of 2022.
In parallel, an Antibiotic stewardship program was initiated in the Central Denmark Region, which comprises the
University Hospital and four regional hospitals. A multidisciplinary ASP working group consisting of 2 clinical
microbiologists, 2 infectious diseases physicians and 2 clinical farmacists equal to a total of one full-time
equivalent was established. I was also part of this working group. In this regional programme, a variation of
different supportive tools were being implemented. One of the tools were establishing antimicrobial
stewardship (AMS)-teams consisting of 2 health care workers from each clinical department (usually a doctor
and a nurse) were selected as part of the organisation as local ambassadeurs and role models.
These AMS-teams were supported by meetings with the members from the working group once or twice during
the first year and the working group fascilitated a network meeting with AMS-team members from similar
specialites, who were put together in groups in order to exchange experiences. In order to make even stronger
interventions for my AMS-project, I had a more frequent contact with the AMS team from the ED, approximately
every month by either mail, phone or personal contact.

Stage 2 Measurement
The usage of STPs was estimated with data from the regional digital surveillance data base. The number of
patients who were treated by a STP was a nominator and the denominator was the number of patients with a
diagnosis leaving the ED and where a STP was available. The percentage was used as a process indicator. The
main aim was a 50% adherence to STPs for specific diagnoses. In order to validate the digital survellance system
a point prevalence study (PPS) was performed prior to intervention and every 6 months by auditing the electronic
patient records. Data was collected from the medical records for a minimum of 19 days each of these months.
We audited a total of 1116 (435 prior to intervention, 338 after 6 months and 343 after 12 months, respectively)
medical records. The usage of STPs overall and specifically for pneumonia was assessed. The usage of specific
STPs for pneumonia was approximately 30 % during the first part of 2021, as shown in Figure 1. A legal permit
was recieved for data collection and in November 2022 the first PPS was performed.
Furthermore, the usage of Piperacillin/tazobactam was estimated with data from the regional data base. Figure
3 shows the number of prescriptions of Piperacillin/tazobactam pr. 100 bed days from January 2021 to February
2024. These data were extracted from the digital survaillance data base.

124
Stage 3 Barriers and facilitators
Based on the COM-B behavioural change model, the Capabilities (C), Opportunities (O) and Motivations (M) were
assessed by conducting a survey. I selected a few questions from a survey which was executed in 2019 amongst
healthcare workers in EU/EEA countries. The aim of this survey was to enable prudent behaviour on antibiotic
use. The selected questions were sent via an online tool (Survey eXact) to doctors, who were on duty in the ED.
Furthermore the doctors were asked about potential barriers for prudent antibiotic use in the ER: ”What is in
your oppinion the greatest challenge in prudent antibiotic use?”.
A list of 240 doctors were contacted by e-mail. This list completed ED specialists, specialists from Internal
Medicine, Geriatry and junior doctors either employed in the ED or collaborating departments, including free
lance employees.

Figure 2. The results of the survey

Results for the question: ”I have sufficient knowledge about how to use antibiotics appropriately for my current
practice.”

125
A total of 49 doctors responded to the questionnaire (43% junior doctors, 49 % senior doctors). The overall results
of this survey show that 78% agree or strongly agree that they have sufficient knowledge. In contrast, only 68 %
feel supported to not prescribe antibiotics when they are not needed. These results are similar to the results of
the Europien survey.
These results were supplemented by written answers, as shown in Table 2: ”What is in your oppinion the greatest
challenge in prudent antibiotic use?”.

Table 2 Sources of behaviour and results of questionnaire

Sources of behaviour Citations from questionnaire Interventions
Capabilities ”No acces to Mecillinam.” E-learning (training)
When the stardard prescription Teaching sessions
package doesn’t fit the patient”. Easy acces to STPs and regional guidelines.”-
”Doctors don’t follow guidelines.” Potential facilitator

Motivations ”A lack of evidence of the usage of Meatings with doctors, where the motivation
STPs” factors are being adressed.
”Fear” Posters
”The need to to something”
”Sometimes patients are pressing for
treatment”
”Just to be sure”
Opportunities ”everybody is prescribing Tazocin Teaching sessions of nurses and doctors.
(broadspectred) antibiotics, regarded Junior doctors and new doctors in the ED are
as the ”house wine” instead of introduces to the STPs and regional
relating it to the actual case and guidelines.
whether prudent antibiotics could be Developing a pamflet.
used instead.”
Diagnostic examinations and
treatment are executed in an inflated
pace, where observation-time is
lacking”

126
 ”The need for action in order to keep
the flow, where we don’t have time
for observation.
”A large personell turnover”.
”Disagreements between diffentent
specialties”
”A lack of diagnostic tests (blood
cultures, urinary cultures etc.before
adminitration of AB makes it hard to
de-escalate”

The AMS-team in the ED was interviewed in January 2023. Different barriers were described in this interview and
informal discussions with several doctors during the feedback. Especially senior doctors were prone to prescribe
from habit, many doctors felt safer using broadspectered antibiotics and some prescribied from different
conseptions which were different from the guidelines, for example regarding a severe pneumonia. Lack of time
was described as a major barrier. These interviews revealed similar barriers as shown in Table 3. Potential
facilitators were described as easy acces to guideines and e-learning program and the fact that junior and senior
doctors also were influenced by the good habits from other departments with a stronger AS profile .

Stage 4 Development and execution of the stewardship intervention

As shown in Table 3, I planned various interventions in order to meet the different barriers found in the survey
and the interviews. The interventions are shown in a timeline in Figure 1. The black lines show the the major
interventions etc. of the regional programme and the blue lines show the interventions of the local project in the
ED (my AMS-project). At the kick-off meeting in October 2022 AMS-teams from the whole region were
introduced to antibiotic stewardship (the why, the what and the how) and all the implementation tools as
described, inclusive the STPs. The participants had group-work sessions where their local interventions and
barriers were being discussed. An e-learning session of prudent antibiotic usage was available initially, but it was
not updated for at least 10 years. The regional program initiated a review and an update, which was available in
October 2023. The session contained 3 cases with interactive questions and would be completed in
approximately 45 minutes. Furthermore, technical support regarding STPs and the digital surveillance data base,
developing a a website with access to literature and teaching material, contact information, ongoing e-mail
correspondance and meetings with members of the regional working group were used as interventions and
supportive tools.
In the ED project audit was planned every 6 months and one day every two weeks feed-back was given to the
doctors in written or personal contact. Furthermore a poster was developed and teacing sessions were executed.

Results:
A total of 1116 (435, 338 and 343, respectively) patients in the ED were evaluated. At least one third of these
patients recieved antibiotics (37-46%).
The usage of STPs increased overall from 41 % to 67 % and for pneumonia 36% to 53%. These results are
comparable to the data from the digital data base as shown in Figure 1. The target of 50% was reached.
Moreover, a total of 18 patients were prescribed Piperacillin/Tazobactam despite having a CURB score of 2 or
less.
An overall reduction of Piperacillin/Tazobactam was not seen (data not shown) in the audit, but as shown in
Figure 3, a trend of increasing usage of Piperacillin/Tazobactam during 2021 and the first part of 2022, and since
the last part of 2023 a stagnation is seen.

Table 2
November 2022 April 2023 November 2023
Number of patients 435 338 343
Patients on antibiotics 201 146 126
Patients with STP/ Available STP 65/160 (41%) 63/110 (57%) 65/92 (67%)
Patients with pneumonia and STP/Available 15/42 (36%) 19/39 (49%) 16/39 (53%)
STP for pneumonia

127
Prescribed Piperacillin/Tazobactam 11 7
with CURB<3

Stage 5 Evaluation of the stewardship intervention

This report describes a small intervention project within a larger regional program, which was initiated almost
simultaniously, thus evaluating the effect of each intervention seperately is very diffucult.
Overall, the interventions in the ED were on a more detailed and personal level and the regional progam relied
mainly on building up an organisation with the supportive tools.
The data from the ED showed increasing usage of STPs and a stagnation of the usage of Piperacillin/Tazobactam.
We experienced that the succes depended of on-going feed-back, teaching-sessions and audits.
The APEASE criterias were used to evaluate the interventions. Acceptability was usually high. In a busy setting,
teaching sessions are grately appreciated. For best results and acceptance, the personal contact and feed-back
was the most effective.
Practicability was also high with easy acces to STPs etc. and some doctors even mentioned that the STPs made it
easier to start treatment. The doctors were forced to or nudged to treat by specific indications. In contrary, the
practicability of using personal feed-back was a problem and turned out to be the most challenging intervention,
since I personally had too little time planned for this. Doing a collaboration was an ED doctor in the AMS team
helped. Moreover, the ED doctor was the consultant responsible for the junior doctors education, resulting in,
for example, a teaching session at start of each internship. The e-learning program was a very practical
intervention, all though critics thought it being to time-consuming (app. 45 minutes).
Affordability was an obvious challange. In the regional program a total of 1 FTE was given pr. year aiming to cover
at least 43 different departments on 5 different hospitals. No ekstra time nor salery was given to the project in
the ED.
The ED is the first place in the hospital where many patients are admitted. I would hypothesize that the antibiotic
treatment that is initiated in th ED often continous during the admission for a majority of patients. Therefore,
the impact of this treatment to the total antibiotic usage is essential.
In contrast, doctors on duty in the ED with good prescribtion habits may be important fascilitators, since they
often also work in other departments. as many there was probably synergistic effect from the departments of
Internal Medicine to the ED, since many of the doctors were on-duty in the ED, thus assuming these doctors had
good prescrition habits. Economic support to antibiotic stewardship programmes depends of regional budgets
and policies, thus wil this infuence equity. And, even in Denmark, different antibiotic guidelines are used. The
latest national strategy was publiced i 2012 and a new strategy for antibiotic stewardship is pending. Therefor,

128
the hope is to have a more homegenic approach and preferably a prioritation of more ressources allocated to
antibiotic stewardship.

Conclusion:
The usage of STPs was a useful as a process indicator and an intervention tool, all though a reduction of the total
amount of antibiotics was not realistic within this timeframe. The main target of at least 50% usage of STPs for
specific diagnoses was reached. A target of more than 60% is not realistic since a significant number of the
patients have conditions such as immunodefiencies, allergies and kidney failure where the STPs are not
applicable. Moreover, some of the patients who were not treated with a STP, had the right treatment according
to the regional guidelines. Overall, the usage of STPs is a useful digitally based supportive tool for succesful
antibiotic stewardship, but cannot stand alone. Ongoing feedback, education and advice is crucial for prudent
antibiotic use in a busy clinical setting. The most important barriers in capability was lack of skills in the usage of
STPs. Here on-going teaching and training may be the most effecient intervention. Lack of motivation do to fear
or bad habits is more challencing to address as is the lack of opportunities in a busy clinical setting.
I have learned that it is important to chose a very specific target for implementation in order to have succes, but
it is crcial to make ”friends”. Therefore, by using the usage of STPs as objective, this agenda supplemented the
more important feed-back and discussions, thus adressing potential barriers and building up a network. This
potential succes depended of my personal skills for example the ability for diplomacy and my medical and
technical knowledge. By using ”pneumonia” as objective had limitations. A small part of the patients who were
treated with antibiotics had pneumonia. Therefore, other issues should be adressed in the ED: the usage of dip-
stix, changing the regional guidelines for for example sepsis away from Piperacillin/tazobactam, shortening the
treatment in the STPs for sepsis. Last but not least, this project shows the need for more bed-side consultations,
especially in a busy setting as the ED where approximately half of the patients recieve antibiotics.
Personally, the greatest challenge was prioritizing sufficient time for interventions in my AMS project, when I
also was part of the regional programme and, as mentioned, seperately evaluating different interventions was
difficult. Nevertheless, these two years have given me lots of knowldge and courage to continue working with
antibiotic stewardship in my daily work.

129
 ESCMID AMS Certificate Program
My Project Report

Name
Trine Langfeldt Hagen
Institution
Aalborg University Hospital Thisted
Country
Denmark
Tutor
Nikolaos Spernovalisis
Title of Project
Reducing the use of Piperacillin/Tazobactam for common infections – a Danish quality improvement study

Date of submission
March 6th 2024

Abstract
Background
During a 4-year period we noticed a doubling of Piperacillin/Tazobactam (PIT) consumption in our hospital,
which was not supported by the local antibiotic guideline. This quality improvement study aimed to study
prescription behavior. We hypothesized that an intervention might increase guideline adherence, thus
decreasing the use of PIT as first line treatment for common infections.

Methods
Baseline data on PIT consumption were collected from April 2022 to March 2023 as defined daily doses
(DDD)/100 bed-days.
We collected baseline data on guideline adherence from January to March 2023 and performed an audit by
reviewing patient charts.
A multifaceted intervention was carried out from April 1st to December 31st, 2023, consisting of teaching
sessions, feedback to prescribers, discouraging use of urinary dipsticks, creating pocket guidelines, and revision
of the antibiotic guideline.
Inclusion criteria: Immuno-competent patients ≥ 18 years being prescribed PIT for community-acquired
infections. Patients with immunosuppression or at risk of hospital-acquired infection were excluded.
We compared the number of PIT prescriptions between baseline and intervention by using unpaired t-tests.
Change in adherence to guidelines was tested using the Chi-squared test.
The primary outcome was reduction in PIT use and secondary outcome was increased guideline adherence.

Results
Three-hundred and sixty-nine patients were included. During the intervention, we observed a 26% decrease in
PIT use (DDD/100 bed days) and 41% reduction in the number prescribed PIT doses/week.
Correctly prescribed PIT doses increased from 49.6% to 53.8% after the intervention. The correct prescriptions
were differentiated between senior (75%), junior (57.7%) and locum doctors (43.5%).

Conclusion
By using a multifaceted intervention strategy to encourage prescribers to follow local guidelines, we observed a
substantial reduction in the use of PIT for community-acquired infections in immuno-competent patients.
Forming a formal antimicrobial stewardship team may play an important role in maintaining vigilance towards
antibiotic prescribing behavior.

130
 ESCMID AMS Certificate Program
My Project Report

Stage 1 Introduction: description of the setting and the AMS program in place and identification of the
problem
Background:
Aalborg University Hospital Thisted is a small acute care hospital located in the North-West of Denmark. The
hospital administration is centralized at the tertiary University Hospital in the city of Aalborg, 100 km away. My
department (Acute and Internal Medicine) is the largest department in the hospital, consisting of approx. 40
beds for in-patients (internal medicine, neurorehabilitation and palliative care) and 12.000 yearly contacts to
the outpatient clinic.
Currently, the department’s doctor staff consists of 13 senior physicians of different internal/general medicine
specializations and 20-22 junior physicians in various stages of their specialization. The department takes part
in the education of 4th year medical students from Aalborg University.
The regional antimicrobial guideline has been readily available in an online guideline-compilation for > 10 years
and is revised regularly by the Specialist Committee of Infections (ID Physicians, clinical microbiologists,
pharmacologists, pediatricians and representatives for the internal medicine departments from each hospital in
the region), of which I am a participant.
Denmark has a well-developed program of surveillance and reporting microbiological findings in national
databases (DANMAP) and all microbiological results are available in a national electronic system (MiBA) with
direct access through the electronic patient charts. Unfortunately, we do not receive data about local
resistance patterns. We have a regional IPC Unit that provides a vast number of guidelines.
Currently, my hospital has no funding for a formal AMS team, but I and another ID physician supervise and
teach AMS principles to our colleagues, including antimicrobial prescribing, and we often consult with our
colleagues in the Clinical Microbiology Department or Department of Infectious Diseases, both located in
Aalborg. There is the opportunity of participating in a weekly online multi-disciplinary conference on
severe/complicated infections. In the ID department in Aalborg there is some AMS-funding (0,5 FTE for 560
beds) for doing “antibiotic rounds” on selected departments, but still no designated AMS team consisting of
both ID physician, medical microbiologist and pharmacists.

My problem:
In my hospital we have seen a significant increase in the use of empirical broad-spectrum antibiotics, especially
Piperacillin/Tazobactam (PIT), for community-acquired infections. This is unfortunately also the case all over
Denmark, and we are now seeing nationwide increasing PIT resistance in the gram-negative bacteria.
Consumption data over 4 years showed a doubling of the percent-vise distribution of PIT from 7,5 to 15% of all
antibiotic prescriptions at our hospital, as well as the highest consumption percent-vise of all the hospitals in
the region. With Denmark being a very low-resistance setting, this development was considered deeply
concerning and an intervention was deemed necessary.

Stage 2 Measurement
To ascertain the extent of my problem I did a baseline measurement of the appropriateness of PIT
prescriptions. I was provided with patient data for all PIT prescriptions in my department for a 3-month period
(July 1st-Sept 30th, 2022) with the help from the Clinical Pharmacology Department. Then I did an audit of the
electronic patient charts, including patients according to in- and exclusion criteria mentioned below (defined as
according to the regional antimicrobial guideline):
• Included population: adult (18 years) immuno-competent patients with community-acquired
infections being admitted to the medical service via the emergency department of Aalborg University
Hospital Thisted, who was prescribed PIT empirically
• Exclusion:
o Immuno-suppression defined as: Active chemotherapy, hematologic cancer diagnosis,
immuno-modulatory treatments (i.e Methotrexate, Prednisolone > 15 mg daily for > 3
weeks), other immunosuppressive treatments (i.e. TNF--inhibitors, monoclonal antibodies)
o Possible hospital-acquired infection (in-patient at the hospital > 48 hours at first sign of
infection or hospital-stay/invasive procedures within 14 days prior to admittance),
hemodialysis patients
o Patients transferred from other hospitals with an active prescription of PIT
Appropriateness of prescriptions according to guidelines was measured by defining correct indications for PIT:

131
 ESCMID AMS Certificate Program
My Project Report

• Sepsis of unknown origin (only if aminoglycosides were contraindicated) – qSOFA assessment

• Septic shock
• Severe pneumonia (CURB-65 score 3-5)
• Complicated UTI or urosepsis (only if aminoglycosides were contraindicated)
• Aspiration pneumonia
Qualitative measure: patient populations with guideline non-concordant prescription of PIT/patient population
prescribed PIT.

Quantitative measures: DDD/100 bed-days for PIT and number of prescriptions of PIT for community-acquired
infections in immuno-competent patients.
Results of the baseline data:
Total Adherence to Non-adherence p-value
guidelines to guidelines
Number 110 57 53
Age (years) 77 (12) 76 (14) 77 (9) 0.80
Sex (male) 62% (68) 58% (33) 66% (35) 0.38
Correct dosing according to guidelines 96% (106) 93% (53) 100% (53) 0.049
Length of treatment (days) 3 (1-4) 3 (2-5) 2 (1-4) 0.031
Re-evaluation within 48 hours 83% (82) 84% (42) 82% (40) 0.75
Narrowing in relation to susceptibility 58% (30) 54% (13) 61% (17) 0.63
Narrowing in relation to guideline 62% (53) 69% (27) 55% (26) 0.19
Sepsis 20% (22) 86% (19) 6% (3) <0.001
Septic shock 7% (8) 12% (7) 2% (1) 0.036
Severe pneumonia 19% (21) 95% (20) 2% (1) <0.001
CURB-65-score registered 13% (3) 11% (2) 17% (1) 0.72
Complicated urinary tract infection 12% (13) 16% (9) 8% (4) 0.17
Urosepsis 6% (7) 9% (5) 4% (2) 0.27
COPD exacerbation 5% (5) 2% (1) 8% (4) 0.15
Aspiration pneumonia 2% (2) 2% (1) 2% (1) 0.97
Blood culture performed 89% (98) 88% (50) 91% (48) 0.63
Blood culture growth 16% (16) 16% (8) 16% (8) 0.93
Urine culture performed 75% (83) 74% (42) 77% (41) 0.65
Urine culture growth 39% (34) 40% (18) 38% (16) 0.86
Expectorate culture performed 37% (18) 39% (11) 33% (7) 0.67
Atypical pneumonia PCR performed 13% (6) 19% (5) 5% (1) 0.17
Death at 30 day follow-up 29% (32) 34% (19) 25% (13) 0.28
Readmission due to complications 1% (1) 0% (0) 2% (1) 0.30
Readmission due to treatment failure 8% (7) 9% (4) 7% (3) 0.74
Data are presented as mean (SD) or median (IQR) for continuous measures, and % (n) for categorical measures

The baseline audit showed no significant differences between the adherence and non-adherence group
regarding the demographic data. The mean age was 77 years. In 53% of the cases PIT had not been prescribed
according to guidelines. 83% of prescriptions were re-evaluated within 48 hours and the median treatment
length was 3 days.
Prescriptions were significantly more often correct when the indication was either severe pneumonia (95%) or
sepsis (86%) – both conditions had PIT as first-choice antibiotic according to the guideline.

132
 ESCMID AMS Certificate Program
My Project Report

Most patients had blood cultures (89%), or urine cultures (75%) performed, and there was a high rate of
positive urine cultures (39%) in both groups. The performance of expectorate culture or registration of CURB-
65-score in case of suspected pneumonia was low (<20%).
30-day mortality was 29% with no significant difference between the groups.

Stage 3 Barriers and facilitators

The baseline results were shared with the other physicians in the department in a morning teaching session, where I
also presented the key elements of AMS. My colleagues were very surprised to see the low level of adherence. Junior
and senior doctors attended the session together and we had an open discussion about the difficulties in
prescribing according to guidelines. Since not all the doctor staff were working on this day, I followed up by
sending out an e-mail to the entire staff of physicians including the results of the data, and with encouragement
to reply to the e-mail, if they had any questions or comments. I also included our team of locum doctors that
frequently take shifts at the department. Several responded to the e-mail addressing their concerns, but none
expressed unwillingness to change their prescribing behavior.

After these discussions I identified possible barriers by using the Flottorp framework. Facilitators for being able to
reach a higher degree of compliance by developing an intervention strategy were identified using the COM-B model.

Barriers:
Factors relating to the individual healthcare professional
- Knowledge and skills
o Lack of knowledge about resistance patterns and the spectrum of the individual antimicrobials
o Knowledge about own practice relative to the recommendation – some doctors was not aware
that they were prescribing PIT more often than other colleagues
o Perception: Many doctors believe that urinary tracts infections in the elderly usually presents with
unspecific symptoms and let results of urinary dipstick tests impact their decision making
- Cognitions - expected outcomes and emotions:
o Worries about the antimicrobial “not being enough for their fragile patients”
o Previous experiences with patients that deteriorated in spite of guideline concordant prescribing
- Professional behavior
o Nature of behavior: several doctors commented on their prescribing behavior as a habit they had
acquired while working at other hospitals. “We always prescribe this way”, “The guidelines are
different at my hospital”
- Professional interactions:
o Communication and influence: The junior doctors want to follow guidelines but if a senior doctor
tells them to prescribe differently, they find it difficult to challenge the orders. They also shared
that especially locum doctors had a tendency to overprescribe PIT themselves and to advise the
junior doctors to follow their lead
o Team processes: The junior doctors often felt pressured by the nursing staff to “do something”,
especially at very busy times in the ER
Facilitators:
- Motivation and education: Most of the doctors wanted to improve their prescribing behavior and were
positive toward getting more feedback, even the senior doctors. They were interested in more teaching
sessions about AMS, AMR and antimicrobials in general, naming that they just want the best outcome for
their patients and acknowledging the lack of specific skills in these areas. They named the possibility of
always being able to have consultations on the phone as an incentive for learning to be better prescribers.
The junior doctors suggested having pocket guidelines to help with the management of infections,
especially containing information about how to do microbiological sampling.
- Role models: The junior doctors mentioned the importance of senior doctors being role models when
prescribing antimicrobials.
- Social interactions and persuasion: All the doctors have very good relationships with each other and want
to help each other to give the best patient care. As an established ID physician in the department, I have

133
 ESCMID AMS Certificate Program
My Project Report

gained the respect of my colleagues on a professional level, which was crucial to be able to give feedback
on prescribing behavior. This also includes the team of locum doctors.
- Incentative: Positive feedback and praise as well as a reward for reaching the final goals was identified.

Stage 4 Development and execution of the stewardship intervention

The intervention was tailored by using the results of the baseline data. For example – we rarely used the CURB-65-
score, when assessing severity of pneumonia, and more than one focus of infection was often suspected due to the
results of urinary dipsticks, even if the results did not correlate with clinical findings, thus leading to more broad-
spectrum antibiotic choices.
The intervention consisted of several approaches, some as one-time teaching sessions, others as repeated sessions in
order to handle the knowledge gap of AMS. The junior doctors requested pocket guidelines on the management of
UTI and pneumonia, and these were developed, tested and put into production during late summer/early fall 2023.
The nursing staff were also introduced to these and were interested in having the guidelines too.
As an ongoing intervention I held teaching sessions for all the nursing staff regarding discontinuing the use of the
urinary dipstick, and a plan of the complete removal of the dipstick after a 3-months trial period of trying to reduce
the use was put in place.

The details of the accomplished interventions are as follows:

• Follow-out audit on patients prescribed PIT for community-acquired infections, including data on time to
re-evaluation, de-escalation and length of treatment.
• Post-prescription audit and feedback to the doctors (written or in person when applicable)
• Pocket guidelines for the management of pneumonia and UTI. The UTI pocket guideline has since been added to
the regional electronic UTI guideline
• Discouraging the use of the urinary dipstick in the wards and emergency department
o Restriction on the use – only when prescribed by the doctors
o Teaching session for the nursing staff on all wards and the ER
o Separate teaching session for the doctors
• Quarterly teaching sessions for all newly hired doctors and medical students in the antimicrobial guideline
• Revision of the regional antibiotic guideline and the UTI guideline –discouraging the use of PIT for
most infections and the use of urinary dipstick in general
• Reminder e-mails to all doctors (including locum doctors) when guidelines were revised
Below is a timeline of the interventions, though it really was an ongoing process with more elements added over a
course of 6 months. The primary focus was to address the knowledge gap by providing extensive education.

The data extraction and patient chart audit was repeated for the entire year of 2023, and I was provided with
new consumption data to follow any changes before/during/after the intervention.

134
 ESCMID AMS Certificate Program
My Project Report

The initial aim of the intervention was to see a 30% reduction in the consumption of PIT, a total reduction in the
number of prescriptions by 30% and an increase in appropriateness of the prescriptions by 20%.

Stage 5 Evaluation of the stewardship intervention

The evaluation of the intervention was performed by retrieving sales data restricted to ATC group J01 for the
department from the hospital pharmacy’s sales database measured in DDD. The number of bed days for the
departments were retrieved from the central patient administration system.
I performed an audit, using patient chart review from January 1st to December 31st, 2023, in order to have data
before and after the intervention was started.

Statistics: Unpaired t-tests were used for comparing the number of PIT prescriptions between baseline and
intervention.

369 patients were included. The mean (±standard deviation) PIT use per month was reduced during the
intervention as compared to baseline 10.29±3.17 vs. 13.99 ±2.80 DDD/100 bed days, p=0.013.
The number of prescribed PIT doses/week were reduced during the intervention as compared to baseline
6.05±2.77 vs. 10.23±4.28 doses/week, p<0.001.
Overall, 52.3% of PIT prescriptions were adherent to the guideline. Senior doctors prescribed PIT correctly in
73% (7/26), junior doctors in 55% (121/221), and locum doctors in 44% (54/124) of cases.
Correctly prescribed PIT increased from 49.6% to 53.8% (p=0.44) after the intervention. After the intervention
senior doctors prescribed PIT correctly in 75.0% (12/16), junior doctors in 57.7% (79/137), and locum doctors in
43.5% (37/85) of cases.

Junior and locum doctors were the main prescribers and had the lowest levels of adherence to guidelines, thus
emphasizing the importance of including both groups in the feedback. Most often the locum doctors function
as consultants for the junior doctors on the shifts, thus having a significant impact on the prescribing behavior
of the junior doctors, which was also addressed by the junior doctors during the initial barrier interview.

Conclusions

135
 ESCMID AMS Certificate Program
My Project Report

Implementing changes is very time consuming and requires a continuous follow-up on the interventions and
multiple mini-evaluations using the plan-do-study-act concept in order to succeed. Since extra resources were
not available for this project, I had to do a lot of the work in my spare time and in between other tasks. Even
though I could not be available for my colleagues continually through the project, I was “saved” by the great
relationship with the prescribers, who sought out my advice regularly and showed a willingness to improve
their prescribing behavior. My colleagues still use me and my other ID colleague very much for antimicrobial
advice and we have a lot of conference discussions about the use of antimicrobials. This has led to the
beginning of a change in culture, where we encourage our colleagues to seek advice with “the experts”
whenever doubts arise, in order to facilitate learning, and we are hopeful that this will further improve
prescription behavior.
I had to adapt the project several times. Unanticipated leaves of absence from some colleagues led to the
transfer of many other tasks to my desk, which in turn made it harder to find the time for the project. This
meant more “e-mail leadership” and less one-on-one feedback, so I am relieved to find that the project
succeeded in spite of this.

With the knowledge I have gained by attending the ESCMID AMS Certificate, I will be able to work towards the
formation of a formal stewardship team, albeit without the possibility of having a clinical microbiologist
physically present at my hospital. I have a very good relationship with the physicians at the Clinical
Microbiology Department, the ID Department, the Clinical Pharmacology Department, the hospital pharmacy
and the local pharmacists, and this will be essential in the further strengthening of the stewardship area locally
– and may even inspire my colleagues in other hospitals. The network of colleagues from all over the world has
been an invaluable resource, and I feel very privileged to have attended this high-quality certificate program.

136