Safety Science 49 (2011) 695–708

Contents lists available at ScienceDirect

Safety Science
journal homepage: www.elsevier.com/locate/ssci

A systemic methodology for risk management in healthcare sector

Anna Corinna Cagliano ⇑, Sabrina Grimaldi 1, Carlo Rafele 2
Department of Production Systems and Business Economics, Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy

a r t i c l e i n f o a b s t r a c t

Article history: The recent biomedical, technological, and normative changes have led healthcare organizations to the
Received 27 May 2010 implementation of clinical governance as a way to ensure the best quality of care in an increasingly com-
Received in revised form 9 November 2010 plex environment. Risk management is one of the most relevant aspects of clinical governance and
Accepted 8 January 2011
approaches put forward in literature highlight the necessity to perform comprehensive analyses intended
Available online 18 February 2011
to uncover root causes of adverse events.
Contributing to this field, the present paper applies Reason’s theory of failures to work out a systemic
Keywords:
methodology to study risks impacting not only directly but also indirectly on patients. Also, the steps of
Healthcare
Risk management
such approach are organized around Human Reliability Assessment phases, in order to take into account
Clinical errors the human component of healthcare systems. This framework is able to foster effective decision making
Patient safety about reducing failures and waste and to improve healthcare organizations’ maturity towards risk man-
agement.
The developed methodology is applied to the pharmacy department of a large Italian hospital. An
extensive validation in different healthcare settings is required to fully prove benefits and limitations.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction adopted the concept of clinical governance. Clinical governance

aims to ensure that patients receive the best quality of care. It
In recent years healthcare systems have been involved in a includes systems and processes for monitoring and improving
number of different changes, ranging from technological to norma- services, risk management, clinical audit, clinical effectiveness
tive ones, all asking for increased efficiency. In addition, the bio- programs, staff management, education training and continuous
medical progress in the last decades has contributed to raise the personal development, and the use of information to support
level of organizational complexity in hospitals, which is given by healthcare delivery (Sale, 2005). Among the different aspects of
many different factors, such as multiple professional experiences, clinical governance, risk management is crucial since it addresses
non-uniform management models, patient specificity, surgery the clinical risk impacting on patients. Literature shows that clini-
complexity, reduced inpatient days, and a growing number of cal risk management does not always take a systemic perspective.
healthcare service users due to an increase in average lifetime. Moreover, it does often not rely on the understanding of people
As a result, medicine complexity, driven by innovations in both sci- acting in the investigated processes, nor gives it a valuable support
ence and technology, stresses the need for new managerial models to decision making.
(Bridges, 2006). Thus, this context highlights the necessity to de- This paper operationalizes Reason’s theory of failures by devel-
velop systemic approaches able to detect waste and errors and to oping a methodology to investigate healthcare processes and re-
suggest organizational and/or technological solutions for continu- lated risks impacting either directly or indirectly on patients.
ous improvement. The work provides a systemic approach based on expert knowl-
To this end, following the success of the application of Kaizen edge and able to sustain continuous improvement. With the pur-
principles to the manufacturing sector (Liker, 2004; Liker and pose of explaining how it works, the methodology is applied to
Hoseus, 2008), international healthcare organizations, such as the pharmacy department of a large hospital. However, more case
The Joint Commission on Accreditation of Healthcare Organiza- studies are needed to completely assess the relevance of the frame-
tions and The World Health Organization, have developed and work to the healthcare sector.
The paper is organized as follows: Section 2 highlights the need
for a systemic perspective on healthcare risk and presents Reason’s
⇑ Corresponding author. Tel.: +39 0110907206; fax: +39 0110907299.
theory of latent failures. Section 3 discusses the importance of
E-mail addresses: [email protected] (A.C. Cagliano), sabrina.grimaldi@
errors to clinical risk, as well as the features characterizing a
polito.it (S. Grimaldi), [email protected] (C. Rafele).
1
Tel.: +39 0110907289; fax: +39 0110907299. successful methodology for managing it. The proposed methodol-
2
Tel.: +39 0110907286; fax: +39 0110907299. ogy and its application are presented in Section 4 and Section 5

0925-7535/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ssci.2011.01.006
696 A.C. Cagliano et al. / Safety Science 49 (2011) 695–708

respectively. Benefits and limitations of the approach, together Incomplete

with future research lines, are discussed in Section 6. procedures
Triggers
Institution
Organization
2. Healthcare risk: need for a systemic perspective Responsibility
Profession shifting
Similarly to any other complex system, the complexity of
healthcare systems generates adverse events if not controlled (Vin- Team
Production
cent, 2006). An adverse event may be defined as an unintended in-
pressures Individual
jury or complication resulting in disability, death or prolonged
hospital stay that is caused by healthcare management rather than Technical
by the patient’s underlying disease process (Ross Baker et al.,
Attention
2004). An intrinsic characteristic of medical care is the fact that,
Distractions
whenever it is delivered, patients run the risk to suffer from a dis- Regularity
ease as an unwilling consequence of treatments (Thomas et al., narrowness Deferred
2000). Thus, the probability of errors and adverse events in general Mixed Maintenance
cannot be eliminated in healthcare organizations. However, it can messages
be controlled by the application to risk management phases of a
recursive process of continuous improvement inspired by the Plan, DEFENSES Inadequate Accident
training
Do, Check, Act (PDCA) paradigm (Tonneau, 1997). According to The
Project Management Institute, risk management includes the pro- Fig. 1. Reason’s model (adapted from Barach (2002)).
cesses concerned with risk management planning, identification,
analysis, response, monitoring, and control. The aim is to increase
the probabilities and impacts of positive events and to decrease the (Harms-Ringdahl, 2009). When there are deficiencies in these bar-
probabilities and impacts related to adverse events (Project Man- riers, they are not able to block the unexpected flow of energy and
agement Institute, 2004). Risk management has been adopted to originate an adverse event that may be classified as a ‘‘near miss’’
cover all healthcare risks, both clinical and non clinical ones. (almost an event), an ‘‘incident’’ (event without damage), or an
The present work focuses on clinical risk, which has been de- ‘‘accident’’ (event with damage) according to its severity
fined by different authors. Wilson and Tingle refer to clinical risk (Hollnagel, 2004). Deficiencies are represented by latent and active
as clinical error to be at variance from intended treatment, care, failures (Reason, 2001). On the one hand, latent failures alone are
therapeutic intervention or diagnostic result (Sale, 2005). Kohn not able to cause full-blown symptoms, only if connected to other
et al. (1999) define clinical risk as the probability that a patient factors and under facilitating conditions they originate adverse
is affected by an adverse event voluntarily or involuntarily caused events. On the other hand, active failures represent immediate trig-
by medical treatments. However, clinical risk is not only due to gering events, they are related to people acting in a system and
medical activities directly impacting on patients but it is reliant their detection often implies the identification of an individual
on a larger set of activities and professionals. It can be determined responsibility.
by many factors relating to the system, the environment, and the The existence of both direct and indirect causes for adverse
interplay of individuals operating in the processes connected to events in any social–technological complex system, as highlighted
the delivery of care (Kohn et al., 1999). This research takes such by Reason’s theory, stresses the need for taking a systemic perspec-
broader perspective on clinical risk, including all events that may tive to risk, in order to have a global view on how the interrelations
affect patients’ safety both directly and indirectly. among technical, human, and organizational factors cause or pre-
Within clinical risk, medical errors are particularly important vent negative events. This necessity is even more evident for clin-
since they may occur during multiple hospital processes, from ical risk, since healthcare systems are human intensive and their
therapy prescription, thorough preparation, distribution, and ultimate goal is providing a medical service ensuring the safety
administration (Vincent, 2001). Several studies performed in US, of the entire population.
Australia, New Zealand, and Europe (Davis et al., 2001; Leape
et al., 1991; Vincent et al., 2001; Wilson et al., 1995) reveal that
about fifty percent of adverse events taking place in healthcare sys- 3. Managing clinical risk by working on errors
tems may be prevented. This highlights a strong need for under-
standing the triggering events of medical errors as well as their A systemic perspective is not the only feature characterizing a
correlations in order to decrease the probability of occurrence of successful methodology for managing clinical risk. Preventing risk
these errors by working on all their possible causes. requires to understand how to strengthen those procedural,
The theory of latent failures put forward by Reason is relevant administrative, physical, and individual barriers intercepting and
to this end (Reason, 2002). According to such author, adverse blocking the energy flow responsible for deviations. To this end,
events are seldom determined by a single error, being it either it is useful to work on what constitutes such energy flow, that is,
human or technological, but more often they are the result of a according to Reason’s theory, on errors.
chain of errors and events where the person responsible for the Error taxonomies put forward in literature provide relevant in-
final error is only the last causal link. In other words, adverse sights on how to cope with adverse events occurring in the health-
events are produced by many factors, such as organizational, pro- care sector. Several classifications of errors have been developed
fessional, personal, and technical ones. Reason’s model defines an (Baysari et al., 2008; Baysari et al., 2009; Cosby, 2003; Shorrock,
adverse event as an unexpected release of energy that may be 2002; Wieman and Wieman, 2004). Among them, Predictive Hu-
prevented by erecting barriers between the source of energy and man Error Analysis (PHEA) is one of the most detailed (Embrey,
the person or the object to be protected (Fig. 1). In this situation, 1992; Hollnagel, 1998). PHEA groups error modes according to
the word ‘‘barrier’’ refers to a wide range of preventive/protective the following activities: planning, execution, control, information
measures including protection devices, security systems, working retrieval, communication, and selection. For example, in a health-
procedures, training, supervision, and emergency plans care setting, prescription errors may be related to planning, if the
 A.C. Cagliano et al. / Safety Science 49 (2011) 695–708 697

proper drug is not prescribed, to execution, if physicians’ handwrit- The present work suggests a risk management methodology
ing is not easy to be interpreted, or to communication, if errors oc- integrating all the three error perspectives by Lucas and Reason
cur when transcribing therapies. This scheme suggests that errors into an inductive approach able to perform a comprehensive inves-
usually have roots grounded in different areas, thus assuming mul- tigation of clinical processes in order to identify criticalities (inef-
tiple viewpoints is essential to manage them. fectiveness) and waste (inefficiency).
Our research builds on Lucas (1997) and Reason’s (2001) per-
spectives for approaching error reduction. These authors look at er-
rors by taking into account four dimensions, namely organizational 4. A systemic methodology for clinical risk management
(systemic), individual, technical, and psychological ones. The pres-
ent work focuses only on the first three perspectives since the last 4.1. Overview of the methodology
one is related to merely psychological issues.
The organizational perspective states that error determinants are Lucas and Reason’s perspectives to error reduction allow to de-
to be found within the system at issue. To be more precise, errors scribe the relevant characteristics of our methodology:
are made by people but their behaviours are highly influenced by  The suggested approach complies with the organizational/sys-
the working environment and the organizational processes. Active temic perspective since it first identifies the most critical parts
failures, such as cognitive, skill set (interpretive and procedural), of the entire clinical system as well as those barriers that are
task-based and personal impairment (Cosby, 2003), have to be most vulnerable and prone to cause adverse events. After that,
traced back to latent failures residing for instance in planning the methodology focuses on a specific healthcare process and
and working procedures (e.g. poor shift programming exposing on the analysis of the associated criticalities, in order to identify,
anaesthetists to many consecutive working hours, with the conse- assess, and control risks related to the activities of such process
quent risk of a decreased level of attention; lacking of integration both directly and indirectly impacting on patients.
among informative systems leading to scarce communication  The individual perspective is assured by involving the actors of
among the actors of the healthcare delivery process). the investigated process in every step of the methodology, in
The individual perspective focuses on the characteristics of peo- order to create that background of information necessary to
ple responsible for errors, such as motivation, personality, and handle all risk management phases.
interpersonal relationships. Also, it encourages learning from er-  The technical perspective is addressed since our methodology
rors by discussing them and communicating possible safety prob- supports decision making in order to define the most effective
lems and adverse events (McDonald and Mayer, 2008). Thus, this and efficient organizational and/or technological improvements,
perspective stimulates the emergence of the experience developed according to the maturity of single organizations towards risk.
by people while working in a given system, of the ‘‘big message’’
coming from experts. In other words, such perspective promotes The practical translation of these features was inspired by
expert knowledge elicitation, a formal process of obtaining infor- Human Reliability Assessment (HRA), an approach aimed to identify
mation able to make a person’s knowledge and believes explicit errors and weaknesses by examining both a system and people
(Garthwaite et al., 2005). Expert knowledge elicitation has been working in it. HRA takes a systemic perspective by looking at the hu-
widely applied not only to risk identification but also to risk assess- man contribution to technical and organisational settings (Embrey,
ment being translated into probability distributions (O’Hagan and 2000). In this way, it provides a class of techniques that are very
Oakley, 2004). In a healthcare system, characterized by human- powerful to improve reliability and safety in the healthcare sector.
based activities, managing criticalities by means of an individual The main phases of HRA include data collection, task descrip-
perspective allows to fortify barriers against the flow of energy tion, task simulation, human error identification and analysis,
associated with adverse events. In this case, experts are repre- and human error quantification (Lyons et al., 2004). Following
sented by nurses, physicians, radiographers, pharmacists, recep- them, we propose a methodology structured according to four pro-
tionists, cleaners, porters or other professional figures. gressive steps, namely:
Finally, according to the technical perspective, adverse event
reduction can only be pursued by means of strong automation, be- 1. Context analysis.
cause the technical components of a system are more reliable than 2. Process mapping.
human beings (Lucas, 1997). However, similarly to what happened 3. Risk identification and assessment.
in the manufacturing industry about forty years ago (Janssen et al., 4. Failure modes and waste analysis (FMEA–Waste analysis).
1995; Norman, 1990), healthcare systems require a gradual intro-
duction of technological innovations such as Computer Physician Our methodology integrates and organizes project, risk, and
Order Entry or Electronic Patient Records. Technology is a powerful waste management approaches, thus enabling them to contribute
tool to increase safety in healthcare processes, but without a well to a systemic analysis of risk, which is not possible when they are
structured organization it may make existing working practices applied in isolation. It is important to highlight that a project and
more complicated, resulting in fewer benefits than expected. There- a process have resembling structures, even though a project is tem-
fore, an accurate analysis of processes to understand criticalities and porary and unique while a process is usually ongoing and repetitive
waste is necessary before their reengineering through technology. (Project Management Institute, 2004). As a matter of fact, a project
The relevance of Lucas and Reason’s approach has been wit- may be defined as a set of activities aimed to produce a given out-
nessed by many authors (Harms-Ringdahl, 2009; Henriksen et al., come (Harvard Business School, 2004) and a process is a set of inter-
2008; Paz Barroso and Wilson, 2000; Wiegmann and Shappell, related activities whose goal is transforming inputs to create
2001). However, adopting one of these perspectives alone is not outputs (Johansson et al., 1993). In the case of healthcare processes
so beneficial. Instead, a combination among the first three ones the activities will be directed towards the correct delivery of care to
leads to an effective methodology to clinical risk management. patients. Therefore, the similarity of structure between a project
Thanks to the individual perspective, human resources will make and a process allows us to apply project management tools to the
emerge that sunk information necessary to adopt a systemic stand- investigation of risk in healthcare processes. In addition, healthcare
point and to comprehend the links between causes and effects of processes are characterized by several cross-functional projects,
adverse events. This enables the definition of appropriate measures such as those related to innovation. Furthermore, the fourth step
for error reduction, also based on technical innovation. of the methodology makes use of a HRA technique, Failure Modes
698 A.C. Cagliano et al. / Safety Science 49 (2011) 695–708

and Effects Analysis (FMEA), which is by far one of the most exten- placed within a matrix structure where horizontal lanes represent
sively adopted in the healthcare sector in the last decades. It has different process phases and vertical lanes correspond to actors
been employed in many fields, such as reducing risk in blood trans- performing activities. This structure is called cross-functional flow
fusions (Burgmeier, 2002), intravenous drug infusions (Apkon et al., chart (Damelio, 1996).
2004), improving drug distribution systems (Lyons, 2009; McNally All the pieces of information related to single activities repre-
et al., 1997), and drug prescription in hospitals (Saizy-Callaert et al., sented by flow charts are summed up by process sheets (Fig. 2).
2001). Moreover, FMEA has been recently endorsed by the health- These tables contain the following details:
care industry and the Joint Commission on Accreditation of Health-
care Organizations as a tool for reducing risk to patients (Brown  name or code of both process phase and activity at issue;
et al., 2008; Ookalkar et al., 2009; Stalhandske et al., 2003).  actors performing the activity;
The following section gives a detailed description of each step of  inputs (information, materials, preliminary actions, orders,
the developed methodology. etc.);
 detailed description of operations required by the activity;
4.2. Description of the steps of the methodology  duration and frequency;
 controls to monitor activity progress;
4.2.1. Context analysis  tools necessary to perform both the activity and related
Context analysis is aimed to select and become familiar with the controls;
critical healthcare process to be investigated. Also, this is the phase  outputs (other activities, information, and data).
when the various actors responsible for the process at issue get in-
volved into the working group committed to perform the analysis. 4.2.3. Risk identification and assessment
Thanks to expert knowledge elicitation and careful consider- The third phase of the proposed approach moves from the
ation of documents, such as working procedures, organizational understanding of the analysed process to the identification of re-
charts, responsibility maps, and shift plans, the working team gets lated risks, again by using project risk management tools.
a first knowledge of process activities and related flows of both First of all, risk sources are identified by using the Risk Break-
quantitative data and clinical and organizational information. This down Structure (RBS), defined as a source-oriented grouping of
forms the base for the identification of criticalities, which are project risks that organizes and defines the total risk exposure of
investigated more in depth by means of two of the traditional the project. Each descending level represents an increasingly de-
tools employed in risk identification: checklists and interviews tailed definition of sources of risk to the project (Hillson, 2002;
(Grimaldi and Rafele, 2008). Checklists are a useful way of keeping Project Management Institute, 2004). In the present methodology
trace of the lessons learnt from previous events and may be pur- RBS levels are determined based on the knowledge of the process
posefully employed in self-assessment processes and reviews at issue gained during the first two steps, Context analysis and Pro-
(Bartlett et al., 2004). On the other hand, interviews are often used cess mapping, and also according to the experience of the members
for risk identification sessions when it is not possible to make the of the working team. The number of levels should be set so that the
working team meet together. They are usually conducted into a RBS is both comprehensive, that is it includes all possible risk
confidential environment, where the interviewee is encouraged sources, and easy to understand and use to control risks.
to express his idea honestly and without fear of reprisal or blame. A first general classification divides risk sources into internal
and external ones. The risk sources in the first class may be suc-
4.2.2. Process mapping cessfully prevented and managed, whereas those in the second
In this second step of the methodology, typical project risk class are out of the process actors’ control and can be treated only
management tools are used to obtain a more in depth definition with assurance coverage or by avoiding them, for example by mod-
of single process activities, also including the identification of ac- ifying activities where they may occur. Internal risk sources are of
tors in charge for them. The process is divided into phases that particular importance since they can be controlled. According to
are analysed and in turn decomposed into activities, until a satis- Roth (1993), they originate from those elements representing the
factory level of detail has not been reached. In order to accomplish foundations of a healthcare delivery system. These are related to
this task effectively, the Activity Breakdown Structure (ABS) may the three processes enabling healthcare systems to transform in-
be used. The ABS is a tree structure coming from the Work Break- puts into outputs, namely clinical, management, and ancillary pro-
down Structure (WBS) (Project Management Institute, 2001), but it cesses (Vissers, 1998) (Fig. 3).
is process-oriented instead of being product-oriented. The lowest Roth’s model has been used in this work as a guide to find out
level of the ABS contains elementary process activities. However, the main areas of a healthcare delivery system where internal risk
since WBS and ABS have been developed in the context of project sources could be identified:
planning, they lack the time dimension. Therefore, this has been in-
cluded in the proposed methodology by making use of process flow  human resources, with their various tasks, their individual
charts (Graham, 2004). Flow charts allow to locate activities in the knowledge and professional skills. They operate in a specific
lowest ABS level according to a logical-time sequence by means of organizational structure able to plan and program the activities
priority links. In addition, activities in a flow chart may be spatially forming clinical processes;

PHASE ACTIVITY
Process Actors
Inputs
Description
Duration and Frequency
Tests
Tools
Outputs

Fig. 2. Process sheet.

 A.C. Cagliano et al. / Safety Science 49 (2011) 695–708 699

INPUTS GENERIC PROCESSES OUTCOMES

PATIENT CLINICAL PROCESSES

HEALTH
DEMAND - treatment modality
(perceived need) - treatment protocol STATUS
- provider
PROVIDERS

PROCESSES
- patient encounters

ANCILLARY
- number CLIENT
- specialty ENABLING
- teaching PERCEPTION
TECHNOLOGY
- reputation
MANAGEMENT
PURCHASERS PROCESSES COST
(finances) - infrastructure
PERFORMANCE
-structure
SUPPLIERS
- information

Fig. 3. Roth’s model (adapted from Roth (1993)).

 physical and technological supports, used by resources to perform The risk sources in the lowest RBS level are linked to the activ-
their activities. They may be either medical or related to infor- ities in the lowest ABS level by means of the Risk Breakdown Ma-
mation or plant technology; trix (RBM) (Hillson, 2003; Hillson et al., 2006). The RBM allows for
 communication/information, as the basis of the relationships risk identification by simply putting crosses into its cells meaning
among resources and between them and technological sup- that given risk sources impact on given activities.
ports. It plays a strategic role in managing healthcare complex- As a further step, depending on the quantity of information elic-
ity. In fact, similarly to any other complex system (Gandolfi, ited from experts, risk evaluation is performed by estimating prob-
1999), in the healthcare one interactions among professionals abilities of occurrence and impacts with either qualitative or
are more important than individual competencies and activities quantitative scales. The first time the proposed methodology is ap-
to determine the success of a clinical treatment; plied to a case, available information will not be sufficient to make
 physical structure, with all the tools necessary to support clini- a quantitative risk evaluation viable. Only after iterating the pres-
cal, technological, and managerial processes within a healthcare ent method a number of times, a healthcare organization will have
delivery system. developed that risk culture making possible to deepen the analysis
through quantitative techniques such as reactive (after an adverse
It can also be noticed that there is a correlation between the event) or proactive (before an adverse event) indicators (Körvers
foundations of a healthcare delivery system and the barriers re- and Sonnemans, 2008).
quired to intercept and block adverse events according to Reason’s The RBM gives a classification of risky events enabling to select
theory. As a matter of fact, adverse events are caused by the simul- the most critical ones, which require a more detailed analysis in or-
taneous action of deficiencies in the different processes character- der to define an adequate risk response supporting continuous
izing a healthcare delivery system. Therefore, by integrating both improvement efforts. For this purpose, the proposed methodology
Roth’s and Reason’s theories, the barriers existing in a generic integrates FMEA as a Human Reliability Assessment technique
healthcare system may be classified as follows: (Lyons et al., 2004) to further investigate the critical links among
risk sources and activities.
 O – Organization.
 T – Technology. 4.2.4. Failure modes and Waste analysis (FMEA–waste analysis)
 C – Communication. Even if a great number of sheet structures to support FMEA have
 S – Structure. been proposed in literature, the present methodology suggests
specific FMEA tables in order to have a more effective integration
These represent the macro areas forming the second RBS level among FMEA and the RBM. Such tables have been conceived with
as far as internal risk sources are concerned (Fig. 4). the aim of highlighting not only the ineffectiveness of a system but

Fig. 4. RBS structure.

700 A.C. Cagliano et al. / Safety Science 49 (2011) 695–708

EFFECTS SUGGESTED
METHODS
IMPROVEMENT
FAILURE MODE RISK CAUSE TO
COD ACTIONS AND
DESCRIPTION SOURCES DESCRIPTION Immediate Final DETECT
TAKEN
ERRORS
MEASURES

Fig. 5. FMEA table.

METHODS
CLASSIFICATION TO DETECT
DESCRIPTION EFFECTS SUGGESTED
AND DESCRIPTION SOURCES WASTE AND TAKEN
COD OF SOURCES OF IMPROVEMENT
OF MODES OF OF WASTE PEOPLE IN MEASURES
OF WASTE WASTE ACTIONS
WASTE CHARGE OF
THIS TASK

Fig. 6. Waste table.

also its inefficiency. To this end, the study of failure modes has FMEA and waste analysis are integrated into FMEA and Waste
been enhanced by a waste analysis driven by the seven classes of tables, which add the adverse events identified by the RBM to
waste defined by the Toyota Production System (Ohno, 1988). the process sheets developed in the second phase of the
These sources of waste have been adapted to healthcare process methodology.
analysis (Gray, 2007; Zidel, 2006) as follows: Each failure mode associated with an activity is characterized in
FMEA tables by the following pieces of information (Fig. 5):
 Overproduction: doing more than customer requirements. For
example, a similar behaviour may be a consequence of mixing  failure mode code;
drugs in anticipation of patient needs or of hospitalizing  failure mode description;
patients when they could be given medical care at their homes.  risk sources, classified into internal and external ones, as dis-
 Waiting times: whenever no activity is performed, waiting for cussed in the third step of the proposed methodology;
the next event happening, such as waiting for bed assignment,  description of causes determining the failure;
waiting for discharge, waiting for treatment, waiting for diag-  effects;
nostic tests, waiting for supplies, waiting for approval, waiting  most effective methods to detect the failure;
for a physician or a nurse, and long waiting times between cases  suggested improvement actions and degree of success of
in operating rooms. already taken measures.
 Transportation: moving medications, patients, and supplies
without adding value to the process. It is important to observe that causes determining failures can
 Overprocessing: performing unnecessary activities leading to an also be other failure modes. In this context, the 5 Whys method
inefficient use of resources, with the consequence of rising pro- may be applied (Zidel, 2006). This is an approach enabling to ex-
cess costs. Examples of such behaviour include using high plore cause and effect relationships by asking five questions, in or-
skilled resources for repetitive activities that could be per- der to determine the root causes of a failure mode. When multiple
formed by less trained people, employing unnecessary auxiliary failure modes need to be considered to define patients’ exposure to
staff, such as technical, catering or laundry personnel, conduct- risks, our methodology links them through the logical AND opera-
ing redundant tests, and subjecting patients to multiple bed tor. Also, failure mode effects are classified into immediate and fi-
moves. nal ones. The first ones include all those effects impacting on the
 Queues/stock: everything waiting for an event, thus increasing analysed organization, thus increasing costs and waste, but not
costs and taking up room, such as medical devices, drugs, and affecting patients. The second ones are those impacting on pa-
other materials bought by specific departments and stocked tients, both directly and indirectly through correlation with other
for a long time, patients in emergency departments waiting failure modes.
for hospitalization, patients waiting for undergoing diagnostic Waste tables (Fig. 6) rely on six of the healthcare sources of
tests, and prescriptions awaiting transcription. waste defined based on the Toyota Production System principles:
 Movements: unnecessary movements that may generate a waste overproduction, waiting times, transportation, overprocessing,
of time or, in some cases, even hurt people, such as repetitive queues/stock, and movements. The seventh source of waste, pro-
searching for documents and supplies and nurses taking care cess defects, errors, and re-work, concerns all those situations in
of patients at multiple hospital floors. which the occurring of failure modes may cause an activity to be
 Process defects, errors, and re-work: generally defined as activi- performed again, with consequent greater cycle times and costs.
ties not adding value either to the process or to patients, such This kind of waste is not considered by Waste tables since it is al-
as medication errors, wrong-site surgery, improper labelling ready extensively analysed among failure modes in FMEA tables.
of specimens, using multiple sticks for blood draws, and injuries As for failure modes, sources of waste may be classified into
caused by either defective drugs or patient intolerances to spe- internal (related to organization, technology, communication, and
cific treatments. structure) and external ones.
 A.C. Cagliano et al. / Safety Science 49 (2011) 695–708 701

Finally, both FMEA and Waste tables detail people in charge of patients and understanding failure modes and waste, thus stimu-
detecting failure modes and waste and possible improvement ac- lating an improvement in the overall level of service.
tions. In addition, both the tables allow to keep trace of the success
of corrective actions already undertaken. This because FMEA and 5.1. Context analysis
waste analysis are not static approaches to be performed only
once, but, on the contrary, they are recursive processes to be ap- First of all, the working team in charge of analysing the logistics
plied overtime to constantly monitor how system outcomes react process of the Pharmacy Department was formed. It included both
to both internal strategies and external inputs. the authors and hospital representatives.
Expert knowledge elicitation was performed by interviewing
5. Applying the methodology to a hospital pharmacy pharmacists and logistics managers. Gathered information, as well
department as provided documentation, allowed to analyse all the procedures
currently in place for the portion of drug process managed by the
In order to exemplify how the developed methodology should Pharmacy. In addition, operational activities were directly observed.
be used to analyse healthcare processes, its application to the drug In this phase the working team was able to uncover the main issues
management process at a hospital pharmacy department is that were deepened by the later steps of the methodology.
detailed.
This case was selected since drug and other material manage- 5.2. Process mapping
ment is one of the most cross-functional processes taking place
in a hospital. In fact, it involves many activities performed by dif- The ABS and the flow charts describing both operational and
ferent departments, including central pharmacy and operating informational flows revealed three phases in the investigated pro-
rooms, starting from when materials are sourced from suppliers cess. First, after physicians prescribe therapies, floor personnel in
until they are employed to deliver care to patients. As a conse- charge of material management requests necessary drugs and med-
quence, effectiveness and efficiency are strongly influenced by ical devices to the Pharmacy Department mainly thorough a com-
the way such process is globally managed. Some authors proved puterized procedure (Material Request Issue by Floors). Second, in
that logistics and sourcing costs represent a big portion of the total the Pharmacy Department, after approval by the chief pharmacist,
costs for a hospital (Linch, 1991). Moreover, adverse events due to requested materials are picked from shelves and placed into bas-
incorrect drug administration (Adverse Drug Events) are common kets to be delivered to floors by means of trolleys. In a similar
causes for injury among hospitalized patients and may be origi- way, orders are placed to suppliers after validation by the chief
nated by any part of the drug management process (Cohen, pharmacist. Incoming products are inspected to check their compli-
2007). Thus, we chose to analyse the logistics process of a phar- ance with orders (Pharmacy Request and Material Management).
macy department since adverse events taking place downstream Third, before leaving the Pharmacy Department, outgoing packages
in the drug management process may find their root causes within are checked by pharmacists, afterwards they reach the destination
this converging point for materials in a hospital. wards together with a copy of the order, and finally the material re-
The present case study focuses on a 1372 bed teaching hospital ceipt confirmation is signed by ward personnel and filed in the
located in Torino (Italy). This is the oldest operating hospital in Pharmacy Department (Material Request Fulfilment).
town, and the largest in Piedmont Region of Italy, spread over Phases were in turn decomposed into activities to form an ABS
142,000 square meters, 14 clinical departments, and 5822 employ- (Fig. 7). We decided to develop process sheets only for critical activ-
ees, with 1030 physicians and 2063 nurses among them (Cagliano ities (see the application of the fourth phase of the methodology).
et al., 2009). Also, this is one of the most complex hospitals in Italy
as far as organizational flows are concerned. To be more precise, 5.3. Risk identification and assessment
the application of the proposed clinical risk management approach
was aimed to study central pharmacy’s drug supply to the hospital The previous phases of the methodology served as a basis to
wards, with the purpose of identifying possible sources of risk for identify risks related to the process under consideration, particularly

PROCESS ACTIVITY
PROCESS PHASE
DESCRIPTION CODE
Starting the Computerized Material Request Procedure ABS 1.1
Computerized Material Request Creation ABS 1.2
1. Material Request Issue by Floors
Material Request Check and Validation ABS 1 .3
Sending Material Requests to Pharmacy ABS 1.4
Material Request Receiving ABS 2 . 1
Material Request Validation by Pharmacists ABS 2.2
Substituting Therapeutic Equivalents for Unavailable Products ABS 2.3
Defining Picking Lists ABS 2.4
List Fulfilment & Material Picking ABS 2.5
2. Pharmacy Request and Material Ma te r i a l P a ck i ng ABS 2.6
Management Checking Reorder Levels for Products ABS 2 .7
Receiving Incoming Materials ABS 2.8
Material Storing ABS 2.9
Incoming Material Data Entry ABS 2.10
Filing Incoming Material Documents ABS 2.11
Returning Products to Suppliers ABS 2.12
Preparing Trolleys for De livery to Floors ABS 3. 1
Outgoing Package Sample Quality Inspection ABS 3.2
Material Delivery to Floors ABS 3.3
3. Material Request Fulfilment
Checking Pending Material Requests ABS 3.4
Material Receiving by Floors ABS 3.5
Product Transaction Registration ABS 3.6

Fig. 7. ABS for drug management process.

702 A.C. Cagliano et al. / Safety Science 49 (2011) 695–708

LEVEL 0 LEVEL 1 LEVEL 2 LEVEL 3 LEVEL 4 RBS CODE

Planning Activities RBS 1.1
Organizational Planned Work
RBS 1.2
Structure Schedule
Workload RBS 1.3
Working Procedures
Knowledge and RBS 1.4
Compliance
Training/Know How RBS 1.5
Availability of
Personnel in Charge
Human Resources RBS 1.6
of Supervising
Activities
1. Organization Controls RBS 1.7
Planned Work
Schedule Knowledge RBS 1.8
and Compliance
Determining the
RBS 1.9
Kinds of Products
Determining Product
RBS 1.10
Quantity
Operations Computerized
RBS 1.11
Procedures
Transcription RBS 1.12
Stocking Products RBS 1.13
Moving Products RBS 1.14
Continuity of Service RBS 2.1
Intranet RBS 2.2
Internal Risk Data Transfer RBS 2.3
Sources Information System
Data Backup and
RBS 2.4
Network Records
Managing Antivirus RBS 2.5
2. Technology Handling Systems RBS 2.6
Revolving Shelves RBS 2.7
Elevators RBS 2.8
RBS for Pharmacy
Equipment Computers and
Logistics Process RBS 2.9
Palmtops
Boxes for Delivering
RBS 2.10
Products
Information
Exchanges According RBS 3.1
to Procedures
Information Exchanges
Variations in
RBS 3.2
Quantity
3. Communication Variations in Quality RBS 3.3
Communicating
Communicating RBS 3.4
Variations
Variations and
Feedback RBS 3.5
Decisions
Decision Making RBS 3.6
Ordinary
RBS 4.1
Maintenance Plans
Layout Extraordinary
RBS 4.2
4. Structure Maintenance Plans
Workplace Safety RBS 4.3
Service Interruptions RBS 4.4
Networks
Service Continuity RBS 4.5
Delivery Lead Times RBS 5.1
Del ive ry Poi nt s R BS 5 .2
Quality of Delivered
RBS 5.3
5. Product Supplying Products
Documentation
Management RBS 5.4

External Risk Delivered Items RBS 5.5

Sources Supplier Assets RBS 6.1
6. Finance
Contract Specifications RBS 6.2
Guidelines by Regional
RBS 7.1
Council
7. Environment Social Issues RBS 7.2
Epidemiological Events RBS 7.3
Natural Events RBS 7.4

Fig. 8. RBS for drug management process.

by combining the information gathered during Context analysis and In such a way, sources of risk were identified and classified
Process mapping with the experience of the working team compo- according to a RBS. This structure was then intersected with
nents about both risk and healthcare process management. elementary activities in the ABS to give the RBM for the drug
A.C. Cagliano et al. / Safety Science 49 (2011) 695–708 703

Fig. 9a. RBM for drug management process (part 1).

 704
A.C. Cagliano et al. / Safety Science 49 (2011) 695–708
Fig. 9b. RBM for drug management process (part 2).

Fig. 9c. RBM for drug management process (part 3).

 A.C. Cagliano et al. / Safety Science 49 (2011) 695–708 705

ACTIVITY
PHASE
2.5 List Fulfilment & Material Picking
Process Actors Warehouse personnel
Picking lists (both computer and paper
Inputs
based)
Warehouse personnel prepare materials
requested by floors according to picking
lists. This task is performed by
Description
2. Pharmacy Request following a logistics path allowing
and Material optimizing the sequence of picking
Management operations
Duration and Frequency Accordi ng to the defined schedule
Matching between requested quantities
Tests
and delivered ones.
Revolving shelves
Tools
Forklifts
Outputs Material Packing

Fig. 10. Process sheet for List Fulfilment & Material Picking.

EFFECTS SUGGESTED
METHODS TO IMPROVEMENT
FAILURE MODE RISK CAUSE
COD DETECT ACTIONS AND
DESCRIPTION SOURCES DESCRIPTION Immediate Final ERRORS TAKEN
MEASURES
Two different
Verbal pharmacy operators
X AND Floor personnel
communication only should always
double-check units

Two different
Misunderstanding of Lack of both verbal
pharmacy operators
FM1 units expressing C and written X AND Floor personnel
should always
requested quantities communication
double-check units

Scarce
communication
X AND Floor personnel
among Pharmacy
warehouse personnel
Floor
Wrong identification
T X --- personnel/visual
of warehouse location
check
Picking of the wrong
Products difficult to
FM2 items to be delivered
be identified (e.g.
to floors Use of optical
O similar packages; --- AND Floor personnel
barcode reading
same packages, but
different dosage)
Documentation not No communication
updated according to about changes in Administrative
FM3 C X ---
changes in quantities quantities requested control
requested by floors. by floors
Unavailability of Pharmacy
personnel in charge of X AND warehouse
Lack of controls on
controlling personnel
the matching between
FM4 O Training courses.
requested and Pharmacy
Having operators be
delivered quantities Lack of staff training X --- warehouse
supported by
manager
qualified personnel.

High number of
Pharmacy Having picking lists
picking lists received
X --- warehouse follow a precise path
FM5 Loosing picking lists O by the Pharmacy
personnel within the Pharmacy
Department every day
Department

Confusion X ---
Inadequate Maintenance Careful ordinary
X ---
maintenance service plans maintenance plans
Training warehouse
Pharmacy personnel to deal
Unwary operations X --- warehouse with machine
manager breakdowns
FM6 Machine breakdowns T
effectively
Training warehouse
Pharmacy personnel to deal
Inadequate staff
X --- warehouse with machine
training
manager breakdowns
effectively
FM7 Staff unavailability O Unexpected absences X ---

Fig. 11. FMEA table for List Fulfilment & Material Picking.
706 A.C. Cagliano et al. / Safety Science 49 (2011) 695–708

METHODS
CLASSIFICATION
TO DETECT
AND DESCRIPTION EFFECTS SUGGESTED
SOURCES WASTE AND TAKEN
COD DESCRIPTION OF OF SOURCES OF OF IMPROVEMENT
OF WASTE PEOPLE IN MEASURES
MODES OF WASTE WASTE ACTIONS
CHARGE OF
WASTE
THIS TASK

Not
Having a pharmacist,
optimized
Pharmacy or another
Poor coordination picking
warehouse professional figure,
O among warehouse paths.
personnel. monitor both picking
personnel Inadequate
Pharmacists paths and task
task
assignment
assignment
Useless motions by
W1 Not
pharmacy operators
optimized
Having a pharmacist,
picking Pharmacy
or another
paths, thus warehouse
T Wrong picking lists professional figure,
operators personnel.
monitor picking
take longer Pharmacists
paths
to pick
items

Fig. 12. Waste table for List Fulfilment & Material Picking.

management process at the Pharmacy Department. Developed RBS resources working in this hospital department are subjected. The
and RBM are presented in Figs. 8 and 9a–c respectively. effects may be both immediate and final. As an improvement
A first correlation between sources of risk and elementary action, two different pharmacy operators should always double-
activities was established by putting crosses in the correspond- check units. Moreover, the picking of the wrong items to be deliv-
ing RBM cells. The limited information available to the working ered to floors is given by both technological and organizational
team in this first application of the methodology did not issues. These may include the wrong identification of either ware-
allow to quantify risks by evaluating their probabilities of occur- house location or package to be picked. Also this failure mode may
rence and their impacts on activities. However, the fourth step of have both immediate and final effects according to the risk source
the methodology may be applied also with a qualitative risk generating it, and can be prevented by the use of optical barcode
evaluation. reading. Finally, loosing picking lists is determined by organiza-
tional issues, such as the high number of picking lists received by
5.4. Failure modes and waste analysis (FMEA–Waste analysis) the Pharmacy Department every day. Related effects do not affect
patients but process time and costs, since they imply that picking
The analysis of the RBM and further interviews to the Pharmacy lists are prepared again. As a consequence, improvement actions
management revealed that 8 out of the 22 identified elementary having picking lists follow a precise path within the Pharmacy
activities may be considered critical so that it is worth investigat- Department are highly recommended.
ing them by means of FMEA and Waste tables. They are namely The analysis of List Fulfilment & Material Picking activity re-
Computerized Material Request Creation, Material Request Check vealed the following kind of waste: useless motions by pharmacy
and Validation, List Fulfilment & Material Picking, Material Pack- operators (Fig. 12). It is determined by both organizational issues
ing, Material Storing, Outgoing Package Sample Quality Inspection, (e.g. poor coordination among workers) and technological ones
Material Delivery to Floors, and Product Transaction Registration. (e.g. wrong picking lists). The effect is the same: operators do not
First, failure modes (FM) and kinds of waste (W) impacting crit- follow optimized paths, thus taking longer to pick items, with
ical activities were numbered according to the following notation: the risk of getting in one another’s way. As a solution, it is sug-
FM1, FM2, . . . , W1, W2, . . . After that, they were put in the corre- gested to have a pharmacist, or another professional figure, moni-
sponding RBM rows, under the sources of risk generating them tor picking paths.
(Fig. 9a–c). A same failure mode or kind of waste may appear It can be observed that for this activity, as well as for the other
multiple times in a RBM row if it may affect an activity as a conse- ones being analysed, the number of failure modes is far greater
quence of more than one source of risk. than the number of kinds of waste. As a matter of fact, many poten-
The description of the application of the fourth step of the pro- tial sources of waste are related to the class ‘Process defects, errors,
posed methodology to the case will focus on List Fulfilment & and re-work’, which, as explained before, is addressed by FMEA ta-
Material Picking. Several failure modes and kinds of waste were bles. Therefore, they are considered as failure modes. It is the case
defined for this activity. of loosing picking lists, which asks for additional operational activ-
Figs. 10 and 11 show the process sheet and the FMEA table for ities such as reintegrating the stock of products that have been
List Fulfilment & Material Picking activity. As far as the effects of wrongly picked and delivering the correct materials to floors.
failure modes are concerned, the symbol X means that the failure
mode at issue has some kind of effect, whereas the logical AND
operator indicates that a failure mode, together with other failure 6. Discussion
modes, has a final effect on patients. Following the discussion of
some of the identified failure modes. First of all, the misunder- Based on the results of its application to the drug management
standing of units expressing the quantities of materials wards or- process at a hospital pharmacy, strengths and weaknesses of the
der to the Pharmacy Department is due to the fact that relevant developed approach to clinical risk management are here discussed.
information is often communicated verbally, and double-checking First of all, the systemic feature of the suggested methodology is
is sometimes impossible because of the heavy workloads to which assured by the adoption of the RBM. The RBM frames all risk
 A.C. Cagliano et al. / Safety Science 49 (2011) 695–708 707

sources into the specific activities characterizing the process at is- The flexibility of our methodology potentially allows the inte-
sue. Furthermore, it gives a global view of criticalities, making it gration with risk management approaches already established in
easy to define correlations among different failure modes in order the healthcare sector, such as for instance Fault Tree Analysis
to trace at the root all the determinants of adverse events. This is (FTA), Hazard and Operability Study (HAZOP), and Incident Report-
crucial in healthcare since the occurring of an adverse event that ing (Armitage et al., 2007; Lyons, 2009). These techniques may
may hurt hospitalized patients is often linked to multiple interre- work at the level of single RBM cells by performing either qualita-
lated failure modes giving rise to a failure mode chain. For exam- tive (e.g. HAZOP, Incident Reporting) or quantitative analyses (e.g.
ple, in the drug management process, the administration of a Monte Carlo simulation), according to the availability of data and
wrong medicine may be due to a picking error by the pharmacy the degree of organizational maturity towards risk management.
operator that has not been detected before the drug arrives at Also, multiple RBM cells may be considered in order to understand
the patient’s bed (Hollnagel, 2004). The systemic perspective of the root causes of a failure mode or of a kind of waste. FTA could be
the RBM enhances the effectiveness of FMEA because it supports applied for this purpose, since it is not limited to the investigation
a more comprehensive understanding of the relationships between of a single system but usually crosses system boundaries. To be
causes and effects of failure modes. Furthermore, the RBM provides more precise, FTA would break down the top event to find out
not only a systemic but also a schematic representation of critical- the parallel and sequential combinations of basic faults responsible
ities, thus making the proposed methodology a valid communica- for it. To this end, the use of logical operators to link failure modes
tion tool for organizational members. in FMEA tables is a first attempt to correlate different risky events.
In addition, the methodology revealed to be extremely flexible Moreover, the role that Key Performance Indicators (KPIs) could
since it is able to work at different levels of detail according to have in the approach as pre-warning signals anticipating the
the specific case and the information available. occurrence of adverse events should be investigated. In particular,
In the developed clinical risk management approach, first pro- RBM cells could be associated with proper metrics able to capture
cess criticalities are identified by means of a reactive analysis the impact of the symptoms of a risk source manifestation on the
based on past adverse events. Usually, such events have not been performance of a given activity.
recorded, thus expert knowledge elicitation is used to encourage Although combing the mentioned approaches with our methodol-
the emergence of process actors’ experience about inefficiencies ogy increases the knowledge about the origins of patients’ exposure
and ineffectiveness. As a further step, thanks to the mapping of to risks and allows a better planning of proper countermeasures, it
the discrepancies in the system barriers (failure modes and kinds may require healthcare organizations additional efforts to develop
of waste), the RBM methodology, integrated with FMEA and waste new skills about the management of risk and safety. Nevertheless,
analysis, is able to make operators aware of both risks and waste we believe that this stream of research deserves future attention
existing in a healthcare process. Therefore, in a sense, the proposed because it contributes to enhance the suitability of the methodology
methodology also constitutes a valid tool for stimulating a struc- discussed in the paper for a variety of settings.
tured analysis of criticalities, which is absolutely important in a
highly human-based context like the healthcare one. 7. Summary
Moreover, the present framework could support decision mak-
ers in setting correct priority areas for intervention and may be a The growing healthcare complexity requires management ap-
part of Health Technology Assessment programs. This is guaran- proaches taking into account multiple points of view. Based on
teed by the identification of improvement actions in the last step Reason’s theory of failures, the paper suggests a methodology giv-
of the method. ing a systemic perspective on clinical risk by integrating existing
Finally, the developed clinical risk management method may be tools coming from different fields, such as process mapping, pro-
applied overtime to review the effectiveness of the implementa- ject risk management, and quality management. Moreover, be-
tion of corrective actions to limit risks and waste. To this end, cause of the human-centred nature of healthcare systems, the
the RBM and FMEA and Waste tables will be updated, and, if nec- steps of such methodology have been developed according to Hu-
essary, new corrective actions will be developed and adopted. As a man Reliability Assessment methods.
consequence, the RBM and FMEA and Waste tables also prove to be The first application to the logistics process of a pharmacy
useful means of communication among people involved in the department in a large hospital highlighted that our method can
improvement process. effectively support not only risk analysis, but also decision making,
The implementation of the methodology in the case hospital thus increasing organizations’ maturity towards clinical risk. Fu-
revealed great difficulty in gathering all the pieces of information ture research efforts will be focused on an extensive test of the pre-
necessary to fully apply the four steps, due to a scarce aptitude for sented approach in various healthcare contexts.
risk management and, as a consequence, for supporting such a
comprehensive organizational analysis by both personnel and infor- Acknowledgements
mational systems. As a matter of fact, this first application to the
logistics process of a pharmacy department was limited to risk iden- This research is part of the National Relevance Research Project
tification, without performing any quantitative evaluations. To this 2008 (PRIN Project 2008) ‘‘Analysis of models and solutions for
end, it stimulated an increase in the level of maturity towards risk hospital logistics’’.
of the studied organization, thus enabling future deeper analyses. The authors are grateful to the Chief Executive and to the phar-
Overall, the application of the proposed methodology may serve macy personnel of San Giovanni Battista Hospital, Torino, Italy, for
as a first step towards a deeper understanding of risk and waste in their active collaboration, to the Agency for Healthcare Services of
healthcare processes and the definition of the most appropriate Piedmont Region of Italy for sponsoring the present research, and
measures to reduce them. It may be the foundation of a quantita- to Dr. Carlo Enrico Carcangiu and Dr. Elio Sgherzi for contributing
tive risk evaluation by numerically determining the probabilities of to the development of the case study.
occurrence of risks as well as their impacts on process activities.
However, in order to prove the full benefits and limitations of References
the suggested approach and understand if it requires further con-
Apkon, M., Leonard, J., Probst, L., DeLizio, L., Vitale, R., 2004. Design of a safer
ceptual refinements, an extensive application to a variety of approach to intravenous drug infusions: failure mode effects analysis. Quality
healthcare settings is needed. and Safety in Healthcare 13, 265–271.
708 A.C. Cagliano et al. / Safety Science 49 (2011) 695–708

Armitage, G., Newell, R., Wright, J., 2007. Reporting drug errors in a British acute Liker, J.K., Hoseus, M., 2008. Toyota Culture. The Heart and Soul of the Toyota Way.
hospital trust. Clinical Governance. An International Journal 12 (2), 102–114. McGraw-Hill.
Barach, P., 2002. Lessons from the USA. In: Emslie, S., Knox, K., Pickstone, M. (Eds.), Linch, B.S., 1991. Just in time and stockless program for hospital: fad or trend?
Improving Patient Safety: Insights from American, Australian and British Hospital Material Management Quarterly 12, 17–21.
healthcare. ECRI and Department of Health, London, UK. Lucas, D., 1997. The causes of human error. In: Redmill, F., Rajan, J. (Eds.), Human
Bartlett, J., Campbell, P., Chapman, C., Close, P., Davey, K., Desai, P., Greenwood, M., Factors in Safety-Critical Systems. Reed Educational and Professional Publishing
Groom, H., Hillson, D., Hopkinson, M., Gerdes, R., Major, E., Newland, K., Ltd., Oxford, UK.
Simister, S., Williams, T., 2004. Project Risk Analysis and Management Guide, Lyons, M., 2009. Towards a framework to select techniques for error prediction:
second ed. APM Publishing Limited, Buckinghamshire, UK. supporting novice users in the healthcare sector. Applied Ergonomics 40, 379–
Baysari, M.T., McIntosh, A.S., Wilson, J.R., 2008. Understanding the human factors 395.
contribution to railway accidents and incidents in Australia. Accident Analysis Lyons, M., Adams, S., Woloshynowych, M., Vincent, C., 2004. Human reliability
and Prevention 40, 1750–1757. analysis in healthcare: a review of techniques. International Journal of Risk and
Baysari, M.T., Caponecchia, C., McIntosh, A.S., Wilson, J.R., 2009. Classification of Safety in Medicine 16 (4), 223–237.
errors contributing to rail incidents and accidents: a comparison of two human McDonald, T., Mayer, D., 2008. Full disclosure is more than saying ‘‘I’ sorry’’. Focus
error identification techniques. Safety Science 47, 948–957. on Patient Safety 11 (1), 1–3.
Bridges, J.F.P., 2006. Lean systems approaches to health technology assessment. A McNally, K.M., Page, M.A., Sunderland, V.B., 1997. Failure-mode and effects analysis
patient-focused alternative to cost-effectiveness analysis. PharmacoEconomics in improving a drug distribution system. American Journal of Health-System
24 (2), 101–109. Pharmacy 54 (2), 171–177.
Brown, A., Eatock, J., Dixon, D., Meenan, B.J., Anderson, J., 2008. Quality and Norman, D.A., 1990. The ‘problem’with automation: inappropriate feedback and
continuous improvement in medical device manufacturing. The TQM Magazine interaction, not ‘over-automation’. Philosophical Transactions of the Royal
20 (6), 541–555. Society of London, Series B 327, 585–593.
Burgmeier, J., 2002. Failure mode and effect analysis: an application in reducing risk O’Hagan, A., Oakley, J.E., 2004. Probability is perfect, but we can’t elicit it perfectly.
in blood transfusion. Joint Commission Journal on Quality Improvement 28 (6), Reliability Engineering and System Safety 85 (1–3), 239–248.
331–339. Ohno, T., 1988. Toyota Production System: Beyond Large-scale Production.
Cagliano, A.C., Rafele, C., Singh, M., 2009. Improving hospital supply chains by Productivity Press, New York.
managing risk flows. In: Proceedings of the 16th International Annual EurOMA Ookalkar, A.D., Joshi, A.G., Ookalkar, D.S., 2009. Quality improvement in
Conference, 14–17 June 2009, Göteborg, Sweden. haemodialysis process using FMEA. International Journal of Quality &
Cohen, M.R., 2007. Medication Errors. American Pharmacists Association, Reliability Management 26 (8), 817–830.
Washington, DC. Paz Barroso, M., Wilson, J.R., 2000. Human error and disturbance occurrence in
Cosby, K.S., 2003. A framework for classifying factors that contribute to error in the manufacturing systems (HEDOMS): a framework and a toolkit for practical
emergency department. Annals of Emergency Medicine 42 (6), 815–823. analysis. Cognition, Technology & Work 2, 51–61.
Damelio, R., 1996. The Basics of Process Mapping. Productivity Press, New York, NY. Project Management Institute, 2001. Practice Standard for Work Breakdown
Davis, P., Lay-Yee, R., Shug, S., Briant, R., Scott, A., Johnson, S., Bingley, W., 2001. Structures. PMI, Philadelphia, PA.
Adverse events regional feasibility study: indicative findings. The New Zealand Project Management Institute, 2004. A Guide to the Project Management Body of
Medical Journal 114 (1131), 203–205. Knowledge (PMBOK Guide), third ed. Project Management Institute, Newtown
Embrey, D.E., 1992. Quantitative and qualitative prediction of human error in safety Square, PA.
assessments. In: I. Chem. E. Symposium Series 130. I. Chem. E., London, UK. Reason, J., 2001. Understanding adverse events: the human factor. In: Vincent, C.
Embrey, D., 2000. Data Collection Systems. Human Reliability Associates, Wigan, (Ed.), Clinical Risk Management: Enhancing Patient Safety, second ed. BMJ
Lancashire, UK. Books, London, UK.
Gandolfi, A., 1999. Formicai, imperi, cervelli. Introduzione alla scienza della Reason, J., 2002. Combating omission errors through task analysis and good
complessità. Bollati Boringhieri Editore, Torino, Italy. reminders. Quality and Safety in Health Care 11 (1), 40–44.
Garthwaite, P.H., Kadane, J.B., O’Hagan, A., 2005. Statistical methods for eliciting Ross Baker, G., Norton, P.G., Flintoft, V., Blais, R., Brown, A., Cox, J., Etchells, E., Ghali,
probability distributions. Journal of the American Statistical Association 100 W.A., Hébert, P., Majumdar, S.R., O’Beirne, M., Palacios-Derflingher, L., Reid, R.J.,
(470), 680–700. Sheps, S., Tamblyn, R., 2004. The Canadian Adverse Events Study: the incidence
Graham, B.B., 2004. Detail Process Charting. Speaking the Language of Process. John of adverse events among hospital patients in Canada. Canadian Medical
Wiley & Sons, Inc., Hoboken, NJ. Association Journal 170 (11), 1678–1686.
Gray, J.A.M., 2007. How to Get Better Value Healthcare. Offox Press, Oxford, UK. Roth, A.V., 1993. World class health care. Quality Management in Health Care 1 (3),
Grimaldi, S., Rafele, C., 2008. A tool selection and support methodology for project 1–9.
risk management. In: Proceedings of the 22nd IPMA World Congress ‘‘Project Saizy-Callaert, S., Causse, R., Thebault, A., Chouaid, C., 2001. Analysis of mode of
Management to Run’’, 9–11 November 2008, Rome, Italy. failure, their effects and criticality: improving of the hospital drug prescribing
Harms-Ringdahl, L., 2009. Analysis of safety functions and barriers in accidents. process. Therapie 56 (5), 525–531.
Safety Science 47 (3), 353–363. Sale, D., 2005. Understanding Clinical Governance and Quality Assurance. Palgrave
Harvard Business School, 2004. Managing Projects Large and Small – The Macmillan, New York.
Fundamental Skills for Delivering on Budget and on Time. Harvard Business Shorrock, S.T., 2002. Error classification for safety management: finding the right
School Press, Boston, MA. approach. In: Proceedings of a Workshop on the Investigation and Reporting of
Henriksen, K., Dayton, E., Keyes, M.A., Carayon, P., Hughes, R., 2008. Understanding Incidents and Accidents, 17–20 July 2002. The Senate Room, University of
adverse events: a human factors framework. In: Hughes, R.G. (Ed.), Patient Glasgow.
Safety and Quality: An Evidence-based Handbook for Nurses. Agency for Stalhandske, E., DeRosier, J., Patail, B., Gosbee, J., 2003. How to make the most of
Healthcare Research and Quality, US Department of Health and Human failure mode and effect analysis. Biomedical Instrumentation & Technology 37
Services, Rockville, MD. (2), 96–102.
Hillson, D., 2002. The risk breakdown structure (RBS) as an aid to effective risk Thomas, E.J., Studdert, D.M., Burstin, H.R., Orav, E.J., Zeena, T., Williams, E.J.,
management. In: Fifth European Project Management Conference, PMI Europe Howard, K.M., Weiler, P.C., Brennan, T.A., 2000. Incidence and types of
2002, 19–20 June 2002, Cannes, France. adverse events and negligent care in Utah and Colorado. Medical Care 38 (3),
Hillson, D., 2003. Effective Opportunity Management for Projects: Exploiting 261–271.
positive risk. Marcel Dekker, New York. Tonneau, D., 1997. Management tools and organization as key factors towards
Hillson, D., Grimaldi, S., Rafele, C., 2006. Managing project risks using a cross risk quality care: reflections from experience. International Journal for Quality in
breakdown matrix. Risk Management 8, 61–76. Health Care 9 (3), 201–205.
Hollnagel, E., 1998. CREAM: Cognitive Reliability Error Analysis Method. Elsevier, Vincent, C., 2001. Clinical Risk Management: Enhancing Patient Safety, second ed.
London, UK. BMJ Books, London, UK.
Hollnagel, E., 2004. Barriers and Accident Prevention. Ashgate Publishing Limited, Vincent, C., 2006. Patient Safety. Elsevier Science Limited, London, UK.
Aldershot, UK. Vincent, C., Neale, G., Woloshynowych, M., 2001. Adverse events in British
Janssen, W., Wierda, M., Van Der Horst, R., 1995. Automation and the future of hospitals: preliminary retrospective record review. British Medical Journal
driver behavior. Safety Science 19 (2–3), 237–244. 322, 517–519.
Johansson, H.J., Mchugh, P., Pendlebury, J., Wheeler, W.A., 1993. Business Process Vissers, J.M.H., 1998. Health care management modelling: a process perspective.
Reengineering: Breakpoint Strategies for Market Dominance. John Wiley & Sons, Health Care Management Science 1, 77–85.
Chichester, UK. Wiegmann, D.A., Shappell, S.A., 2001. Human error perspectives in aviation. The
Kohn, L., Corrigan, J., Donaldson, M. (Eds.), 1999. To Err is Human: Building a Safer International Journal of Aviation Psychology 11 (4), 341–357.
Health System. National Academy Press, Washington, DC. Wieman, T.J., Wieman, E.A., 2004. A systems approach to error prevention in
Körvers, P.M.W., Sonnemans, P.J.M., 2008. Accidents: a discrepancy between medicine. Journal of Surgical Oncology 88, 115–121.
indicators and facts! Safety Science 46, 1067–1077. Wilson, R.M., Runciman, W.B., Gibberd, R.W., Harrison, B.T., Newby, L., Hamilton,
Leape, L.L., Brennan, T.A., Laird, N., Lawthers, A.G., Localio, A.R., Barnes, B.A., Hebert, J.D., 1995. The quality in australian health care study. The Medical Journal of
L., Newhouse, J.P., Weiler, P.C., Hiatt, H., 1991. The nature of adverse events in Australia 163 (9), 458–471.
hospitalized patients. Results of the Harvard Medical Practice Study II. The New Zidel, T.G., 2006. A lean toolbox-using lean principles and techniques in healthcare.
England Journal of Medicine 324 (6), 377–384. Journal for Healthcare Quality 28 (1).
Liker, J.K., 2004. The Toyota Way. 14 Management principles from the World’s
Greatest Manufacturer. McGraw-Hill.