Preventing Posttraumatic Stress in ICU Survivors:

A Single-Center Pilot Randomized Controlled

Trial of ICU Diaries and Psychoeducation
Maia S. Kredentser, MSc1; Marcus Blouw, MD, MHA, FRCPC2; Nicole Marten, RN3;
Jitender Sareen, MD, FRCPC4; O. Joseph Bienvenu, MD, PhD5; Jennifer Ryu, MD6; Brooke E. Beatie, MA1;
Sarvesh Logsetty, MD, FRCSC, FACS4,7; Lesley A. Graff, PhD8; Shauna Eggertson, RN, BN9;
Sophia Sweatman, MD10; Braeden Debroni, MD11; Nina Cianflone12; Rakesh C. Arora, MD, PhD, FRCSC7,13;
Ryan Zarychanski, MD, MSc, FRCPC2; Kendiss Olafson, MD, MPH, FRCPC2

1
Department of Psychology, University of Manitoba, Winnipeg, MB, Ms. Kredentser has received doctoral research awards from the Canadian
Canada. Institutes of Health Research and the University of Manitoba Centre on
2
Department of Internal Medicine, Max Rady College of Medicine, Rady Fac- Aging and Wu Scholarship. Dr. Blouw, Ms. Kredentser, and Drs. Sareen,
ulty of Health Sciences, University of Manitoba, Winnipeg, MB, Canada. Bienvenu, Ryu, Logsetty, and Olafson received funding from Manitoba
Medical Services Foundation. Drs. Blouw, Sareen, Graff, Logsetty, and
3
Department of Internal Medicine, Section of Critical Care, St Boniface Olafson received support for article research from the Canadian Institutes
General Hospital, Winnipeg, MB, Canada. of Health Research. Dr. Sareen disclosed that he was a consultant for UpTo-
4
Department of Psychiatry, Max Rady College of Medicine, Rady Faculty Date and has previously held Johnson and Johnson stock. Ms. Beatie has
of Health Sciences, University of Manitoba, Winnipeg, MB, Canada. received doctoral research awards from the Social Sciences and Humani-
5
Department of Psychiatry and Behavioral Sciences, Johns Hopkins Uni- ties Research Council of Canada and the University of Manitoba Centre
versity School of Medicine, Baltimore, MD. on Aging. Drs. Ryu, Sweatman, and Debroni received a research stipend
from the University of Manitoba for completing the Med II Summer Research
6
Department of Family Medicine, Faculty of Medicine, University of Program while working on this article. Dr. Arora has received an unrestricted
Ottawa, Ottawa, ON, Canada. educational grant from Pfizer Canada Inc and honoraria from Mallickrodt
7
Department of Surgery, Max Rady College of Medicine, Rady Faculty of Pharmaceuticals for work unrelated to this article. The remaining authors
Health Sciences, University of Manitoba, Winnipeg, MB, Canada. have disclosed that they do not have any potential conflicts of interest.
8
Department of Clinical Health Psychology, Max Rady College of Medi- For information regarding this article, E-mail: [email protected]
cine, Rady Faculty of Health Sciences, University of Manitoba, Winni-
peg, MB, Canada.
9
Melanoma Clinical Research Unit, Duke Cancer Institute, Duke Univer- Objectives: Critical illness can have a significant psychological
sity School of Medicine, Durham, NC. impact on patients and their families. To inform the design of a
10
Department of Pediatrics, Faculty of Medicine, University of Toronto, larger trial, we assessed feasibility of ICU diaries and psychoedu-
Toronto, ON, Canada.
cation to prevent posttraumatic stress disorder, depression, and
11
Department of Family Medicine, Max Rady College of Medicine, Rady Fac-
anxiety following ICU stays.
ulty of Health Sciences, University of Manitoba, Winnipeg, MB, Canada.
Design: Four-arm pilot randomized controlled trial.
12
College of Dentistry, Rady Faculty of Health Sciences, University of Man-
itoba, Winnipeg, MB, Canada. Setting: A 10-bed tertiary ICU in Winnipeg, MB, Canada.
13
Cardiac Sciences Program, St. Boniface General Hospital, Winnipeg, Patients: Critically ill patients greater than 17 years old with pre-
MB, Canada. dicted ICU stays greater than 72 hours and mechanical ventilation
Ethical approval: All procedures performed in studies involving human par- duration greater than 24 hours.
ticipants were in accordance with the ethical standards of the institutional
Interventions: Patients were randomized to usual care, ICU diary,
and/or national research committee and with the 1964 Helsinki declara-
tion and its later amendments or comparable ethical standards. psychoeducation, or both ICU diary and psychoeducation.
Clinical Trials Registration: Registration Number NCT02067559. Measurements and Main Results: Our primary objective was to
Supplemental digital content is available for this article. Direct URL citations determine feasibility measured by enrollment/mo. Secondary out-
appear in the printed text and are provided in the HTML and PDF versions comes included acceptability of the ICU diary intervention and
of this article on the journal’s website (http://journals.lww.com/ccmjournal).
psychological distress, including patients’ memories 1 week post
Supported, in part, by a grant from the Manitoba Medical Service Foundation
ICU using the ICU Memory Tool, posttraumatic stress disorder
(grant number 8-2014-07) to Dr. Blouw. Supported, in part, by Canadian Insti-
tutes of Health Research Foundation grant number 333252 (to Dr. Sareen). (Impact of Events Scale-Revised), depression, and anxiety symp-
Copyright © 2018 by the Society of Critical Care Medicine and Wolters toms (Hospital Anxiety and Depression Scale) 30 and 90 days
Kluwer Health, Inc. All Rights Reserved. post ICU. Over 3.5 years, we enrolled 58 patients, an average
DOI: 10.1097/CCM.0000000000003367 of 1.9 participants/mo. Families and healthcare providers wrote

Critical Care Medicine www.ccmjournal.org 1

Copyright © 2018 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Kredentser et al

a mean of 3.2 diary entries/d (sd, 2.9) and indicated positive atti- METHODS
tudes and low perceived burden toward ICU diary participation.
A majority of patients reported distressing memories of their ICU
Study Enrollment
stay. Those who received the diary intervention had significantly
We completed a single-center, pilot randomized controlled
trial at a tertiary, 10-bed mixed medical-surgical ICU in Win-
lower median Hospital Anxiety and Depression Scale anxiety (3.0
nipeg, MB, Canada.
[interquartile range, 2–6.25] vs 8.0 [interquartile range, 7–10];
Patients were eligible for enrollment if they were mechani-
p = 0.01) and depression (3.0 [interquartile range, 1.75–5.25] vs
cally ventilated, greater than 17 years old, and English speak-
5.0 [interquartile range, 4–9]; p = 0.04) symptom scores at 90
ing. Enrollment occurred within 72 hours of the current ICU
days than patients who did not receive a diary.
admission, with an ICU length of stay greater than 72 hours
Conclusions: ICU diaries are a feasible intervention in a tertiary
and duration of mechanical ventilation greater than 24 hours
Canadian ICU context. Preliminary evidence supports the efficacy
as predicted by the ICU treatment team. Exclusion criteria
of ICU diaries to reduce psychological morbidity following dis-
included no caregiver/family available, terminal illness with life
charge. (Crit Care Med 2018; XX:00–00)
expectancy of less than 6 months, preexisting known cognitive
Key Words: anxiety; depression; intensive care; intensive care unit
impairment (per chart), or reason for admission involving sui-
diaries; postintensive care syndrome; posttraumatic stress disorder
cide attempt/intentional toxic overdose, meningitis/encephali-
tis, status epilepticus, anoxic encephalopathy, traumatic brain
injury, or coma due to another etiology.

M
ore patients are surviving their critical illnesses (1–3); Consent for study enrollment was provided initially by a
however, survivors often experience significant mor- proxy/substitute decision maker and subsequently by patient
bidity after an ICU stay (4), referred to as “postint- participant once they regained capacity. The study was approved
ensive care syndrome” (5), which includes high rates of anxiety by the St. Boniface Hospital Research Review Committee and
(25–46%), depression (43%), and posttraumatic stress disorder the University of Manitoba research ethics boards.
(PTSD) (20%) (6–8). PTSD is twice as prevalent among ICU
Randomization
survivors compared with the general population (6, 7, 9–11).
Patients were randomized 1:1 in three blocks of 20, using com-
Underlying disease, mechanical ventilation, delirium, restraint,
puter-generated randomization to one of four arms at the time
and sedative use may prevent the formation of cohesive memo-
of enrollment: 1) usual care, 2) ICU diary, 3) psychoeducation,
ries, resulting in a fractured narrative of the trauma, known to be
and 4) ICU diary and psychoeducation (ICU diary + psycho-
associated with PTSD and other mental health syndromes (4, 12).
education). Treatment allocation was concealed using sequen-
An empirically supported treatment for PTSD is the cre- tially numbered, opaque envelopes (26).
ation of a narrative to consolidate traumatic memories. In
an effort to prevent PTSD in ICU survivors, some ICUs have Study Interventions
implemented ICU diaries (13, 14), bedside notebooks where Prior to study commencement, approximately 75% of the ICU
daily events are recorded by healthcare staff and patients’ rela- nurses attended an educational in-service on ICU diaries.
tives (13–16). This chronological record of illness may aid in In the “usual care” arm, no follow-up psychological support
consolidating memories into an autobiographical narrative or education was provided.
(13, 17) and has evidence for reducing PTSD, anxiety, and In the “ICU diary” arm, diaries were created using a bound
depression in patients and their caregivers, with PTSD preva- empty journal. Instructions for diary use were provided at the
lence reduced as much as 50% through early ICU diary use time of randomization and were available for the duration of
(12, 18–20). The majority of this research has been conducted the ICU admission (14). Family members and ICU staff were
in Europe and has typically used a usual care comparison group, invited to write in the ICU diary at any time. Research staff
although other interventions, including a self-help rehabilita- took photographs of the patient using an instant camera.
tion manual (21) and nurse-led follow-up clinics (22), show The diary was stored securely for 30 days post ICU discharge
benefit in reducing post-ICU psychopathology (23). or until a participant achieved capacity to provide consent
It is unclear whether ICU diaries themselves reduce symp- (whichever occurred later), at which time the diaries were
tomatology or if a psychoeducational follow-up after ICU dis- offered to participants or to their family in the case of patient
charge might have similar or added benefit, with potentially less death. Diaries were reviewed with each participant by a mem-
effort for staff and families. A successful diary program requires ber of our research team.
support and acceptance from ICU staff and families. Social, cul- In the “psychoeducation” arm, participants and their fol-
tural, and practical factors are relevant, requiring testing in North low-up physicians were given a brochure 30 days after ICU
American contexts, with research needed before ICU diaries can discharge or after return of the ability to provide consent,
be recommended as a standard of care (24, 25). The goal of the whichever occurred later. The brochure described common
present research was to assess the feasibility and acceptability of ICU procedures and experiences (e.g., delirium); symptoms
ICU diaries in a Canadian context and to provide foundational of PTSD, anxiety and depression; and follow-up/emergency
information comparing ICU diaries with psychoeducation, to resources (Supplemental Digital Content 1, http://links.lww.
inform a larger, multicenter trial. com/CCM/D869).

2 www.ccmjournal.org XXX 2018 • Volume XX • Number XXX

Copyright © 2018 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
 Clinical Investigation

In the “ICU diary + psychoeducation arm,” participants subscale scores greater than or equal to 8 (suggestive of signifi-
received both the ICU diary and psychoeducation interven- cant levels of anxiety or depression) (31) at 30 days and 90 days
tions as above. post-ICU discharge.

Data Collection Statistical Analyses

We assessed severity of illness using Sequential Organ Failure We report continuous variables using means and sds or medi-
Assessment scores for the first 24 hours of ICU admission (27). ans and interquartile range (IQR), depending on distribution.
We assessed delirium (Confusion Assessment Method for the We explored between-group differences in categorical and con-
ICU) (28), continuous sedative infusion use, and level of seda- tinuous outcome measures using chi-square and Kruskal-Wal-
tion (Richmond Agitation-Sedation Scale [RASS]) (29) daily. lis/Mann-Whitney U tests, respectively. Analysis of all outcomes
We defined coma as a RASS score of less than or equal to –3 was completed according to intention to treat. We analyzed the
for greater than 6 hours of the calendar day. The ICU has stan- data using SAS software Version 9.3 (37) and SPSS Version 24
dardized daily awakening and spontaneous breathing trials (38). All tests of significance reflect a two-sided alpha of 0.05.
protocols.
Following ICU discharge, we collected patient memories RESULTS
using the ICU Memory Tool (ICU-MT) (30) at 1 week and Between May 30, 2014, and November 30, 2016, 1,242 ICU
validated self-report measures of depression, anxiety (Hospital patients were screened; 152 patients met eligibility criteria
Anxiety and Depression Scale [HADS]) (31), and PTSD (13%) (Fig. 1). Common reasons for ineligibility were no
(Impact of Event Scale-Revised [IES-R]) (32) symptoms at mechanical ventilation (n = 560), mechanical ventilation
both 30 and 90 days. Measurement instruments were collected duration less than 24 hours (n = 275), postcardiac arrest
first, followed by intervention-specific questions, as such, (n = 154), life expectancy < 6 mo (n = 89), predicted ICU stay
outcome assessors were not blinded to treatment allocation. less than 72 hours (n = 81), and preexisting cognitive impair-
Thirty-day outcomes were routinely assessed prior to provid- ment (n = 58). Patients’ sociodemographic and illness-related
ing the intervention(s). information is presented in Table 1.

Outcomes Primary Outcome

Primary Outcome: Feasibility of Enrollment and Interven- One hundred families were approached for study enroll-
tion Delivery. The primary feasibility outcome was the abil- ment, and consent was obtained for 58 patients (58% consent
ity to enroll and retain patients in the study, to determine a rate). Approximately one quarter of families provided a rea-
realistic enrollment rate per month to power a larger trial. We son for declined consent (Fig. 1). Overall, 64% of participants
anticipated being able to recruit two patients per month, with (37/58) completed the study, 21% died (n = 12), 10% withdrew
a completion rate of 60%. We also tracked protocol adherence (n = 6), 3% were lost to follow-up (n = 2), whereas one
and anticipated being able to conduct follow-up interviews remained in the ICU. The enrollment rate was 1.9 participants/
at 7, 30, and 90 days post ICU discharge, with delivery of the mo. Median (IQR) days from ICU discharge to the 7-, 30-, and
intervention(s) at 30-day follow-up (33–35). 90-day follow-ups were 8 (7–11), 36 (31–42), 97 (92–105),
Secondary Outcomes: Acceptability. An additional feasibil- respectively. Diaries and psychoeducation brochures were
ity outcome was diary participation, with the goal of greater claimed on a median of 40 days (IQR, 32–50 d) and 36 days
than or equal to one diary entry per day after enrollment and (IQR, 29–44 d) post ICU discharge, respectively.
the attitude and perceived burden toward the diary by family/ Ninety-three percent of participants randomized to the diary
caregivers and ICU staff. At the time of diary entry comple- arm claimed their diary (28/30); 24 were ICU survivors, and four
tion, all healthcare workers and family members were invited were family members of participants who died. The two unclaimed
to complete a diary evaluation form (created by our research diaries belonged to individuals who died and whose families were
group), which included estimated entry time (min) and ratings lost to follow-up. Ninety-six percent of participants randomized to
of perceived burden and attitude toward completing the entry psychoeducation received the brochure (22/23; one refusal), which
(Supplemental Digital Content 2, http://links.lww.com/CCM/ was faxed to 21 of 23 follow-up physicians (91%).
D870). We hypothesized low burden and positive attitude
toward diaries by both families and staff. Secondary Outcomes
Psychological Distress. As this was a feasibility study, it was Acceptability. Participants’ families and friends wrote a mean
not powered to detect differences between treatment groups. 1.7 (sd, 1.7) diary entries per day, whereas staff (nurses, phy-
However, we undertook exploratory analyses to determine sicians, and allied health) wrote a mean 1.5 (sd, 1.4) entries
levels of psychological distress—symptoms of anxiety, depres- per day (Table 2). Diaries were well accepted by family and
sion, and PTSD—across and within treatment arms between healthcare providers, who reported low burden and positive
30 and 90 days. We also examined the proportion of patients attitudes toward the intervention, with 48% of entries taking
with IES-R mean item scores greater than or equal to 1.6 (sug- less than 5 minutes (Table 2).
gestive of significant trauma-related distress in an acutely ill Most ICU diary recipients (85%; 22/26) read their diary
population) (36) and the proportion of patients with HADS again after the initial meeting with the study nurse, with a

Critical Care Medicine www.ccmjournal.org 3

Copyright © 2018 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Kredentser et al

frequency of one to two (27%; 6/22), three to five (53%; 12/22), that the most helpful aspect of the diary intervention was
six to nine (5%; 1/22), and 10–20 (14%; 3/22) times within 90 reviewing the diary with study personnel at the time of receipt.
days after ICU. Many participants reported that their family or Psychological Distress. Forty-five participants (96%) com-
friends also read the diary (81%; 21/26). Participants reported pleted the ICU-MT 1 week post ICU. Frequently, patients had

Figure 1. Consolidated Standards of Reporting Trials diagram.

TABLE 1. Patient Demographic and Illness Characteristics by Treatment Group

Diary +
Usual Care Psychoeducation Diary Psychoeducation Total
Characteristics (n = 14) (n = 14) (n = 15) (n = 15) (n = 58) p

Patient characteristics
Age, yr, mean (sd) 49.9 (16.9) 54.3 (17.7) 59.3 (15.5) 55.7 (14.0) 54.9 (16.0) 0.5
Male, n (%) 7 (50) 10 (71) 7 (47) 11 (73) 35 (60) 0.3
Prior ICU admission, n (%) 3 (21) 3 (21) 2 (13) 2 (13) 10 (17.2) 0.9
Preexisting comorbidities, n (%)
  Neurologic disease 4 (29) 5 (35) 3 (20) 3 (20) 13 (22.4) 0.9
  Psychiatric disease 1 (7) 1 (7) 2 (13) 2 (13) 6 (10.3) 0.9
   Cardiac New York Heart 0 0 0 0 0 0
Association III/IV
  Respiratory disease 0 1 (7) 2 (13) 2 (13) 5 (8.6) 0.5
  Renal disease 1 (7) 1 (7) 4 (26) 0 5 (8.6) 0.3
  Liver disease 0 0 0 0 0 0
  Drug abuse 5 (36) 2 (14) 1 (7) 4 (27) 12 (20.6) 0.2
(Continued )

4 www.ccmjournal.org XXX 2018 • Volume XX • Number XXX

Copyright © 2018 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
 Clinical Investigation

TABLE 1. (Continued). Patient Demographic and Illness Characteristics by Treatment Group

Diary +
Usual Care Psychoeducation Diary Psychoeducation Total
Characteristics (n = 14) (n = 14) (n = 15) (n = 15) (n = 58) p

Acute illness characteristicsa

Sequential Organ Failure 9.2 (3.5) 8.6 (3.8) 9.1 (3.7) 8.2 (4.8) 8.8 (3.9) 0.9
Assessment scoreb, mean (sd)
Primary reason for admission, n (%)
  Respiratory failure 13 (93) 10 (71) 11 (73) 10 (67) 44 (76) 0.4
  Cardiovascular 1 (7) 1 (7) 3 (20) 2 (13) 7 (12) 0.7
  Sepsis 0 3 (21) 1 (7) 3 (20) 7 (12) 0.2
ICU mortality, n (%) 4 (29) 2 (14) 2 (13) 1 (7) 9 (15.5) 0.4
Hospital mortality, n (%) 4 (29) 3 (21) 2 (13) 2 (29) 11 (19) 0.7
Sedation use duration, d, median (IQR)
   Duration of continuous sedation 9.5 (8.5–12) 2.5 (0–12.25) 5.0 (2–7) 5.0 (3–10) 5 (2.75–11) 0.01
   Duration of continuous 6.5 (0.75–9.5) 0.0 (0–3.5) 2.0 (0–6) 2.0 (0–6) 2 (0–7) 0.2
benzodiazepine
   Duration of continuous narcotic 9.0 (6.5–11.25) 1.0 (0–8.5) 5.0 (2–7) 5.0 (2–9) 5 (1–10) 0.03
   Duration of continuous propofol 6.0 (1–8.25) 1.5 (0–3.25) 1.0 (0–3) 3.0 (0–5) 2 (0–5) 0.05
Ventilator duration, d, median (IQR) 14.0 (8.5–22) 6.0 (3.75–18.5) 7.0 (3–11) 6.0 (4–11) 8 (4.75–14.5) 0.05
Coma duration, d, median (IQR) 8.0 (5.5–13.25) 2.0 (0.75–4.25) 4.0 (0–7) 4.0 (1–5) 4 (1–8) 0.02
Delirium duration, d, median (IQR) 2.0 (0.75–5.25) 2.0 (0–4) 0.0 (0–3) 2.0 (1–5) 2 (0–4) 0.9
ICU length of stay, d, median (IQR) 19.0 (9–24.5) 10.5 (6.75–23.75) 9.0 (7–13) 11.0 (6–18) 12 (7–20.5) 0.2
Hospital length of stay, d, median (IQR) 24.0 (20–47.5) 30.0 (12–44.5) 21.0 (12–61) 23.0 (9–48) 24 (12–49) 0.8
IQR = interquartile range.
a
There were no significant differences in illness characteristics between those who completed the study and those who withdrew/were lost to follow-up.
b
Within first 24 hr of admission.
Boldface values indicate statistically significant p ≤ 0.05.

TABLE 2. Use of ICU Diary and Evaluation Survey Results by Family Members and
Healthcare Providers
Family Friend Nurse Physician Allied Health Total
Diary Entries (n = 78) (n = 4) (n = 98) (n = 12) (n = 6) (n = 198)

Time per entry, min, n (%)

1–5 17 (21.8) 3 (75.0) 64 (65.3) 9 (75.0) 2 (33.3) 95 (48.0)
5–10 28 (35.9) 1 (25.0) 31 (31.6) 2 (16.7) 4 (48.0) 66 (33.3)
15–20 15 (19.2) 0 3 (3.1) 0 0 18 (9.1)
> 20 18 (23.1) 0 0 1 (8.3) 0 19 (9.6)
Number of entries per day, mean 1.70 (1.74) (0–8) 1.50 (1.40) (0.6–8) 3.20 (2.90)
(sd) (minimum–maximum) (1.15–16)
Perceived burdena, mean (sd) 0.69 (1.39) 2.0 (2.45) 1.6 (0.19) 1.75 (1.48) 1.0 (0.63) 1.25 (1.7)c
Attitudeb, mean (sd) 9.57 (1.46) 7.33 (2.52) 8.12 (2.17) 8.25 (1.49) 9.0 (1.41) 8.71 (2.0)d
“How burdensome did you find it to complete the last entry?” (0 = not burdensome to 10 = highly burdensome).
a

b
“What is your attitude toward completing diary entries?” (0 = very negative to 10 = very positive).
c
p < 0.01.
d
p < 0.001.

Critical Care Medicine www.ccmjournal.org 5

Copyright © 2018 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Kredentser et al

no memory of being admitted to the hospital (40%; 18/45) The distribution of IES-R and HADS scores at 30 and
or pre-ICU hospitalization (58%; 18/45). Although almost all 90 days post ICU by treatment group is shown in Figure 2.
participants (95.6%) recalled some aspect of their ICU stay, Participants in the ICU diary + psychoeducation interven-
only 15.5% (7/45) reported remembering their whole ICU stay tion had significant decreases in both median HADS depres-
clearly. Common memories included family visiting (80%), sion (5.0 [IQR, 3–7] vs 2.0 [IQR, 1–3]; p = 0.03) and IES-R
confusion (75.6%), breathing tube (66.7%), voices (66.7%), (PTSD) scores (1.0 [IQR, 0.5–1.4] vs 0.4 [IQR, 0.1–0.7];
faces (62.2%), discomfort (57.8%), ICU team rounds (57.8%), p = 0.02) between 30 and 90 days. Across treatment arms,
lights (55.6%), feeling “down” (53.3%), panic (46.7%), hallu- there was a significant difference between 90-day median
cinations (42.2%), feeling that people were trying to hurt them HADS anxiety scores (usual care 7.5 [IQR, 6–9], psycho-
(31.1%), and nightmares (28.9%). education 8.0 [IQR, 8–11], diary 3.0 [IQR, 2–5], and

Figure 2. Boxplot showing median, 25th, and 75th percentiles of Impact of Event Scale-Revised (IES-R) scores and Hospital Anxiety and Depression
Scale (HADS) scores in patients 30 and 90 d post ICU. Whiskers show minimum and maximum responses. p values above the boxplots are 30 and 90 d
comparisons within each group, whereas p values in the top right reflect differences between the groups at each time using the Kruskal-Wallis test.

6 www.ccmjournal.org XXX 2018 • Volume XX • Number XXX

Copyright © 2018 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
 Clinical Investigation

diary + psychoeducation 3.0 [IQR, 2–7]; p = 0.05) and trend- than our 30 days post-ICU discharge target, and this timeline
ing differences between 90-day median HADS depression may need to be reassessed for a future larger trial.
scores (usual care 4.5 [IQR, 2–7], psychoeducation 5.0 [IQR, Our rate of recruitment of eligible patients (58%) was
4–9]), diary 3.0 [IQR, 3–6], and diary + psychoeducation 2.0 slightly lower and attrition (13%) slightly higher in our study
[IQR, 1–3]; p = 0.06). than it has been in previous trials of diaries (18, 20) These
Across treatment arms, 20.5% and 32.4% of participants studies used a waitlist-control design, where all participants
had significant anxiety symptoms, whereas 20.5% and 13.5% eventually received the diary intervention. Reasons for refusal
had significant depressive symptoms at 30 and 90 days, may reflect reluctance to participate in research in general, a
respectively. Significant PTSD symptoms were present in population for whom diaries could be harmful (40) or a “pref-
12.8% of participants at 30 days and 13.5% of participants erence effect” (41) related to desire to be placed in a specific
at 90 days. In a post hoc analysis, we explored the effect of treatment arm, and these issues must be considered in a future
the diary by comparing participants who had received the trial.
diary intervention with those who did not and also com- All ICU survivors randomized to diary intervention com-
pared participants who were exposed to psychoeducation pleted the study, suggesting increased engagement and per-
with those who were not. Participants who received the diary ceived high value of the intervention. Attrition in fact only
intervention had significantly lower median HADS anxiety occurred in the usual care and psychoeducation arms. This
(3.0 [IQR, 2–6.25] vs 8.0 [IQR, 7–10]; p = 0.01) and depres- may reflect higher levels of distress in participants who with-
sion (3.0 [IQR, 1.75–5.25] vs 5.0 [IQR, 4–9]; p = 0.04) scores drew and may be a function of avoidance that characterizes
at 90 days. Figure 3 shows that the diary intervention was PTSD and is often present in depression and anxiety (42),
associated with a lower pro-
portion of significant anxiety
symptoms at 90 days (19.2%
vs 63.6%; p = 0.008), whereas
the proportion of patients
with significant PTSD symp-
toms at 90 days was lower
in those who received psy-
choeducation (0% vs 26.3%;
p = 0.04).

DISCUSSION
This single-center, pilot ran-
domized controlled trial
comparing ICU diaries and
psychoeducation for preven-
tion of PTSD in ICU survi-
vors informs the feasibility
of a larger randomized con-
trolled trial. We enrolled
1.9 patients per month in a
10-bed ICU, a recruitment
rate in line other ICU trials
in Canada (39). Despite the
challenges of high ICU mor-
tality and 90-day post-ICU
follow-up protocol, we had a
study completion rate above
our goal of 60%. ICU staff
and family members were
engaged in the diary interven-
tion, entering more than three
entries per day, well above our
goal of greater than or equal to
one. The delivery of the diary Figure 3. Proportion of patients reporting significant mental health symptoms as defined by Impact of Event
Scale-Revised score greater than or equal to 1.6 or Hospital Anxiety and Depression Scale (HADS) subscale
and psychoeducation inter- scores greater than or equal to 8. A, Compared participants exposed to diary versus no diary. B, Compares
ventions took slightly longer participants exposed to psychoeducation versus no psychoeducation.

Critical Care Medicine www.ccmjournal.org 7

Copyright © 2018 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Kredentser et al

but this is speculative. This study also echoes previous research REFERENCES
showing that unpleasant memories of the ICU (e.g., endo- 1. Zimmerman J, Kramer A, Knaus W: Changes in hospital mortality for
United States intensive care unit admissions from 1988 to 2012. Crit
tracheal tubes, harm, hallucinations) and subsequent psy- Care 2013; 17:R81
chological distress are common after critical illness (19, 22). 2. Moran J, Bristow P, Solomon P, et al: Mortality and length-of-stay out-
Family, friends, and healthcare providers had positive atti- comes, 1993–2003, in the binational Australian and New Zealand
tudes toward, and low perceived burden with, the interven- intensive care adult patient database. Crit Care Med 2008; 31:46–
61
tion, suggesting high levels of acceptability in the Canadian 3. Hutchings A, Durand MA, Grieve R, et al: Evaluation of modernisation
setting. of adult critical care services in England: Time series and cost effec-
Rates of significant depression, anxiety, and PTSD in tiveness analysis. BMJ 2009; 339:b4353
our sample were higher than in the general population (43, 4. Bienvenu OJ, Jones C, Hopkins RO: Psychological and Cognitive
Impact of Critical Illness. New York, NY, Oxford University Press,
44), at both 30- and 90-day follow-ups, demonstrating the 2017
importance of prevention and intervention in this popula- 5. Needham DM, Davidson J, Cohen H, et al: Improving long-term out-
tion. Additionally, although psychological outcomes must be comes after discharge from intensive care unit: Report from a stake-
holders conference. Crit Care Med 2012; 40:502–509
interpreted with caution as this pilot study was not powered
6. Nikayin S, Rabiee A, Hashem MD, et al: Anxiety symptoms in survivors
to assess between-group differences, promising results include of critical illness: A systematic review and meta-analysis. Gen Hosp
significant reductions between 30 and 90 days in depression Psychiatry 2016; 43:23–29
and PTSD symptomatology with combined diary and psy- 7. Parker AM, Sricharoenchai T, Raparla S, et al: Posttraumatic stress
disorder in critical illness survivors: A metaanalysis. Crit Care Med
choeducation interventions and significantly lower rates of 2015; 43:1121–1129
depression and anxiety scores at 90 days with exposure to the 8. Rabiee A, Nikayin S, Hashem MD, et al: Depressive symptoms after
diary intervention. Additionally, there did not appear to be any critical illness: A systematic review and meta-analysis. Crit Care Med
adverse effects of the diaries on patients’ psychological distress, 2016; 44:1744–1753
an important factor given the relatively limited evidence for 9. Myhren H, Ekeberg Ø, Stokland O: Health-related quality of life
and return to work after critical illness in general intensive care unit
the intervention (45). Baseline rates of sedation, an important patients: A 1-year follow-up study. Crit Care Med 2010; 38:1554–
risk factor for subsequent psychological distress (46), were 1561
significantly higher in the usual care arm. We did not see sig- 10. Davydow DS, Gifford JM, Desai SV, et al: Posttraumatic stress disor-
der in general intensive care unit survivors: A systematic review. Gen
nificant differences in distress across treatment arms at 30 days Hosp Psychiatry 2008; 30:421–434
and suspect that this baseline imbalance is a reflection of our 11. Wade DM, Howell DC, Weinman JA, et al: Investigating risk factors
small sample size. for psychological morbidity three months after intensive care: A pro-
With our enrollment and study completion rate, a sub- spective cohort study. Crit Care 2012; 16:R192
12. Jones C, Bäckman C, Capuzzo M, et al; RACHEL group: Intensive
sequent multicentered trial would require 120 ICU beds to care diaries reduce new onset post traumatic stress disorder follow-
enroll 800 patients over 3 years. This study would be pow- ing critical illness: A randomised, controlled trial. Crit Care 2010;
ered to detect a 50% reduction in PTSD, assuming a baseline 14:R168
PTSD rate of 13%. This recruitment target is similar to a large 13. Bäckman CG, Walther SM: Use of a personal diary written on the ICU
during critical illness. Intensive Care Med 2001; 27:426–429
ongoing ICU diary trial in France (47).
14. icu-diary.org. 2012. Available at: http://www.icu-diary.org/diary/start.
Limitations of the present study include a small sample html. Accessed April 1, 2014
size and use of self-report symptom measures, which are 15. Egerod I, Storli SL, Åkerman E: Intensive care patient diaries in Scan-
screening instruments and not diagnostic tools, rather than dinavia: A comparative study of emergence and evolution. Nurs Inq
2011; 18:235–246
diagnostic interviews. Many ICU patients did not meet our
16. Combe D: The use of patient diaries in an intensive care unit. Nurs
eligibility criteria, which may limit generalizability of inter- Crit Care 2005; 10:31–34
vention to general ICU population. In a larger trial, it will 17. Roulin MJ, Hurst S, Spirig R: Diaries written for ICU patients. Qual
be important to blind outcome assessors to treatment arm. Health Res 2007; 17:893–901
Strengths of the trial include the use of a randomized con- 18. Åkerman E, Ersson A, Fridlund B, et al: Preferred content and useful-
ness of a photodiary as described by ICU-patients–a mixed method
trolled trial design and inclusion of a psychoeducation analysis. Aust Crit Care 2013; 26:29–35
intervention. 19. Garrouste-Orgeas M, Coquet I, Périer A, et al: Impact of an intensive
In summary, to our knowledge, this is the first study in care unit diary on psychological distress in patients and relatives*.
Canada to determine the feasibility of an ICU diary inter- Crit Care Med 2012; 40:2033–2040
20. Knowles RE, Tarrier N: Evaluation of the effect of prospective patient
vention in our healthcare system. Patients, families, and staff diaries on emotional well-being in intensive care unit survivors: A ran-
were engaged and supportive of ICU diaries. The significant domized controlled trial. Crit Care Med 2009; 37:184–191
reduction in depression and anxiety symptoms at 90 days 21. Jones C, Skirrow P, Griffiths RD, et al: Rehabilitation after critical ill-
post ICU discharge in those who received the diary (with or ness: A randomized, controlled trial. Crit Care Med 2003; 31:2456–
2461
without psychoeducation) is compelling and warrants further 22. Schandl A, Bottai M, Hellgren E, et al: Gender differences in psycho-
investigation of the two interventions in a larger trial. Further logical morbidity and treatment in intensive care survivors–a cohort
research will also allow for an extension of the study paradigm study. Crit Care 2012; 16:R80
to examine the impact of diaries on mental health and coping 23. Roberts MB, Glaspey LJ, Mazzarelli A, et al: Early interventions for
the prevention of posttraumatic stress symptoms in survivors of
outcomes in family members and staff, and patients’ experi- critical illness: A qualitative systematic review. Crit Care Med 2018;
ences of the intervention. 46:1328–1333

8 www.ccmjournal.org XXX 2018 • Volume XX • Number XXX

Copyright © 2018 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
 Clinical Investigation

24. Ullman AJ, Aitken LM, Rattray J, et al: Diaries for recovery from critical 36. Bienvenu OJ, Gellar J, Althouse BM, et al: Post-traumatic stress dis-
illness. Cochrane Database Syst Rev 2014; 12:Cd010468 order symptoms after acute lung injury: A 2-year prospective longitu-
25. Bienvenu OJ, Gerstenblith TA: Posttraumatic stress disorder phenom- dinal study. Psychol Med 2013; 43:2657–2671
ena after critical illness. Crit Care Clin 2017; 33:649–658 37. SAS Institute Inc: SAS/ACCESS® 9.3 Interface to ADABAS: Refer-
26. Dettori J: The random allocation process: Two things you need to ence. Cary, NC, SAS Institute Inc, 2012
know. Evid Based Spine Care J 2010; 1:7–9 38. IBM Corp: Released 2016. IBM SPSS Statistics for Mac, Version
27. Vincent JL, Moreno R, Takala J, et al: The SOFA (Sepsis-related 24.0. Armonk, NY, 2016
Organ Failure Assessment) score to describe organ dysfunction/fail- 39. Burns KE, Zubrinich C, Tan W, et al; Canadian Critical Care Trials
ure. On behalf of the Working Group on Sepsis-Related Problems Group: Research recruitment practices and critically ill patients. A
of the European Society of Intensive Care Medicine. Intensive Care multicenter, cross-sectional study (the Consent Study). Am J Respir
Med 1996; 22:707–710 Crit Care Med 2013; 187:1212–1218
28. Ely EW, Margolin R, Francis J, et al: Evaluation of delirium in criti- 40. Aitken LM, Rattray J, Hull A, et al: The use of diaries in psychological
cally ill patients: validation of the Confusion Assessment Method recovery from intensive care. Crit Care 2013; 17:253
for the Intensive Care Unit (CAM-ICU). Crit Care Med 2001; 41. King M, Nazareth I, Lampe F, et al: Conceptual framework and sys-
29:1370–1379 tematic review of the effects of participants’ and professionals’ prefer-
29. Sessler CN, Gosnell MS, Grap MJ, et al: The Richmond Agitation- ences in randomised controlled trials. Health Technol Assess 2005;
Sedation Scale: Validity and reliability in adult intensive care unit 9:1–186, iii
patients. Am J Respir Crit Care Med 2002; 166:1338–1344 42. Griffiths RD: Sedation, delirium and psychological distress: Let’s not
30. Jones C, Humphris G, Griffiths RD: Preliminary validation of the ICUM be deluded. Crit Care 2012; 16:109
tool: A tool for assessing memory of the intensive care experience. 43. Van Ameringen M, Mancini C, Patterson B, et al: Post-traumatic stress
Clinical Intensive Care 2000; 11:251–255 disorder in Canada. CNS Neurosci Ther 2008; 14:171–181
31. Zigmond AS, Snaith RP: The hospital anxiety and depression scale. 44. Pearson C, Janz T, Ali J: Mental and substance use disorders in Can-
Acta Psychiatr Scand 1983; 67:361–370 ada. Health at a Glance. September. Statistics Canada Catalogue
32. Weiss DS, Marmar CR: The impact of event scale - revised. In: no. 82-624-X, 2013
Assessing Psychological Trauma and PTSD. Wilson J, Keane TM 45. Bolton N: Diaries for recovery from critical illness. Clin Nurse Spec
(Eds). New York, NY, Guilford, 1996, pp 399–411 2016; 30:17–18
33. Hertzog MA: Considerations in determining sample size for pilot stud- 46. Girard TD, Shintani AK, Jackson JC, et al: Risk factors for post-traumatic
ies. Res Nurs Health 2008; 31:180–191 stress disorder symptoms following critical illness requiring mechanical
34. Lancaster GA, Dodd S, Williamson PR: Design and analysis of pilot ventilation: A prospective cohort study. Crit Care 2007; 11:R28
studies: Recommendations for good practice. J Eval Clin Pract 2004; 47. Garrouste-Orgeas M, Flahault C, Fasse L, et al: The ICU-Diary study:
10:307–312 Prospective, multicenter comparative study of the impact of an ICU
35. Thabane L, Ma J, Chu R, et al: A tutorial on pilot studies: The what, diary on the wellbeing of patients and families in French ICUs. Trials
why and how. BMC Med Res Methodol 2010; 10:1 2017; 18:542

Critical Care Medicine www.ccmjournal.org 9

Copyright © 2018 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.