Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Garrouste-Orgeas, 2017

Type of study:

RCT

Setting and country:

12-bed medical-surgical ICU of a tertiary hospital in Paris,

France

“The ICU admits about

800 patients each year and has a nurse-to-patient ratio of 1:2.5,

four senior and four junior physicians on every weekday from

8.30 am to 6 pm, and one senior and one junior physician at

night from 6 pm to 9:30 am and on weekends. Families can visit

around the clock.”

Funding and conflicts of interest: in part by non-profit research organization Fondation de France. One of the authors disclosed lecturing and board participation for farmaceutical companies, and his institution received funding (research grants) from two companies.

Inclusion criteria:

family member:

* close relative who visited the patient and ranked highest in the following order: spouse, grown child, sibling, and other

* family of a ventilated patient for over 48 hours in the ICU

Exclusion criteria:

family member:

* not fluent in French

* Conflict with the ICU team and admission

* refused to participate in the study

* refused recording of their semistructured interviews.

* family members of patients who were organ donors.

N total at baseline:

Intervention: 50, 42 evaluable.

Control: 50, 44 evaluable.

Important prognostic factors:

(characteristics of family members)

Age, median (range)

I: 61 (49-68)

C:56 (45-64)

Sex:

I: 38% M

C: 37% M

Relationship to patient, %:

I: Spouse 39, child 31, parent 15, other 15.

C: Spouse 33, child 47, parent 12, other 9.

History of receiving psychological or psychiatric support,%

I: 40

C: 29

History of treatment with psychoactive agents, %:

I: 3

C:9

Groups comparable at baseline? Not statistically, but there are some differences in percentages.

Furthermore, baseline distress levels not assessed.

Family conferences with proactive participation of a nurse in family conferences.

Each conference was led by

a physician. A conference was held after admission, then on days 3 and 7, and finally once a week until discharge. Additional conferences were held if the patient’s condition worsened or was expected to die (end-of-life conference). The physicians and nurses participating in the conferences were those in charge of the patient on that day.

Physicians could participate in conferences

in both study groups.

All physicians and nurses received training, that included role playing.

The NURSE method was used for bad news communication, and the ask-tell-ask concept was applied in all conferences. VALUE strategy was followed in end-of-life conferences.

At the end of the first conference, family was handed an information brochure.

The physician and nurse each focussed on specific topics. The nurse explained the organization of the ICU, (visitation policy and ICU diary), described the function of machines, tubes and catheters and verified comprehension by the family of the information delivered by the physician.

Family conferences without proactive participation of a nurse in family conferences.

Length of follow-up:

3 months

Loss-to-follow-up:

Intervention, n(%): 7 (14)

n=6 withdrew consent

n=1 loss to follow-up

Control, n(%): 6 (12)

n=5 withdrew consent

n=1 loss to follow-up

Incomplete outcome data:

see loss to follow-up

No statistical differences were found between patients included or analysed in the intervention and control groups.

1. Anxiety,

Evaluated using the HADS

90 days, median, [interquartile range]

I: 4 [1-9]

C: 8 [4.5-12]

p=0.01

HADS score ≥ 8, n(%)

I: 14 (33.3)

C: 23 (52.3)

p=0.08

“Although the intervention group had significantly lower anxiety and depression subscale scores, the prevalence of severe anxiety

and depression symptoms (HADS subscale scores > 8) did not differ significantly between the two groups.”

2. Depression Evaluated using the HADS

90 days, median, [interquartile range]

I: 2 [0-6]

C: 5.5 [1-11.5]

p=0.04

HADS score ≥ 8, n(%)

I: 10 (23.8)

C: 17 (38.6)

p=0.14

See also ‘Anxiety’

3. PTSD

Evaluated using the IES-R

90 days, median, [interquartile range]

I: 21 [9-33]

C: 24 [12.5-45]

p=0.24

IES-R score ≥ 22, n(%)

I: 21 (50)

C: 23 (52.3)

p=0.83

No significant between-group differences were found.

4. Satisfaction

Not reported

5. Quality of life

Not reported

Family members participated in a semistructured interview at patient death or ICU discharge.

“Between-group comparisons relied on the nonparametric

Fisher exact test and Kruskal-Wallis test.”

Sample size estimation: 49 participants in each group needed.

There were 96 conferences in the intervention and 92 in the control group. There was no significant between group difference in conference contents on day 1 and 3.

Nielsen, 2020

Type of study:

multicenter, block-randomized, single-blinded, controlled trial

Setting and country:

four medical-surgical ICUs at two university hospitals and two regional hospitals

Funding and conflicts of interest: Aarhus University, Denmark

Inclusion criteria:

patients:

* ≥18 years of age

* undergoing mechanical ventilation for ≥24h

* staying in the ICU ≥48h

* with a close relative ≥ 18 years of age, as identified by the relative.

Both the patients and the relative:

* Danisch speaking

Exclusion criteria:

patients:

* severe cognitive deficits or neurological damage

relatives continued in the study if the patient died or refused participation.

N total at baseline:

Intervention: 56, 44 analysed

Control:60, 47 analysed

Important prognostic factors:

(characteristics of included family members)

Age, median (range)

I: 56 (25-79)

C: 48 (18-91)

Sex:

I: 23% M

C: 20% M

Ralationship to patient, %:

I: Spouse 38, child 42, parent 11, other 10.

C: Spouse 24, child 49, parent 8, other 19.

Groups comparable at baseline?

No statistically significant differences, but there were more male patients in the control group. Also the relationship to the patient was slightly different between the groups and relatives in the control group were younger. Baseline distress levels not assessed.

Diary (incl. pictures) authored by a close relative for a critically ill patient. The ICU nurses guided the relative on how to write a diary for the ICU patient and later to share the diary with the patient.

No diary

Length of follow-up:

3 months

Loss-to-follow-up:

Intervention, n(%): 9 (16)

3 couples were excluded due to the inclusion criteria.

Control, n(%): 12 (20)

no reasons provided.

1 couple was excluded due to the inclusion criteria.

Incomplete outcome data:

See also the results. No reasons for incomplete outcome data provided.

1. Anxiety,

Evaluated using the HADS

3 months, HADS score ≥ 11, n(%)

I: 5/41 (12)

C: 7/46 (15)

RR: 0.80 (95%CI 0.28;2.33), p=0.76

2. Depression Evaluated using the HADS

3 months HADS score ≥ 11, n(%)

I: 3/41 (7)

C: 5/46 (11)

RR: 0.67 (95%CI 0.17;2.64), p=0.72

3. PTSD

Evaluated using the PTSS-14

3 months, median, [interquartile range]

I: 26 [14-64]

C: 32 [14-77]

Difference: 26.3%, 95%CI: 4.8-52.2, p=0.01

PTSS-14 score ≥ 31, n(%)

I: 14/43 (32)

C: 23/46 (50)

RR: 0.65 (95%CI 0.39;1.09), p=0.13

NNT: 5.9

4. Satisfaction

Not reported

5. Quality of life

Evaluated using the SF-36,

Physical function, mean

I: 86.1

C: 85.1

p=0.73

Role functioning

I: 73.7

C:68.5

p=0.20

Global health

I: 78.2

C:70.8

p=0.10

role emotional

I: 73.5

C: 70.4

p=0.79

Mental health

I: 77.4

C: 70.4

p=0.16

Sample size estimation: 71 participants in each group needed. Planned sample size was not achieved. Sample size estimation only performed for PTSD in patients.

Differences in the number of cases were tested using Fisher exact test.

17% in the control group wrote a diary, but the intensity of the intervention was less.

15 patients in the intervention group and 16 patients in the control group died before follow-up.

Garrouste-Orgeas, 2019

Type of study:

multicenter RCT.

Setting and country: 35 French ICUs

Funding and conflicts of interest: Study was funded through a grant from the Fondation de France and was sponsored by the Saint Joseph Hospital Network. Some authors received grants, personal or consultancy fees and/or nonfinancial support from pharmaceutical companies.

Inclusion criteria:

patients

* aged ≥18 years

* receiving mechanical ventilation for at least 48 hours that was initiated within 48 hours of ICU admission

* family member present during inclusion period and able to visit the patient during ICU stay.

Both family member and patient

* sufficient French-language skills

Exclusion criteria:

patients:

* no family members visiting

* under legal guardianship

* preadmission diagnosis of psychosis or dementia.

* acute neurological diseases, with cardiac arrest at admission, mute or deaf

* whose status was considered by the investigator as highly likely to lead to death or withdrawal of life support within 48 hours

*participating in another trial with a telephone interview after ICU discharge.

N total at baseline:

family members

Intervention: 332, 281 analysed IES-R and 286 analysed HADS

Control: 325, 282 analysed IES-R and 286 analysed HADS

Important prognostic factors:

(characteristics of family members)

Age, median (range)

I: 55 (45-66)

C: 54 (44-66)

Sex:

I: 27% M

C: 28% M

Ralationship to patient, %:

I: Spouce 54, child 33, parent 5, other 9.

C: Spouce 56, child 30, parent 4, other 10.

Groups comparable at baseline? Yes, but c.

ICU dairy filled out by clinicians and family members during ICU stay.

Before discharge from the ICU, a concluding note was written by an ICU clinician. The dairy was given to the patient or family member before discharge. If the patient died, the diary was concluded with a condolence letter and mailed to the family.

Usual ICU care without ICU dairy

Length of follow-up:

3 months

Loss-to-follow-up:

Intervention, n(%): 51 (15)

n=27 declined

n=17 no contact

n=7 no IES-R score

Control, n(%): 43 (13)

n=18 declined

n=14 no contact

n=11 no IES-R score

Incomplete outcome data:

see loss to follow-up

1. Anxiety,

Evaluated using the HADS

3 months, median, [interquartile range],

I: 7 [5-11]

C: 7 [5-11]

Difference (95%CI): 0.28 (-0.5;1.04), p=0.65

HADS score ≥ 8, n(%)

I: 141 (49)

C: 134 (47)

Risk difference (95%CI): 2 (-6 to 11), p=0.56

2. Depression Evaluated using the HADS

3 months, median, [interquartile range]

I: 4 [1-7]

C: 4 [1-7]

Difference (95%CI): 0.05 (-0.67;0.78), p=0.96

HADS score ≥ 8, n(%)

I: 70 (24.5)

C: 67 (23.4)

Risk difference (95%CI): 1 (-6 to 8), p=0.77

3. PTSD

Evaluated using the IES-R

3 months, median, [interquartile range]

I: 20 [11-35]

C: 20 [10-37]

Difference (95%CI: 0.48 (-2.5;3.5), p=0.87

IES-R score ≥ 22, n(%)

I: 134/281 (48)

C: 127/282 (45)

RD: 3 [95%CI -6% to 11%], 0.53

4. Satisfaction

Not reported

5. Quality of life

Not reported

To be included, ICUs had to have more than 8 beds.

3 ICUs used were routinely using an ICU dairy before the beginning of the trial.

Sample size estimation: 176 participants in each group needed. Sample size estimation only performed for PTSD in patients.

“Comparisons between groups were performed using X2 tests, Mann-Whitney Test an multivariable regressions (generalized linear multivariable mixed model).

the final logistic regression model included: presence of diary, age, sex, simplified acute physiology score II, duration of propofol exposure, acute respiratory failure/chronic obstructive respiratory disease, emotional memories, and delusional memories.

Of the 332 patients in the intervention group and 325 patients in the control group, only 193 and 209 patients were alive at 3 months follow-up, respectively.

White, 2018

(PARTNER)

Type of study:

stepped-wedge, cluster RCT

Setting and country: 5 ICUs at 5 hospitals in the UPMC health system, Pittsburgh, USA

Funding and conflicts of interest: Innovation Award from the UPMC Health system and by the Greenwall Foundation

Inclusion criteria:

Patients:

* age of 18 years and older

* lack of decision making capacity as judged by the patient’s attending physician

* receipt of mechanical ventilation for at least 4 consecutive days, an estimated change of death during hospitalization of at least 40% as judged by patient’s attending physician, and/or an estimated change of severe long-term functional impairment of at least 40% as judged by the patient’s attending physician.

One surrogate decision maker of ≥18 years of age – identified as the patients main surrogate - was enrolled.

Exclusion criteria:

Patient:

* no surrogate decision maker

* receiving only comfort-focused treatment at the time that they were eligible for enrolment.

N total at baseline:

Intervention: 429

Control: 677

Important prognostic factors:

(characteristics of family members)

Age, mean (SD)

I: 57.1 (13.7)

C: 56.4 (13.6)

Sex:

I: 34% M

C: 29% M

Relationship to patient, %:

I: Spouse 38, child 38, parent 7, other 18.

C: Spouse 44, child 29, parent 9, other 18.1.

Groups comparable at baseline? Surrogates yes, but there were differences in the demographics of the patients:

“There were differences between the treatment groups in some of the patient

characteristics at baseline, including an older mean age and higher scores for the severity of

acute illness and the number of chronic coexisting

conditions in the intervention group than in

the control group”

Multicomponent family-support intervention delivered by ICU team. Three components:

* advanced communication training that focussed on skills for supporting families of seriously ill patients, provided to 4-6 nurses in each ICU.

* family-support pathway: nurses met with families on a daily basis, according to a standardized protocol, and arranged clinician-family meetings within 48 hours after enrolment and every 5-7 days thereafter.

* intensive support for implementation provided by an quality-improvement specialist.

Usual care

Length of follow-up:

6 months

Loss-to-follow-up:

1106 patients agreed to be contacted for long-term follow-up, 809 completed long-term follow-up.

“Demographic characteristics

were generally similar between patients

whose surrogates completed long-term follow-up

and patients whose surrogates did not”

1. and 2. Anxiety and depression

Evaluated using the HADS

6 months, mean (SD) or mean (95%CI)

I: 11.7 (7.9)

C: 12.1 (8.5)

I adj: 11.7 (10.7 to 12.7)

C adj: 12.0 (11.3 to 12.8)

β: -0.34, 95%CI -1.67;0.99, p=0.61

“There was no significant difference between the two groups in analyses of the

HADS and IES scores as categorical outcomes

with established cutoff points or in analyses of

the scores on the HADS subscales of anxiety and

depression as continuous, dichotomous, and ordinal

outcomes.”

3. PTSD

Evaluated using the IES

6 months, mean (SD) or mean (95%CI)

I: 20.5 (18.1), n=308

C: 20.7 (17.7) n=501

I adj: 21.2 (19.3 to 21.2)

C adj: 20.3 (18.8 to 21.9)

β=0.9, 95%CI -1.66 to 3.47, p=0.49

4. Satisfaction

Not reported

5. Quality of life

Not reported

Each ICU (cluster) began in the control phase and transitioned to the intervention phase at a randomly assigned time (wedge).

Sample size estimation: 1000 participants needed.

The authors used generalized linear mixed models, with ICU and time as random effects.

“Adjusted analyses were performed with regression models. All models were adjusted for patient’s age, modified SAPS III, Elixhauser Comorbidity Index score, use or nonuse of mechanical ventilation, primary diagnosis, and admission source.” Furthermore the analyses of the HADS scores and the IES were additionally adjusted for patient’s vital status at 6 months after discharge, the surrogate’s sex, and the surrogates’ s relationship to the patient.

1106 surrogates agreed to be contacted for long-term follow-up. 809 completed long-term follow-up.

“In-hospital mortality was higher in the intervention

group than in the control group (36.0% vs.28.5%; odds ratio, 1.43; 95% CI, 1.10 to 1.87; P = 0.008), but mortality at 6 months did not differ significantly between the two groups (60.4% and 55.4%, respectively; adjusted odds ratio, 1.18; 95% CI, 0.93 to 1.50; P = 0.17).”

Carson, 2017

Type of study:

multicenter RCT

Setting and country: 4 ICUs: urban tertiary care center, 2 tertiary care and a community hospital, USA

Funding and conflicts of interest: National Institute of Nursing Research.

Inclusion criteria:

Patients

* Adult ≥21 years requiring ≥ 7 (or 10 days in the first year) days of mechanical ventilation uninterrupted for 96 hours or longer

* not expected to be weaned or to die within 72 hours.

Family members

* responsibility of health care decision making for the patient

Exclusion criteria:

Patients

* mechanically ventilated at an outside hospital for longer than 7 days

* chronic neuromuscular disease, trauma or burns.

* surrogate decision maker not available or lacked English proficiency

* primary physician refused to grant permission to approach the patient or family

* the investigators were the attending physicians.

N total at baseline:

Intervention: 184, 163 analysed

Control: 181, 149 analysed

Important prognostic factors:

(characteristics of family members)

Age, mean (95%CI)

I: 51 (48.8;52.8)

C: 51 (48.6;52.7)

Sex:

I: 30% M

C: 28% M

Ralationship to patient, %:

I: Spouce 44, child 32, parent 14, other 10.

C: Spouce 37, child 33, parent 13, other 18.

Treated for anxiety in the past, n(%)

I: 38 (21)

C:42 (25)

Treatment for depression in the past, n(%)

I :54 (29)

C: 53 (29)

Groups comparable at baseline? No statistically significant differences.

At least 2 structured family informational and emotional support meetings led by palliative care clinicians, separated by 10 days. Furthermore provision of an informational brochure.

Scheduling of additional support and information team meetings between these time points was possible.

The palliative care team conducted pre-meetings with ICU physicians to review each patient’s condition, prognosis, and previous discussions of goals of care. Prognosis was also based on the ProVent 14 score.

The ICU clinicians could attend the meetings, or were provided with feedback from the palliative care team.

Provision of an informational brochure and routine family meetings conducted by ICU teams

“Clinicians were able to formally consult palliative care clinicians at their discretion even if randomized to the usual care control group, and this was encouraged if they needed assistance with symptom management or for transfer to hospice.”

Length of follow-up:

Immediately after the second support and information meeting, at 10 days after randomization for the control group (unless the patient had died), 3 months (median 105 days after randomization)

Loss-to-follow-up:

Intervention, n(%): 21 (11)

n=15 refused to participate, n=6 unavailable

Control, n(%): 32 (18)

n=15 refused to participate

n=17 unavailable

Incomplete outcome data:

IES-R

Intervention:2

Control: 4

1. Anxiety

Evaluated using the HADS

Baseline, mean (SD):

I: 9.5 (4.8)

C: 9.8 (4.7)

3 months, mean (SD) or mean (95%CI)

I: 7.2 (4.6)

C: 6.4 (4.7)

I adj: 7.2 (6.6;7.9)

C adj: 6.4 (5.7;7.1)

Difference: 0.8, 95%CI -0.1 to 1.8, p=0.9

HADS score ≥ 8, % (95%CI)

I adj: 44 (35;53)

C adj: 31 (23;40)

OR: 1.72, 95%CI 1.00;3.00, p=0.05.

2. Depression

Evaluated using the HADS

Baseline, mean (SD)

I: 6.6 (4.0)

C:6.7 (4.4)

3 months, mean (SD) or mean (95%CI)

I: 4.9 (4.2)

C: 5.0 (4.5)

I adj: 5.0 (4.4;5.6)

C adj: 5.0 (4.3;5.6)

Difference: 0, 95%CI -0.9;0.9, p=0.93

HADS score ≥ 8, % (95%CI)

I adj: 24 (17;31)

C adj: 22 (16;30)

OR: 1.09, 95%CI 0.62;1.92, p=0.77.

3. PTSD

Evaluated using the IES-R

3 months, mean (SD) or mean (95%CI)

I: 25.6 (18.0)

C: 20.7 (18.3)

I adj: 25.9 (22.8;29.0)

C adj: 21.3 (18.0;24.6)

Difference=4.6, 95%CI 0.01 to 9.10, p=0.05

Adjusting for additional variables did not change the interpretation of the results

IES-R score ≥ 33,% (95%CI)

I: 34 (27;42)

C: 25 (18;33)

OR: 1.56 [95%CI 0.90 to 2.60], p=0.10.

4. Satisfaction

As evaluated using the 24-items Family Satisfaction in the Intensive Care Unit survey. Total score, mean (95%CI)

I: 81.1 (78.3;83.9)

C:84.3 (81.3;87.3)

difference= -3.1, 95%CI -7.3 to 1.0, p=0.13.

5. Quality of life

Not reported

“The primary surrogate decision maker was determined through discussions with the ICU physicians, nurses and social workers, by review of the medical record, and by asking individual family members”

Estimated sample size: 150 family members in each study arm.

“For the primary analysis, the HADS score was adjusted for the baseline score and for multiple surrogate respondents. The IES-R was adjusted for multiple respondents.”

“…Family Satisfaction in the Intensive Care Unit survey were adjusted for multiple surrogate respondents and study site.”

“The proportion of patients meeting diagnostic cutoffs for anxiety and depression (scores ≥8 for the HADS Anxiety and Depression subscales for both anxiety and depression, adjusting for baseline and for multiple surrogate respondents) and PTSD (scores >33 on the IES-R, adjusting for multiple surrogate respondents) were compared using generalized linear models allowing for random effects”

“Eighty-two percent of family surrogate decision makers in the intervention group participated in at least 1 support and information team meeting, and there was an average of 1.4 meetings per surrogate.”

34 surrogates did not receive the intervention (n=22 surrogates unavailable, n=8 patients died, n=2 patients discharged before meeting, n=2 surrogates withdrew).

1,9 family meetings were conducted independently by the ICU teams for families in the intervention group after randomization that were separate from the support and information team meetings. This was similar as the number of family meetings (2.1) conducted by the ICU teams for families in the control group.

“Thirteen percent of patients in the intervention group had a formal palliative care consultation outside the study protocol compared with 22% of patients in the control group; however, adjusting for this variable had no significant effect on the between-group comparison.”

“Potential explanations for this lack of benefit may relate to the high perceptions of quality of communication, emotional support, and family satisfaction in the usual care control”

Hospital mortality, n(%)

I: 49 (38)

C: 51 (40)

Curtis, 2015

Type of study:

parallel-group randomized trial

Setting and country: 5 ICUs from 2 hospitals, one academic level-one trauma center and one community-based hospital, Seatle, USA.

Funding and conflicts of interest:

“Funded by the National Institute of Nursing Research (R01 NR05226).”

Inclusion criteria:

Patients

* patients ≥18 years

* in ICU for more than 24 hours

* mechanically ventilated at enrolment

* SOFA score greater than or equal to six or diagnostic criteria predicting a greater than or equal to 30% risk of hospital mortality

* legal surrogate decision maker to consent for patient participation

* family member able to come to the hospital

Family members

* age ≥18years

* able to complete consent process and questionnaires in English

*completion of a

baseline questionnaire within 2 wk and a

follow-up questionnaire within 60 d of

distribution

N total at baseline:

“The 68 randomized patients had 268 family members who participated with 51% (n = 137) associated with control

patients (n = 86) and 49% (n = 131)

associated with intervention patients (n = 82).”

Intervention: 131 (of 150 surveyed)

Control: 137 (of 156 surveyed)

Important prognostic factors:

(characteristics of family members)

I: 49.5 (12.0)

C: 52.4 (14.2)

Sex:

I: 29% M

C: 30% M

Relationship to patient, %:

I: Spouse 30, child 25, parent 21, other 24.

C: Spouse 29, child 29, parent 18, other 14.

Groups comparable at baseline? Yes

Communication-facilitator intervention: the trained facilitators (nurse and social worker) assisted families of patients by providing communication support during the ICU stay.

“It included the following: (1) interviews by facilitators with family to understand the family’s concerns, needs, and communication characteristics; (2) meetings by facilitators with physicians, nurses, or other clinicians offering brief

summary of family concerns, needs, and

communication characteristics; (3) provision of communication and emotional support adapted to the family member’s

attachment; (4) facilitator participation in family conferences; and (5) 24-hour follow-up with the family after discharge to acute care.”

Usual care.

Length of follow-up:

3 and 6 months after the patients died in or was discharged from the ICU.

Loss-to-follow-up:

3-months

Intervention, n(%): 53 (41)

Reasons (describe): n=24 no baseline questionnaire, n=29 no 3-month questionnaire.

Control, n(%), 69 (50)

n=51 no baseline questionnaire, n=18 no 3-month questionnaire.

6-months

Intervention, n(%): 55 (42)

Reasons (describe): n=24 no baseline questionnaire, n=31 no 3-month questionnaire.

Control, n(%), 76 (55)

n=51 no baseline questionnaire, n=25 no 6-month questionnaire.

Incomplete outcome data:

3 months

Intervention, n (%): 3 (4)

n=1 informed consent not signed, n=2 late baseline questionnaire

Control, n (%): 11 (16)

n=7 late baseline questionnaire, n=3 late 3-month questionnaire

6 months

Intervention, n (%): 7 (9)

n=1 informed consent not signed, n=2 late baseline questionnaire, n=4 late 6-month questionnaire.

Control, n (%): 9 (15)

n=7 late baseline questionnaire, n=2 late 6-month questionnaire

1. Anxiety

Evaluated using the Generalized Anxiety Disorder-7

3 months, n=127

I: 2.3

C: 3.0

β: -0.7, 95%CI -2.9 to 1.4, p=0.5

6 months, n=117

I: 1.8

C: 2.7

β: -0.9, 95%CI -3.1 to 1.3, p=0.4

Both at 3 and 6 months, the scores were adjusted for baseline GAD score.

2. Depression

Evaluated using the Patient Health Questionnaire (PHQ)-9

3 months, n=118

I: 3.1

C: 4.9

β: -1.8, 95%CI -3.9 to 0.3, p=0.1

“Adjustment variables were the patient’s SOFA score at study enrollment, and mortality status at hospital discharge; and the family member’s age, legal

next of kin status, length of acquaintance with the patient, and baseline report of end-of-life treatment discussion with the patient.”

6 months, n=115

I: 2.4

C: 4.7

β: -2.4, 95%CI -4.3 to -0.4, p=0.02

“Adjustment variables were the patient’s age at study enrollment and mortality status at hospital discharge; and the family member’s age, length of

acquaintance with patient, and relationship to patient (dummy indicators for spouse, parent, and child).”

“Adjusted depression scores were lower for family members randomized to

the intervention, but the between-group difference was not statistically significant.”

6 months, n=115 family members

Adjusted depression scores

were significantly lower for the intervention group than for the control group (P = 0.017)”

3. PTSD

Evaluated using the PTSD Checklist Civilian Version (PCL)

3 months, n=126

I: 29.8

C: 31.6

β: -1.8, 95%CI -6.7 to 3.1, p=0.5

6 months, n=114

I: 27.1

C: 30.6

β: -3.5, 95%CI -7.1 to 0.1, p=0.06

4. Satisfaction

Not reported

5. Quality of life

Not reported

About the PHQ-9: “A minimal clinically important difference (MCID) is defined as five points (19, 21–25).”

“For analysis of family reported outcomes, we used clustered regression

models (family members clustered under patients) and included all family members who met inclusion criteria and who had complete data on the outcome, its value at baseline, and any

variables confounding the association between intervention and outcome. We used tobit regression for depression and anxiety outcomes (because of floor effects), robust linear regression for PTSD … of stay. All estimation used restricted maximum

likelihood. This was a “complete case” analysis.” “These parameters suggested a total of 350 family members (175 per

group) would enable us to identify family members who benefitted from the

intervention.”

This number was not met.

“The facilitators had an average of 9.4 contacts with family members per family (SD, 6.5 and spent an average of 267 minutes with family members per family (SD, 211).”

“Fourth, it is possible that

there could be contamination of the control

group during this study, which could bias our results toward the null.”

Study reference

(first author, publication year)

Describe method of randomisation¹

Bias due to inadequate concealment of allocation?²

(unlikely/likely/unclear)

Bias due to inadequate blinding of participants to treatment allocation?³

(unlikely/likely/unclear)

Bias due to inadequate blinding of care providers to treatment allocation?³

(unlikely/likely/unclear)

Bias due to inadequate blinding of outcome assessors to treatment allocation?³

(unlikely/likely/unclear)

Bias due to selective outcome reporting on basis of the results?⁴

(unlikely/likely/unclear)

Bias due to loss to follow-up?⁵

(unlikely/likely/unclear)

Bias due to violation of

intention to treat analysis?⁶

(unlikely/likely/unclear)

Garrouste-Orgeas, 2017

“The family members were allocated at random to the intervention or control group, using sequentially numbered, opaque, sealed envelopes. Randomisation was generated by S.R. via a computer. Two main investigators enrolled and assigned the participants”.

Unlikely

Likely

Likely (and risk on contamination, as physicians could led conferences in both groups)

Likely

Unlikely

Unclear, no reasons provided (only that patients withdrew consent)

Likely, family members lost to follow-up were not included in the analysis. Furthermore, there is risk on contamination.

Nielsen, 2020

“Patient-relative dyads were randomized 1:1 to either diary intervention or control group receiving usual care. Block randomisation secured that all sites recruited to both groups. Randomisation was carried out after obtaining informed written consent from relatives and collection of baseline data. Allocation concealment was achieved using opaque envelopes. A designated group of study nurses and the primary investigator conducted recruitment and randomisation, whereas ICU nurses could give information about the diary.”

Unclear, block randomisation, but

not clear if multiple block sizes were used. Furthermore, unclear whether the envelops were sealed.

Likely (“blinding of participants and staff was not possible because of the nature of the intervention”)

Likely (“blinding of participants and staff was not possible because of the nature of the intervention”)

Likely (“blinding of participants and staff was not possible because of the nature of the intervention”)

Unlikely

likely: Only 78% of relatives completed the questionnaire. The % females in the analysed control group was higher when compared to the percentage in the intervention group (87 vs 77%) and the median age was lower. Nielsen: “As lower age and female gender are risk factors for development of PTSD, the possibility of an overestimation of positive effects of the diary should at least be considered”.

Likely, relatives lost to follow-up were not included in the analysis. There was some contamination.

Garrouste-Orgeas, 2019

“the patients were randomized in a 1:1 ratio into 2 groups, stratified by center, using a secure web-based random number generator to select permuted blocks, with a block size of 4.”

Unlikely.

Likely

Likely

Unclear, the outcome assessor who contacted patients and family members was blind to the randomization group, however patients of course provided the answers to the questions.

Unlikely

Unclear, there was some loss to follow-up, but the reasons for the loss of follow-up were only minimally described.

Likely, patients were analyzed according to their randomization group, but family members lost to follow-up were not included in the analysis.

White (2018)

Each ICU (cluster) began in the control phase and transitioned to the intervention phase at a randomly assigned time (wedge). A computer-generated randomization scheme was used. New crossovers occurred every 6 months.

Unlikely

Unclear, it is unclear whether the surrogates new at the time of the intervention were participating in this study.

Likely

Unlikely, the outcome assessor who contacted patients’ surrogates was blind to the randomization group.

Unlikely, although results on the subscales of the HADS were not published separately.

Likely, loss-to-follow up was quite substantial, no reasons provided.

Unclear “All analyses were performed on an intention-to-treat basis” but unknown whether all patients were included in the analysis.

Carson, 2017

“After _enrolment of patients and family members, patients were randomized to the intervention or the control group using a computer-generated, web-based randomization system with blinding of allocation. The randomization was stratified by study site in block sizes varying from 8 to 10. The research coordinator at each study site who had knowledge of group assignments was not involved in collection of the primary outcomes through family interviews.”

Unlikely

Unclear, family members were not blinded, but the results are contra intuitive

Likely

Unclear, observers were blinded to group allocation for the measurements of primary outcomes, however patients of course provided the answers to the questions.

Unlikely

Unclear, there was some loss to follow-up, but the reasons for the loss of follow-up were only minimally described.

Likely

“All analyses were 2-tailed and performed on an intent-to-treat basis.” However, family members lost to follow-up were not included in the analysis.

Curtis, 2015

“Patients were randomly assigned to intervention or control in a 1:1 ratio. Randomization was stratified by hospital in block sizes of six, with results provided to study staff in sealed, opaque, consecutively numbered envelopes.”

unlikely

Likely

Likely

Likely

Unlikely

Likely, Full information maximum likelihood (FIML) estimation models gave similar results as the complete case analyses. However loss to follow-up was quite substantial.

Likely, not reported and high amount of incomplete data.