The Caregiver Pathway Intervention Can Contribute to Reduced Post-Intensive Care Syndrome Among Family Caregivers of ICU Survivors: A Randomized Controlled Trial

Abstract

OBJECTIVES:

Explore short-term effects of “The Caregiver Pathway,” an intervention for family caregiver follow-up, on Post-Intensive Care Syndrome symptoms among families (PICS-F).

DESIGN:

A randomized controlled trial.

SETTING:

A medical ICU at a Norwegian University Hospital.

PARTICIPANTS:

One hundred ninety-six family caregivers of critically ill patients randomized to an intervention (n = 101) or control group (n = 95).

INTERVENTIONS:

“The Caregiver Pathway” four-step model offers individual and structured follow-up, including: 1) mapping family caregivers’ needs and concerns with an assessment tool followed by a conversation with an ICU nurse within the first days at the ICU, 2) a supportive card when leaving the ICU, 3) offer for the family caregivers to receive a phone call after ICU patient discharge, and 4) a follow-up conversation within 3 months.

MEASUREMENTS AND MAIN RESULTS:

Data were collected at baseline and after 3 months and analyzed using linear regression. No significant effects were detected when comparing all participants completing 3-month outcome measurements (n = 144). A subgroups analysis stratified on patient survival, however, showed statistically significant effect for family caregivers of patients surviving the ICU stay receiving the intervention compared with controls. Caregivers of surviving patients reported improved symptoms related to post-traumatic stress disorder, measured by Impact of Event Scale-Revised (B = –8.2 [95% CI, –14.2 to –2.2]; p = 0.008), anxiety (B = –2.2 [95% CI, –4.0 to –0.5]; p = 0.014), and depression (B = –1.5 [95% CI, –2.9 to –0.1]; p = 0.035); measured by the Hospital Anxiety and Depression Scale, subscore physical functioning in health-related quality of life (B = 9.7 [95% CI, 0.3–19.0]; p = 0.043); measured by Short Form 12-Item Health Survey; and hope (B = 2.4 [95% CI, 0.4–4.3]; p = 0.017) and measured by the Herth Hope Index. At 3-month, the model did not appear to improve the outcomes for family caregivers of nonsurviving patients.

CONCLUSIONS:

“The Caregiver Pathway” intervention was associated with reduced symptoms of PICS-F in family caregivers of surviving ICU patients compared with controls.

KEY POINTS.

Question: Could “The Caregiver Pathway” model improve follow-up of family caregivers to critically ill patients and help reduce Post-Intensive Care Syndrome-Family (PICS-F) symptoms for these caregivers?

Findings: A randomized controlled trial at 3 months showed “The Caregiver Pathway” to be associated with reduced symptoms of PICS-F in terms of reduced symptoms of post-traumatic stress disorder, anxiety, and depression for caregivers of surviving patients, compared with controls with surviving patients.

Meaning: Structured, individual follow-up of family caregivers of critically ill patients can reduce PICS-F symptoms for family caregivers of patients surviving ICU stays.

Family caregivers of critically ill patients are in a dire situation and at risk for developing symptoms of post-traumatic stress disorder (PTSD), anxiety, and depression, symptoms known as Post-Intensive Care Syndrome-Family (PICS-F) (1, 2). These symptoms can profoundly impact their quality of life in the time after the incident (1, 3). PICS-F is of major concern for critical care medicine, as the number of people being treated for critical illness is increasing, accompanied by the number of subsequent family caregivers (2, 4). While efforts have been made to meet these dire challenges (5–9), models to follow-up family caregivers are urgently needed (10–13). “The Caregiver Pathway” (14) was developed to meet this disparity, with the model innovatively introducing follow-up of family caregivers in an individual and structured way.

The current study sought to test short-term (i.e., 3-mo) efficacy of the intervention in a randomized controlled trial (RCT), with the aim of reducing the symptoms of PICS-F among family caregivers who received the intervention, compared with family caregivers in the control group. The primary outcomes are symptoms of PTSD, anxiety, and depression and secondary outcomes health-related quality of life (HRQoL), self-efficacy, and hope.

MATERIALS AND METHODS

Study Design

Family caregivers were recruited to a single-center RCT from the medical ICU (MICU) at Oslo University Hospital (OUH), Norway, from April 2021 to August 2023. The inclusion criteria were being able to understand and speak Norwegian, being between 18 and 70 years old, and with the patient expected to be invasive ventilated for at least 48 hours. For each patient, one to three family caregivers could be included. ICU bedside nurses identified potential participants and asked if they would like to receive information about the study. Those interested were introduced to the study by the first author (i.e., an ICU nurse) by phone or face-to-face if present at the ICU. If willing to participate, the family caregivers received an e-mail with a link to the informed consent form or a paper-based form if preferred. When the informed consent was signed and the baseline outcome measurements completed, the first author randomized family caregivers to the intervention or control groups using a computerized randomization program (R-tool: RNGCryptoServiceProvider in Microsoft.net, Block size: 10; Microsoft Corporation, Redmond, WA). If more than one family caregiver for a patient were included in the study, all were allocated to the same group (i.e., intervention or control). The intervention was initiated 1–3 days after randomization.

All family caregivers received standard care, including spontaneous conversations with the ICU nurses and meetings with the physicians. If needed, all participants could be referred for additional support at the hospital (e.g., the social worker). In addition, the intervention group received follow-up in line with “The Caregiver Pathway” delivered by one of 11 dedicated ICU nurses from the MICU. The 11 ICU nurses were trained in communication skills, such as the mnemonic VALUE method; Value family contribution, Acknowledge emotions, Listen, Humanity, and the Eliciting of information (1), received an introduction course (3 hr) about “The Caregiver Pathway” model and participated in regular meetings for education and support three to four times a year. Other staff at the MICU did not know the content of “The Caregiver Pathway.”

The study was conducted in accordance with the Helsinki Declaration of 1975 and approved by the Regional Committee for Medical Research Ethics South-East Norway (Approval number: 199446; Approval date: March 15, 2021; Title: Pårørende til kritisk syke—Veien videre) and the Institutional Review Board at OUH (21/03594) and registered in ClinicalTrials.gov (NCT04839406) prior to study enrollment.

Intervention

“The Caregiver Pathway” four-step model (14) offers individual and structured follow-up of family caregivers of critically ill patients (Fig. 1; and Supplement 1, http://links.lww.com/CCM/H634). In step 1, within the first days at the hospital, the family caregivers are offered an individual conversation through a digital assessment tool, assisted by an ICU nurse, where the family caregivers can articulate how they are doing and mark their needs and concerns. The assessment tool generates a summary to be used in a conversation with the ICU nurse, who then can support and provide follow-up from a social worker, priest, or psychiatrist if required or appropriate. In step 2, the family caregivers receive a preprinted card with information and support when leaving the ICU. In step 3, family caregivers receive a text message offering a phone call if the patient is transferred to a step-down unit. Step 4 entails an individual follow-up conversation with an ICU nurse, following a template, within 3-month post-ICU discharge. Family caregivers are given the opportunity to talk about their ICU experiences, clarify any potential misunderstandings, and eventually process challenging situations. If required or deemed beneficial, a physician is invited to the follow-up conversation. The family caregivers are also asked about their current situation and information about available resources for support locally or nationally are provided if needed.

Figure 1.

The four steps of “The Caregiver Pathway” model adapted from Watland et al (14).

The tailored approach in “The Caregiver Pathway” seeks to ensure appropriate steps are offered to each family caregiver. The clinical course, the length of stay, and the outcome may vary considerably, and all four steps from the model may not be suitable in all cases. For example, in this study, the first conversation (step 1) was not offered if the patient was about to pass away after a short stay at the ICU. The discharge conversation (step 3) was only offered to family caregivers of patients transferred to step-down units. However, the supportive card (step 2) and the follow-up conversation (step 4) were offered to all family caregivers, bereaved or not (Fig. 1; and Supplement 1, http://links.lww.com/CCM/H634).

Data Collection and Outcome Measurements

Demographic data included family caregiver gender, age, and education. The family caregivers also provided information about their relationship to the patient and the patient’s age, gender, and reason for admission to the ICU. Outcome measurements were collected at baseline and after 3 months. The intervention group received the 3-month measurements median 30 days (range, 0–83 d) after the follow-up conversation.

Primary Outcomes

Symptoms of PTSD were measured by the 22-item Impact of Event Scale-Revised (IES-R), with the subscales Intrusion (eight items), Avoidance (eight items), and Hyperarousal (six items). Each item has a score range from 0 to 4, resulting in a total score of 0–88 (15). Higher scores indicate more symptoms of PTSD. Although IES-R was not designed for a universal cutoff (15), a score above 24 has been found to indicate PTSD symptoms as a clinical concern (16), and scores above 30 (16) or 33 (17, 18) have been proposed as a cutoff for high-risk PTSD.

Symptoms of anxiety and depression were measured by the Hospital Anxiety and Depression Scale (HADS) (19, 20), a 14-item measure of psychological distress divided into two subscales: anxiety (HADS-A) (seven items) and depression (HADS-D) (seven items), with a score range of 0–21 for each subscale. Scores below 8 are considered nonclinical, scores from 8 to 10 as suggestive, and scores of 11 and above as probable presence of mood disorder (19).

Secondary Outcomes

HRQoL was measured by the Short Form 12-Item Health Survey. The scale has a Mental Component Score-12 and a Physical Component Score-12 (PCS-12). Summary scores range from 0 to 100, and higher scores indicate better health (21). Self-efficacy was measured by the General Self-Efficacy (GSE) scale (22). Total score ranges from 10 to 40, and higher scores indicate higher perceived GSE. Hope was measured by the Herth Hope Index (23), total scores range from 12 to 48, and higher scores indicate a higher level of hope. Reliability and validity were established for all outcome measurements (15–17, 20–29).

Sample Size

To detect a mean between-group difference of six points on the IES-R (9), and assuming high variation in data with a sd of 13 in both groups, a statistical power of 0.80 and a significance level of 5%, a sample size of 150 participants (75 in each group) was required. A 30% drop-out rate was assumed due to the challenging situation for the participants, and the estimated total sample size was 196 participants.

Statistical Analyses

Independent sample t tests chi-square tests were used to assess the potential between-group differences related to demographics and background variables at baseline. Linear regression analysis, controlling for baseline outcome measure values, was used to explore potential intervention effects among completers of 3-month measures. As the proportion reporting living with the patient was statistically significantly different between the groups, and living with the patient can be associated with higher risk of PICS-F (30, 31), this variable was included as a possible confounder in all analyses. The potential dependency of including three family caregivers from each patient was controlled for using a linear mixed model with fixed effects, with family caregivers from each patient as fixed factor. Regression coefficients size were almost identical compared with the model not including this fixed factor, indicating that the statistical dependencies among family caregivers to the same patient did not confound the effect of the intervention. This factor was therefore not included in the final analysis.

To explore impact based on patient survival, a subgroup analysis with linear regression related to whether caregivers were bereaved or not was performed.

As depression is a risk factor for PTSD and may require additional follow-up (32–34), an exploratory subgroup analysis using linear regression of family caregivers with symptoms of depression of HADS-D less than or equal to 11 or greater than 11 at baseline was also conducted.

Independent samples t tests and chi-square tests compared completers and noncompleters of 3-month measures on baseline demographics and primary and secondary outcomes (i.e., sensitivity analysis). No significant differences were observed, and as imputation of additional data could potentially lead to increased skewness, the analysis was performed including complete cases only. All statistical analyses are conducted with SPSS, Version 25 (IBM Corp, Armonk, NY). p values of less than 0.05 were considered statistically significant.

RESULTS

Study Participants

In total, 486 patients received invasive mechanical ventilation at MICU during the recruitment period. Family caregivers of 129 patients were included in the study (intervention group: 62/control group: 67). Of 240 family caregivers invited to participate (Fig. 2), 198 signed the written informed consent form. The sensitive approach needed in the recruitment of the participants did not permit asking for reasons for not consenting to participation, but some spontaneous comments included “Could not bear it because of the crisis and the severity of the situation” or “Satisfied with the care as it is.” Of 198 consenting participants, one was excluded due to missing baseline outcome measurements and one due to the limitation of maximum three family caregivers for each patient.

Figure 2.

Study recruitment and participant flow.

Of 196 participants included, 101 were randomized to the intervention group and 95 to the control group. The only statistically significant difference between participants at baseline was that a significantly higher proportion of participants in the control group reported living with the patient (Table 1).

TABLE 1.

Demographics of Family Caregivers (n = 196) and Patients

Characteristics	Intervention Group (n = 101)	Control Group (n = 95)	p
Family caregivers
Age, mean (sd)	46.7 (13.3)	46.9 (13.8)	0.942
Gender, female, n (%)	68 (67.3)	59 (62.1)	0.444
Education, n (%)a			0.729
Elementary/high school	37 (36.6)	36 (37.9)
University/college ≤ 4 yr	37 (36.6)	30 (31.6)
University/college > 4 yr	27 (26.7)	29 (30.5)
Healthcare support in the current situation as caregiver, n (%)
Private network of healthcare professionals	26 (25.7)	21 (22.1)	0.551
General practitioner	12 (11.9)	13 (13.7)	0.705
Psychiatrist/psychologist	1 (1.0)	5 (5.3)	0.110
Healthcare at municipality	5 (5.0)	4 (4.2)	1.000b
Relation to the patient, n (%)a			0.377
Partner, spouse	20 (19.8)	25 (26.3)
Parent	14 (13.9)	18 (18.9)
Child	43 (42.6)	36 (37.9)
Sibling	14 (13.9)	12 (12.6)
Other	10 (9.9)	4 (4.2)
Living situation, n (%)
Living alone	16 (20.0)	12 (21.1)	0.880
Living with the patient	20 (19.8)	37 (38.9)	0.003
Patient-related details
Patient’s age, n (%)			0.163
16–30 yr	10 (9.9)	18 (18.9)
30–60 yr	36 (35.6)	34 (35.8)
60 yr old and older	55 (54.5)	43 (45.3)
Patient’s gender, female, n (%)	31 (30.7)	26 (27.4)	0.609
Patient’s diagnosis, n (%)c
Autoimmune disease	10 (9.9)	8 (8.4)	0.720
Sepsis	11 (10.9)	14 (14.7)	0.420
COVID-19	12 (11.9)	17 (17.9)	0.236
Drowning, accident, unknown	6 (5.9)	4 (4.2)	0.749b
Cerebral injury/attack	6 (5.9)	11 (11.6)	0.161
Heart failure/attack	28 (27.7)	12 (12.6)	0.009
Liver failure	4 (4.0)	10 (10.5)	0.097b
Respiratory failure	20 (19.8)	25 (26.3)	0.279
Neurologic disease	3 (3.0)	6 (6.3)	0.32b
Renal failure	11 (10.9)	14 (14.7)	0.420
Acute poisoning, suicidal attempt, psychiatric suffering	12 (11.9)	14 (14.7)	0.556

^aPercentages not 100 due to rounding.

^bFisher exact test.

^cPatients could have several diagnoses.

Seventy-seven participants (76%) in the intervention group and 67 (71%) in the control group completed the 3-month outcome measures, with 22 (29%) in the intervention group and 14 (21%) in the control group reporting the patient had died during their stay at ICU. In the intervention group, step 1 was received by 36% (8/22) of the bereaved and 71% (39/55) of the nonbereaved family caregivers, step 2 were administered to all participants, step 3 received by 2% (4/22) of the bereaved and 53% (29/55) of the nonbereaved, and step 4 was received by 100% (22/22) of the bereaved and 91% (50/55) of the nonbereaved family caregivers.

Psychosocial Outcomes

No statistically significant differences were detected between groups in primary and secondary outcomes after 3 months (n = 144) (Table 2). However, subgroup analyses stratified by patient survival revealed statistically significant differences between the intervention and control group among family caregivers of patients who survived (Table 3). For symptoms of PTSD, adjusted analyses showed that the intervention group (mean IES-R total: 23.0) scored on average eight points lower (B = –8.2 [95% CI, –14.2 to –2.2]; p = 0.008) compared with the control group (mean IES-R total: 31.7). Among caregivers of patients who survived, 24% (13/55) in the intervention group reported IES-R greater than 33, compared with 55% (29/53) in the control group. In addition, statistically significant lower symptoms of anxiety (B = –2.2 [95% CI, –4.0 to –0.5]; p = 0.014) and depression (B = –1.5 [95% CI, –2.9 to –0.1]; p = 0.035), as well as statistically significant higher (PCS-12) (B = 9.7 [95% CI, 0.3–19.0]; p = 0.043) and levels of hope (B = 2.4 [95% CI, 0.4–4.3]; p = 0.017) were detected in favor of the intervention group. No significant differences in symptoms of PTSD were detected among the bereaved caregivers (Table 4), although level of PTSD symptoms were high in both groups (IES-R total intervention group: 37.3; control group: 34.6).

TABLE 2.

Primary and Secondary Outcomes

Outcome	Intervention (n = 77)	Control (n = 67)	Between-Group Differencesa
	Mean (95 % CI)	Mean (95 % CI)	Unstandardized B (95 % CI)	p
Primary outcomes
Symptoms of post-traumatic stress disorder (IES-R), at 3 mo
IES-R total	27.1 (23.1–31.0)	32.3 (28.7–36.0)	–4.8 (–10.2 to 0.6)	0.080
IES-R intrusion	11.5 (9.9–13.1)	13.6 (11.9–1 5.2)	–1.8 (–4.1 to 0.4)	0.108
IES-R avoidance	9.5 (8.0–11.0)	10.8 (9.5–12.2)	–1.3 (–3.4 to 0.7)	0.200
IES-R hyperarousal	6.1 (4.9–7.3)	7.9 (6.6–9.2)	–1.6 (–3.4 to 0.1)	0.060
Hospital Anxiety and Depression Scale-Anxiety			–0.8 (–2.1 to 0.5)	0.205
Baseline	9.9 (8.8–11.0)	10.8 (9.6–12.0)
3 mo	6.1 (5.0–7.1)	7.3 (6.2–8.4)
Hospital Anxiety and Depression Scale-Depression			–0.1 (–1.2 to 0.9)	0.802
Baseline	7.2 (6.2–8.2)	8.1 (6.9–9.2)
3 mo	4.0 (3.1–4.9)	4.6 (3.7–5.5)
Secondary outcomes
Health-related quality of life (Short Form 12-Item Health Survey)
Physical functioning			3.6 (–2.7 to 9.8)	0.263
Baseline	76.0 (71.1–80.8)	73.0 (67.0–79.0)
3 mo	76.0 (70.7–81.2)	69.8 (63.0–76.6)
Mental functioning			2.0 (–5.9 to 10.0)	0.614
Baseline	53.3 (48.1–58.5)	53.9 (47.2–60.5)
3 mo	59.6 (53.4–65.8)	58.0 (51.4–64.5)
General Self-Efficacy			0.3 (–0.9 to 1.5)	0.622
Baseline	31.1 (30.1–32.2)	31.0 (30.0–32.1)
3 mo	31.5 (30.3–32.6)	31.2 (30.2–32.3)
Hope (Herth Hope Index)			0.9 (–0.5 to 2.3)	0.205
Baseline	37.7 (36.5–38.9)	36.7 (35.5–38.0)
3 mo	38.2 (36.8–39.5)	36.7 (35.6–37.9)

IES-R = Impact of Event Scale-Revised.

^aBetween-group differences from baseline to 3 mo (adjusted for baseline values), and at 3 mo for IES-R, are estimated using linear regression with the variable “living with the patient (yes/no)” as covariate.

TABLE 3.

Primary and Secondary Outcomes Among Caregivers to Patients Surviving the ICU Stay (n = 108)

Outcome	Intervention (n = 55)	Control (n = 53)	Between-Group Differencesa
	Mean (95 % CI)	Mean (95 % CI)	Unstandardized B (95 % CI)	p
Primary outcomes
Symptoms of post-traumatic stress disorder (IES-R), at 3 mo
IES-R total	23.0 (18.7–27.2)	31.7 (27.2–36.3)	–8.2 (–14.2 to –2.2)	0.008
IES-R intrusion	9.8 (8.0–11.5)	13.6 (11.6–15.6)	–3.6 (–6.2 to –1.0)	0.006
IES-R avoidance	8.1 (6.5–9.6)	9.9 (8.4–11.4)	–1.7 (–3.8 to 0.4)	0.109
IES-R hyperarousal	5.1 (3.9–6.4)	8.2 (6.6–9.8)	–2.8 (–4.8 to –0.9)	0.005
Hospital Anxiety and Depression Scale-Anxiety			–2.2 (–4.0 to –0.5)	0.014
Baseline	9.6 (8.2–11.0)	10.9 (9.5–12.4)
3 mo	5.3 (4.0–6.5)	7.6 (6.3–8.9)
Hospital Anxiety and Depression Scale-Depression			–1.5 (–2.9 to –0.1)	0.035
Baseline	6.8 (5.6–7.9)	8.5 (7.1–9.8)
3 mo	3.2 (2.2–4.2)	4.9 (3.8–6.0)
Secondary outcomes
Health-related quality of life (Short Form 12-Item Health Survey)
Physical functioning			9.7 (0.3–19.0)	0.043
Baseline	77.8 (72.1–83.6)	72.0 (65.0–79.1)
3 mo	79.1 (73.4–84.8)	68.6 (60.7–76.6)
Mental functioning			6.7 (–3.2 to 16.7)	0.183
Baseline	52.3 (46.0–58.6)	52.4 (44.5–60.3)
3 mo	65.8 (59.2–72.4)	58.2 (50.3–66.0)
General Self-Efficacy			0.6 (–1.1 to 2.4)	0.472
Baseline	31.3 (30.0–32.5)	31.4 (30.1–32.7)
3 mo	32.4 (31.0–33.7)	31.7 (30.4–32.9)
Hope (Herth Hope Index)			2.4 (0.4–4.3)	0.017
Baseline	38.1 (36.7–39.5)	37.0 (35.5–38.6)
3 mo	39.3 (37.8–40.7)	36.7 (35.4–38.1)

IES-R = Impact of Event Scale-Revised.

^aBetween-group differences from baseline to 3 mo, and at 3 mo for IES-R, are estimated using linear regression with the variable “living with the patient (yes/no)” as covariate.

TABLE 4.

Primary and Secondary Outcomes Among Bereaved (Patients Died During the ICU Stay, n = 36)

Outcome	Intervention (n = 22)	Control (n = 14)	Between-Group Differencesa
	Mean (95 % CI)	Mean (95 % CI)	Unstandardized B (95 % CI)	p
Primary outcomes
Symptoms of post-traumatic stress disorder (IES-R), at 3 mo
IES-R total	37.3 (29.5–45.2)	34.6 (29.9–39.2)	3.0 (–7.5 to 13.5)	0.568
IES-R intrusion	15.8 (12.8–18.8)	13.4 (11.0–15.7)	2.7 (–1.4 to 6.8)	0.188
IES-R avoidance	13.0 (9.8–16.2)	14.2 (11.3–17.1)	–1.4 (–6.0 to 3.2)	0.551
IES-R hyperarousal	8.5 (6.0–11.0)	7.0 (5.0–9.0)	1.7 (–1.9 to 5.2)	0.348
Hospital Anxiety and Depression Scale-Anxiety			2.0 (–1.0 to 4.9)	0.192
Baseline	10.6 (8.7–12.6)	10.2 (7.9–12.5)
3 mo	8.1 (6.2–9.9)	6.1 (3.7–8.4)
Hospital Anxiety and Depression Scale-Depression			2.7 (0.1–5.2)	0.043
Baseline	8.2 (6.1–10.2)	6.6 (5.2–8.0)
3 mo	5.9 (4.0–7.8)	3.4 (1.9–4.9)
Secondary outcomes
Health-related quality of life (Short Form 12-Item Health Survey)
Physical functioning			–5.0 (–22.8 to 12.8)	0.574
Baseline	71.3 (61.8–80.9)	76.8 (64.5–89.1)
3 mo	68.2 (56.3–80.0)	74.1 (60.6–87.6)
Mental functioning			–12.7 (–30.8 to 5.4)	0.163
Baseline	55.6 (45.6–65.5)	59.4 (47.3–71.5)
3 mo	44.1 (31.6–56.6)	57.1 (45.1–69.1)
General Self-Efficacy			–0.4 (–3.7 to 2.8)	0.783
Baseline	30.8 (28.5–33.0)	29.8 (27.6–31.9)
3 mo	29.2 (27.0–31.4)	29.7 (27.3–32.1)
Hope (Herth Hope Index)			–1.2 (–5.2 to 2.8)	0.543
Baseline	36.8 (34.3–39.3)	35.5 (33.8–37.2)
3 mo	35.4 (32.6–38.2)	36.6 (34.3–39.0)

IES-R = Impact of Event Scale-Revised.

^aBetween-group differences from baseline to 3 mo, and at 3 mo for IES-R, are estimated using linear regression with the variable “living with the patient (yes/no)” as covariate.

Symptoms of depression at baseline appeared to have an impact, as exploratory analyses, stratified on baseline HADS-D scores revealed family caregivers with HADS-D less than or equal to 11 (mean HADS-D: 6.1) in the intervention group, compared with the control group, to have significantly lower symptoms of PTSD (Supplement 2, Table S1, http://links.lww.com/CCM/H634). No such differences were observed among family caregiver with HADS-D greater than 11 (mean HADS-D: 14.8).

DISCUSSION

At 3 months, considering the sample as a whole, no statistically significant differences between family caregivers receiving the intervention and those in the control group where identified. Subgroup analyses stratified on whether the patients survived the critical illness or not, did detect significantly lower symptoms of PTSD, anxiety, and depression for family caregivers receiving the intervention compared with those who did not, when the patient survived. As the elements of PICS-F (i.e., symptoms of PTSD, anxiety, and depression) can have long-lasting negative impact not only on the family caregivers but on the entire family, profoundly impacting their quality of life (1–4), interventions to alleviate this overwhelming impact are of great importance (10–13). There are also indications that hope plays an important role in preventing PICS-F (35, 36), and levels of hope did increase among family caregivers to patient who survived and received “The Caregiver Pathway” intervention.

“The Caregiver Pathway” aims to improve strategies for effective, structured communication between staff and the family caregivers with emphasis on empathy and acknowledging emotions as this have been suggested to reduce the risk of developing PICS-F (3). The ICU nurses offering “The Caregiver Pathway” in this study were trained in communication. The bedside manner: to be aware of nonverbal interaction as showing empathy, listen actively, and attentive to each individual person, has been described as important for communication with family caregivers (37), all values and skills well-integrated among the ICU nurses delivering the intervention in this study.

When a patient dies at the ICU, the family caregivers are at special risk for developing psychological challenges (38), including major depressive disorder (39), complicated grief (40), and PTSD (41). This naturally indicates a more complicated and distressing situation and could be at least part of the reason why the results in the current study were more impactful for family caregivers whose patient survived. For the bereaved family caregivers, additional steps could be included in “The Caregiver Pathway” model in the future, as a recent study showed how a three-step support strategy decreased the level of complicated grief after six months among bereaved family caregivers after patient critical illness (42). It is also possible that the bereaved family caregivers in the current study might have benefited from more time to recover after the loss, and new measurements (e.g., after 6 or 12 mo) might have captured this improvement. Family caregivers in the intervention group with baseline symptoms of depression above 11 also appeared to have had no statistically significant effect of the intervention in terms of reduced symptoms of PTSD. This could indicate that more severe symptoms may require more extensive follow-up (31–34, 41). As such, the “The Caregiver Pathway” may not be enough support for all family caregivers and should perhaps be combined, for example, with follow-up from local community services or referral to other additional support (e.g., general practitioner, social worker, priest, psychiatrist) if needed.

As the ICU nurses often handle family crises combined with providing close monitoring and care to critically ill patients, identifying family caregiver in need of more support is vital for the family caregivers as well as for the ICU nurses. For models such as in this intervention to succeed; however, it is important for managers and employers to provide resources to the ICU nurses to implement and conduct the model in otherwise hectic work environments (14, 43, 44).

“The Caregiver Pathway” (14) was built on the value of narrative (45, 46) along with individual and structured care. The ICU nurses were aware of the need for articulation of family caregivers’ needs, concerns, and emotions, and as a part of the model took family caregivers aside in the middle of the chaotic ICU environment and simply asked them how they were doing. Being met in such a structured, personalized way both at the hospital and in the time after were shown to be associated with reduced symptoms of PICS-F among family caregivers to patients who survived.

This study has some limitations. First, there is sparse information about the potential eligibility of family caregivers during the study as the ICU nurses only forwarded information of those interested. Second, all family caregivers in the ICU received standard care, which is influenced by the physicians and ICU nurses’ bedside and differences in standard care could have been present. Third, three family caregivers were included from each patient in the current study to capture the experiences of a broad sample of caregivers, independent of relation. Analyses with only one caregiver per patient (e.g., partner) could have impacted the results. However, excluding caregivers already included in the study from the analysis could introduce bias and were, therefore, not conducted. Finally, as this is a single-center study using complete cases, and subgroups analyses were based on small sample sizes, potentially impacting the precision of the estimates (e.g., increasing possibility of type II errors), findings should be interpreted bearing these issues in mind.

CONCLUSIONS

“The Caregiver Pathway” intervention could be associated with less symptoms of PTSD, anxiety, and depression after 3 months for family caregivers of patients who survived, thus contributing to reduced presence of PICS-F for these family caregivers.

ACKNOWLEDGMENTS

We thank all the participants in this study without whom this study would not have been possible. Also, we thank the team at the medical ICU (MICU) at Oslo University Hospital, especially the 11 dedicated ICU nurses conducting the intervention, and the stakeholders at MICU for encouraging and facilitating the study.

Supplementary Material

ccm-53-e555-s001.pdf ^{(1MB, pdf)}