Results
Characteristics of Patients and FDMs
Over the study period from August 2012 to September 2013, a total of 176 FDMs and patients met the eligibility criteria and were approached to participate in the study. Seventy-seven FDMs provided data for all study time points and completed the study. Figure 1 is the flow chart of the study. Individuals completing the study tended to be older than individuals lost to attrition (M = 57.01, SD = 13.82; M = 49.53, SD = 13.16; t (105) = –2.55, p = 0.01). The presence of a living will [X (1) = 4.08, p = 0.04] and durable power of attorney [X (1) = 4.08, p = 0.04] was higher in the sample completing the study. The proportion of African-American FDMs in the attrition group was higher than the sample group approaching statistical significance (p = 0.09) (Table A3, Supplemental Digital Content 1, http://links.lww.com/CCM/B207). The demographic characteristics of the hospitalized ICU patients and their FDMs are summarized in Table 1. The ICU mortality rate was 30%, which increased to 43% by 30 days after hospital discharge.
(Enlarge Image)
Figure 1.
Flow chart of the study. FDM = family decision maker.
Posttraumatic Stress Symptoms
The mean IES-R score for the sample was 23.14 (SD = 15.90). Eighteen FDMs (23%) had IES-R scores above 33, and 32 FDMs (42%) had scores above 24. The IES-R score differed among FDMs based on the disposition of the patient at 30 days (deceased: M = 29.70, SD = 15.83; alive: M = 18.23, SD = 14.23; t (75) = –3.34, p = 0.001). The IES-R score also varied according to the relationship of the FDM to the patient [F (3, 73) = 2.96, p = 0.04] with children of ICU patients having the highest IES-R scores (M = 29.18, SD = 16.04) and other FDMs (siblings, power of attorney, etc.) having the lowest IES-R scores (M = 11.44, SD = 7.38). The difference in the mean scores between these two groups was statistically significant using one-way ANOVA with post hoc testing (p < 0.05). The IES-R scores of spouses (M = 22.11, SD = 15.06) and parents (M = 24.10, SD = 19.31) were not statistically different from each other or the other relationship categories. No other differences were noted in IES-R score among the patient and FDM characteristics.
FDM Coping Strategies and Posttraumatic Stress Symptoms
Problem-focused coping scores were highest among the three coping strategies at enrollment and 30 days after hospital discharge as shown in Table 2. Group problem-focused and emotion-focused coping mean scores decreased significantly over time while avoidant coping scores remained statistically unchanged from enrollment to 30 days after hospital discharge. Change scores for the coping strategies demonstrated moderate variation (Figs. 2, 3, and 4). RCI revealed four participants (5%) had a statistically significant decrease in avoidant coping use between T1 and T2, and three individuals (4%) experienced a significant increase over time (RCI < 1.96 and RCI > 1.96, respectively, p < 0.05). Two FDMs (3%) demonstrated a statistically significant decrease in problem-focused coping over time and two participants (3%) reported a decrease in emotion-focused coping use from T1 to T2 based on RCI analysis with no individuals endorsing a significant increase in either coping strategy. Female FDMs used higher amounts of avoidant coping at enrollment (T1) than men (female: M = 1.60, SD = 0.33; male: M = 1.37, SD = 0.37; p = 0.02), but the difference was no longer detected at 30 days following hospital discharge (female: M = 1.63, SD = 0.44; male: M = 1.56, SD = 0.51; p = ns). No other differences in the use of coping strategies were found among FDM characteristics.
(Enlarge Image)
Figure 2.
Distribution of avoidant coping change scores. Black bars indicate Reliable Change Index of < –1.96 or > 1.96 (p < 0.05).
(Enlarge Image)
Figure 3.
Distribution of problem-focused coping change scores. Black bars indicate Reliable Change Index of < –1.96 (p < 0.05).
(Enlarge Image)
Figure 4.
Distribution of emotion-focused coping change scores. Black bars indicate Reliable Change Index of < –1.96 (p < 0.05).
At enrollment (T1), all of the coping strategies demonstrated moderate correlations with each other but avoidant coping was the only strategy to have a significant positive relationship with IES-R score (r = 0.27, p < 0.01) (Table A4, Supplemental Digital Content 1, http://links.lww.com/CCM/B207). Coping strategies 30 days following hospital discharge (T2) continued to exhibit significant positive correlations with each other and all three coping strategies demonstrated a positive relationship with IES-R score (avoidant coping: r = 0.60; emotion-focused coping: r = 0.23; problem-focused coping: r = 0.38; p < 0.05). Furthermore, longitudinal correlations between the coping strategies at T1 and T2 demonstrated moderately positive correlations within a particular type of coping strategy but weaker positive correlations between different T1 and T2 coping strategies (Table A4, Supplemental Digital Content 1, http://links.lww.com/CCM/B207). Additionally, T1 problem-focused coping had no significant relationship to T2 avoidant coping (r = 0.08, p > 0.05).
Predictors of PTSD Symptom Severity
A stepwise regression model identified avoidant coping change score, problem-focused coping change score, patient death, and the child FDM role as significant predictors of higher IES-R scores (Table 3). Parents serving as FDM for their children approached statistical significance (B = 7.83, p = 0.09). The model was moderately strong in predicting severity of later PTSD symptoms accounting for 40% of the variance in IES-R score [F (6, 70) = 14.05; p = 0.000].
T1 coping strategies were poor predictors of IES-R score with an R of 0.09, F (3, 73) = 2.35, and p = 0.08. When the variables of FDM relationship to patient and patient outcome (deceased/alive) identified in the stepwise regression were added, the model provided a statistically significant explanation of IES-R score variance [F (7,69) = 4.14, p = 0.001] with patient death being the only independent predictor of higher posttraumatic stress symptom severity (Table A5, Supplemental Digital Content 1, http://links.lww.com/CCM/B207). In contrast, T2 coping strategy use did explain a statistically significant amount of the variance in IES-R score [F (3, 73) = 19.41, p < 0.001, R = 0.44]. When patient outcome and FDM relationship to patient were added, the variance in IES-R score explained by the model increased slightly (R = 0.50). Avoidant coping and problem-focused coping at T2 were independent predictors of higher IES-R score, whereas death of the patient was no longer a significant independent predictor of IES-R score (Table A6, Supplemental Digital Content 1, http://links.lww.com/CCM/B207).
Having identified that T2 avoidant and problem-focused coping were significant predictors of PTSD symptom severity, we then tested whether these coping strategies mediated the relationship between patient death and later PTSD symptoms. Using the method described by Baron and Kenny, we tested path A (Fig. 5), demonstrating a significant relationship between death of the patient and IES-R score. In path B, we performed regression analysis between patient death with T2 avoidant coping and T2 problem-focused coping. Patient death was a significant predictor of T2 avoidant coping but not of T2 problem-focused coping (β = 0.03, p = 0.77). To complete the test for mediation, IES-R score was regressed on avoidant coping and patient death shown in path C, demonstrating a mediating effect of avoidant coping between patient death and PTSD symptoms.
(Enlarge Image)
Figure 5.
Mediation by T2 avoidant coping between patient death and posttraumatic stress disorder symptoms.