Background “Self-efficacy” (SE) describes the confidence in one’s ability to perform self-care (e.g. seeking help before decompensation). The Self-Efficacy Scale (two items; 5-point Likert scale) is part of the widely used Kansas City Cardiomyopathy Questionnaire (KCCQ). The clinical utility of this domain in patients with heart failure (HF) has not been defined yet.

Methods This is a post-hoc analysis of the Extended Interdisciplinary Network Heart Failure (E-INH) study, which investigated the long-term effects of a HF nurse-led remote patient management (RPM) program (HeartNetCare-HF^TM [HNC]). E-INH randomized 1,022 patients with left ventricular ejection fraction <40% into HNC vs Usual Care). SE was assessed with the KCCQ-SE Scale (range 0 to 100 with higher values indicating better SE). For prognostic analyses, SE was dichotomized at the median (“low” ≤62.5; “high” >62.5). Depressed mood was measured using the Patient Health Questionnaire (PHQ-9). Disease-specific quality of life (QoL) was assessed by the KCCQ-QoL subscale. Correlates of SE were sought using chi² test and Spearman's coefficient (ρ). Cox proportional hazards models were used for the association between SE and the E-INH study’s combined primary endpoint, i.e. time to all-cause death or rehospitalization.

Results Out of 1022 patients, 855 patients (84%) provided complete SE scores at baseline. At baseline, median SE was 62.5 (quartiles: 50, 87.5), with 52% of patients presenting with low SE. Mean age was 67±12 years, 72% were male. 360 patients (42%) were in NYHA class III or IV. Median scores for QoL and PHQ were 50 (33, 67) and 15 (12, 20), respectively.

Higher SE at baseline was associated with male sex (median SE: men 75 vs. women 62.5; p=0.003), living in a partnership (median SE: in partnership 75 vs. living alone 62.5; p<0.001) and NYHA class (median SE: I/II 75 vs. III/IV 62.5; p<0.001). Further, SE was inversely associated with age (ρ=-0.091, p=0.008), scores for PHQ (ρ=-0.25, p<0.001), and QoL (ρ=-0.27, p<0.001). SE was not associated with the number of living relatives or number of comorbidities. 215 out of 514 patients (42%) showed a relevant change (> 12.5 score points) of SE at the end of the treatment period (18 months).

In the course of 10 years of follow-up, 543 (64%) patients died and 258 (31%) experienced a first rehospitalization event. Overall, low SE predicted death or rehospitalization both after 5 years (HR 1.27, 95%CI 1.05-1.53; p=0.012) and after 10 years (HR 1.28, 95%CI 1.09-1.51; p=0.003). Adjustment for age and sex did not materially alter this association, which was mainly carried by the prognostic utility for death. The age- and sex adjusted risk of death after 10 years for patients with low SE was higher in patients in the usual care arm compared to HNC (HR 1.30, 95%CI 1.04-16.4; p=0.024). Compared to usual care, the mean life-span gained in subjects with low SE in HNC was 269 days in the course of 10 years.

Conclusion SE can easily be measured, as it is composed of 2 items only. Besides age and sex, SE was strongly associated with important factors mediating risk in HF, including depressed mood, QoL, living in partnership, and NYHA functional class. High SE after discharge from hospital appeared to be a powerful factor indicative of a longer life expectancy. Importantly, patients with low SE had a lower mortality risk if supported by a HF nurse led program.

Figure