The impact of self-care behaviour on health-related quality of life across functional dependency phenotipes: insights from a real-world cohort of 1120 patients with chronic heart failure

Abstract Background Poor self-care is associated with higher risk in heart failure (HF). Self-care may influence patient-reported outcomes (PROMs) such as health-realted quality of life (QoL). Functional and instrumental autonomy are determinants of QoL. Literature has shown instrumental and/or functional dependency are associated with worse QoL. Purpose Describe the effect of self-care on QoL across to different levels of dependency and explore the potencial interaction between self-care and autonomy. Methods 1120 patients with chronic HF from DAMOCLES study were analyzed. We used modified European Heart Failure Self-care Behavior Scale 9-item version (EHFSCBS-9) to measure self-care, scores were inverted. To assess QoL, we used Minnesota Living with Heart Failure Questionnaire (MLHFQ). Functional dependency for basic activities of daily living (BADL) was evaluated using Barthel index (BI). Dependency for instrumental activities (IA) was evaluated using Lawton and Brody questionnaire (LBQ). Regarding dependency, patients were grouped in 3 phenotypes: independent (normal BI and LBQ, phenotype 1), dependency for IA (abnormal LBQ, phenotype 2) and dependency for BADL and IA (abnormal BI ± abnormal LBT, phenotype 3). General linear models (GLM) were constructed to explore interaction between dependency phenotypes and poor self-care (defined as EHFSCBS-9 ≤lower tertile) on adjusted MLHFQ scores and to obtain the estimated marginal means of QoL scores across phenotypes and levels of self-care. All models were adjusted by determinants of HF severity. Results Mean age was 72±11 years, 43% were women and mean LVEF was 45±17%. Mean MLHFQ scores were 45±24. Mean EHFSCBS-9 scores were 69±28. There was a significant inverse correlation between EHFScBS-9 and MLHFQ scores (r=−0.158; p-value <0.0001). Unadjusted MLHFQ scores were consistenly higher in the phenotype 3 (51.5±21.3) compared to phenotype 2 (41.8±25.4, p-value<0.001) or phenotype 1 (41.6±25.6, p-value <0.001). In adjusted GLM, poor global self-care (F=15.7; p-value <0.001) was an independent predictor of higher adjusted scores in MLHFQ. Interaction term dependency phenotypes by poor self-care was not significantly associated with MLHFQ scores (F=0.297; p-value = 0.743; Figure 1): the gap in terms of QoL between self-care strata was observed across dependency phenotypes. Estimated adjusted marginal means of MLHFQ scores stratified according to dependency phenotypes and levels of self-care are displayed in Table 1. In phenotypes 2 or 3, MLHFQ scores were lower in patients with preserved self-care. In phenotype 1 differences in QoL across self-care strata were present but not significant. Conclusions Self-care behaviour is an independent predictor of health-related QoL in HF regardless of the level of patient's autonomy. Better self-care was associated with better QoL particularly in patients with dependendecy to IA and/or BADL. Impact of self-care on QoL was attenuated in patients with normal autonomy. Funding Acknowledgement Type of funding sources: None.