The updated european heart failure self-care behaviour scale (EHFScBS): a refined evidence-based patient reported outcome measure for assessing heart failure self-care

Background and aim

Recent evidence and updated guidelines have challenged traditional recommendations regarding fluid and sodium restriction. In light of this, we aimed to adapt the European Heart Failure Self-care Behaviour Scale (EHFScBS) to reflect current evidence-based recommendations by removing items related to fluid and sodium restriction. This study investigates the psychometric properties of this revised version of the EHFScBS.

Methods

Baseline data from a previous randomised controlled study evaluating effects of exergaming (the HF Wii trial) were used for the psychometric testing. Data were collected using validated versions of the EHFScB-9 in Swedish, Dutch, German, Italian, Hebrew and English. The 9 items with self-care statements are rated on a five-point Likert scale. A standardised score from 0 to 100 with a higher score indicating better self-care. The factor structure of the EHFScB-9 was evaluated using Confirmatory Factor Analysis (CFA).

Results

A total of 605 patients were included (mean age 67 years (SD=12), and 71% (n=430) males). The mean time since diagnosis with heart failure was 4 years. In total, 24 patients (9%) were in NYHA-class I, 350 (58%) were in NYHA-class II, 171 (23%) were in NYHA-class III, and 4 patients (1%) were in NYHA-class IV.

The unidimensional models, in which items 5 (fluidrestrction) and item 7 (saltrestriction) were excluded, presented excellent model fit with RMSEA ≤ 0.06 (90% confidence interval 0.04-0.08) and CFI, TLI > 0.95.

The original two-factor 9-item version (factor 1: Adherence to regimen and factor 2: Consulting behavior) presented a good model fit. The unidimensional models, in which items 5 (fluidrestriction) and 7 (saltrestriction) were excluded, both presented excellent model fit. All two-factor models presented moderate to strong standardised factor loadings (all above 0.3), and the factor correlations ranged from 0.65 - 0.80.

For the unidimensional models, the standardised factor loadings were moderate to strong and ranged from 0.30 to 0.90. For the two-factor model, Adherence to regimen, were < 0.7. As expected, a decline in reliability estimates was observed as the number of items was reduced. For the second factor, Consulting behavior, the reliability was > 0.82. For the unidimensional models, the ordinal alpha and composite reliability ranged from 0.81 to 0.85 across the models, and Cronbach's alpha ranged from 0.75 to 0.77

Conclusions

The results opens up for an opportunity to exclude self-care items with less evidence from the EHFScB. Considering the clinical evidence on relevant self-care and based on the psychometric evaluation, both items on salt and fluid restriction can be excluded from the EHFScB. It is recommended to only use the unidimensional version of the 7 item version of the instrument, considering the low reliability for the first factor in the two-factor model. The suggested version of the EHFScB requires further testing in other contexts.