Regional Variation in Mortality and Subsequent Hospitalization of Nursing Residents With Heart Failure

Abstract

Do demographics, clinical characteristics, care structures, mortality, and rehospitalization differ by region among a national sample of nursing home (NH) residents with heart failure (HF)? Retrospective observational study of NH residents with HF by ICD-9 CM codes or Minimum Data Set (MDS) diagnosis, using the Linked Nursing Home/Skilled Nursing Facility Stay File, containing MDS, hospitalization, and mortality data for all residents in a 10% random sample of NHs (n = 1840) during 2003-2004. Facility characteristics, demographics, functional characteristics, comorbidity, and outcomes were described by geographic region. Baseline characteristics and care structures for subjects who experienced a subsequent HF hospitalization or death were compared with those who did not, using unadjusted odds ratios and chi-square tests or Fisher's exact tests for categorical, and t tests or Wilcoxon Rank Sum tests for continuous variables. Predictive logistic regression models for mortality in all subjects and HF hospitalization in subjects with a single NH stay of fewer than 90 days were developed. Time to first HF hospitalization in longer-stay subjects was analyzed using Cox models. Mortality of NH residents with HF exceeds 45%; HF hospitalization exceeds 50% annually. Residing in a facility with 50 or fewer beds or in the rural South were both associated with lower risk of death. Older age and residing in the Midwest were associated with higher risk of HF hospitalization. Age, comorbidity, and functional impairment were not clinically different among regions of the country, but both mortality and HF hospitalization rates differed significantly by region.