16-year trends in 30-day cause specific readmissions following hospitalisation for heart failure in england by sex, socioeconomic status and ethnicity.

Background: Reducing the high cost and burden of early readmissions after hospitalisation for HF has become a health policy priority of recent years, yet prior trend data are conflicting and UK data are scarce. This study investigates 16-year trends in 30-day cause-specific readmissions after hospitalisation for heart failure (HF), by gender, socioeconomic status and ethnicity.

Methods: Using Hospital Episodes Statistics linked to socioeconomic and death data, we identified index admissions for HF in England (2002-2018). Trends in 30-day cause-specific readmissions by gender, socioeconomic status and ethnicity were estimated and patient characteristics of readmission were identified.

Results: There were 698.983 index admissions for HF, median age 81 years [IQR 14], 50% male, 82% White, 3.1% South Asian, 1.7% Black, 16% most and 22% least affluent. In hospital deaths occurred in 108.798 patients (16%), but reduced from 20% to 12% over time (Figure 1a). A further 31.178 deaths (5%) occurred by 30 days. Between 2002-2006 and 2014-2018, age adjusted 30-day readmissions for any cause increased from 19% to 22%, an average increase of 1.4% (95% CI 1.3, 1.5) per annum. Readmissions for HF (6%) and ‘other CVD’ (3%) remained stable but readmissions for non-CVD causes increased from 10% to 13% at a rate of 2.6% (2.4, 2.7) per annum (Figure 1). Rates were similar by gender but there were diverging trends by ethnicity. Black groups experienced a 50% increase in readmissions for HF (6% to 9%, interaction-p 0.03) and South Asian groups had more rapidly increasing rates of readmissions for non-CVD causes (interaction-p 0.04). Non-CVD readmissions (12%) were also more prominent in the least (15%; 15, 15) compared to the most affluent group (12%; 12, 12) (see Table). Strongest predictors for HF readmission were Black ethnicity and CKD, whilst cardiac procedures and cardiologist care during the index admission were protective. For non-CVD readmissions, strongest predictors were non-CVD comorbidities (Cancer, COPD, CVA and depression), whilst cardiologist care was protective.

Conclusions: In HF, despite readmission reduction policies, 30-day readmissions have been increasing. This change is driven by non-cardiovascular causes and impacts the least affluent and ethnic minority groups the most. Readmission reduction programmes require close consideration of patient factors and a multidisciplinary approach to specialist non-cardiovascular care.