Predictors of Frequent Readmissions in Patients With Heart Failure.

BACKGROUND:Patients with heart failure (HF) have a high incidence of hospital readmissions. However risk models that explore predictors of a single readmission may be less useful at identifying the patients with frequent readmissions who contribute to a disproportionately large proportion of morbidity and health care costs. METHODS:A total of 6252 patients enrolled in the Management of Cardiac Failure Program (MACARF) in Northern Sydney Area Hospitals between 1998 and 2015 were randomly divided into derivation and validation cohorts to create and test a risk model for predictors of ≥2 readmissions or death within 1year of initial hospitalisation for HF. RESULTS:Multivariate predictors of frequent (≥2) readmissions or death were a history of ischaemic heart disease and chronic kidney disease, being unmarried, having anaemia, low serum albumin, elevated creatinine, prolonged hospital stay (>7 days), and not receiving beta blockers on discharge. Event rates increased with a higher risk score (p<0.001) and the prediction was similar in the validation and derivation cohorts (p=0.588). The C-statistic was 0.65. CONCLUSIONS:Our risk score may assist in focussing health care resources and interventions by identifying the subset of HF patients at increased risk for a disproportionately high burden of disease.