ModelHeart: a resident-led implementation of metrics to reduce 30-day heart failure readmissions

Hospitals have been penalised for excessive 30-day readmissions via Medicare payment penalties. As such there has been keen interest in finding ways of reducing readmissions. The basis for the study was a retrospective review of heart failure (HF) admissions at Cleveland Clinic Florida from 1 January 2010 to 31 December 2010. The result of this was a set of metrics associated with >30day span between admissions: N-terminal pro-brain natriuretic peptide by at least 23%, fluid balance of -1.3L and sodium 135mEq/L on discharge. The ModelHeart trial was a prospective resident-led validation of these criteria that consisted of education about and implementation of these metrics. A total of 200 patients carrying a diagnosis of HF, admitted between 1 November 2012 and 14 January 2014 were included in the trial. Of the 200 enrolled patients, 94% of discharged patients met at least one criteria, 58% met at least two criteria and 20% met all three. There were forty-eight all-cause 30-day readmissions. 30-day readmission rates between themore than equal to twocriteria cohort and the remaining patients were not significantly different (p=0.71). Overall readmission rates were higher in the 2011-2012 retrospective patient pool (19%) versus the ModelHeart cohort (11%), and proportional differences were significant, (p<0.001). This may suggest that education provided sufficient awareness to alter discharge practices outside of the measured metrics. However, the lack of significant differences between groups with respect to discharge metrics suggests that further study is needed to refine the metrics and that reducing HF readmissions involves a continuum of care that spans the inpatient and outpatient setting.