Applicability of the heart failure Readmission Risk score: A first European study. (1st June 2017)

• Record Type: Journal Article
• Authors: Formiga, Francesc; Masip, Joan; Chivite, David; Corbella, Xavier
• Is Part Of: International journal of cardiology. Volume 236(2017)
• Page Start: 304
• Page End: 309
• Publication Date: 2017-06-01
• Subjects: Heart failure -- Acute heart failure -- Readmission -- Readmission Risk score -- Readmission risk stratification; Cardiology -- Periodicals; Electronic journals; 616.12
• DOI: 10.1016/j.ijcard.2017.02.024
• Languages: English
• ISSNs: 0167-5273

Abstract

Abstract: Background: The Readmission Risk score (RR score) has been considered useful to predict Medicare/Medicaid patients' likelihood of 30-day hospital readmission for heart failure (HF). To our knowledge, the accuracy of this prediction model has not been independently validated in other clinical circumstances in Europe. Methods: From July 2013 to December 2014, all patients who survived to a first admission due to decompensated HF at our tertiary care teaching hospital were retrospectively included in the study. The RR score was calculated in all patients to predict future 30 and 90-day unplanned all-cause readmissions. Results: A total of 679 patients were included, of them, 52 patients (7.6%) were readmitted by any cause within 30 days after discharge, and 98 (14.4%) within 90 days. When compared, the average RR scores for patients readmitted was significantly higher to those who did not, either within 30 days (22.7 vs. 20.1) or 90 days (22.7 vs. 20.1) of discharge. The 30-day C-statistic was 0.649 (95% CI 0.574–0.723) and the 90-day 0.621 (95% CI 0.560–0.681). There was a significant increase in readmission percentages at 30 and 90 days with respect to increasing quartiles of RR score. Conclusion: Our results only support a modest applicability of this predictive model in patients at 30 and 90 days, after a first hospitalization for decompensated HF. Probably, the fact that our readmission rate in patients firstly admitted due to HF was very low, generated a bias in the study, discouraging the use of this score in the de novo HF patients.