Abstract 292: Predictors Of Heart Failure Readmission Rates, A Stanford Experience.

Abstract

Introduction: The prevalence of heart failure (HF) is approximately 5.8 million, one in every five Medicare patients with HF are readmitted within 30 days. Heart failure readmissions cost 17.4 billion dollars annually. A major focus of health policy reform has been to contain costs by preventing readmission. Medicare has proposed criteria for HF reimbursement among which are readmissions within a 30-day period after the index hospitalization will not be reimbursed. The Medicare core measure aggregate has been sparsely validated in the literature and such validation of the core measures is essential. The aim of our study was to evaluate the predictive value of the Medicare core measure aggregate and if this aggregate is an effective proposed benchmark in reduction of HF readmission. Methods: A database was developed using RedCap (Research Electronic Data Capture), to study all patients admitted to Stanford Hospital and Clinics from September 2011 through January 2012. Inclusion criteria included those with a primary or secondary diagnosis of heart failure. Cardiac variables of interest beyond the core measures included among others, demographics, length of stay, and past medical history. Aggregate data from the Redcap database was analyzed by standard methods. Results: 101 patients were indexed during the retrospective cohort analysis. 100 were discharged alive. The overall 30-day readmission rate was 14%. There was no difference in readmission rates between patients who had core measures completed versus those who did not (13.8% vs. 14.3%; p= 0.944). In univariate Chi-square analyses, a phone call to the patient by a nurse within 24-48 hours post discharge was associated with significant reduction in rehospitalization rates (7.1% vs. 22.7%; p=0.026). Patients who were readmitted were found to have significantly more prior hospital admissions (3.0 +/- 2 vs. 1.7 +/- 2; p = 0.011) and prior ED visits (1.6 +/- 2 vs. 3.1 +/- 2; p=0.008). In multivariate logistic regression analysis, the phone call was associated with a 75% reduction in readmission (adjusted odds ratio = 0.254, 95% CI 0.066-0.981; p= 0.047). The predictor variable, systolic dysfunction (EF < 55%) approached significance, with an adjusted odds ratio of 6.962, 95% CI 0.979-49.5; p 0.052. Conclusions: High readmission rates for HF are present despite advances in its management. The achievement of Medicare core measure variables was not associated with a reduction in 30-day HF readmission. The reasons for this finding are still unclear, perhaps hospitalizations are in need to meet Medicare core measure guidelines rather then tailor the discharge care of each patient. It is unexpected and noteworthy that a phone call made by a nurse to the patient post discharge was associated with a 75% reduction in 30-day readmission rates. Predictors of HF readmission remain complex and an ideal aggregate for predictors in readmission have yet to be defined.