The futile cycle of hospital readmission in patients with diabetes.

Abstract

In 2012, the Centers for Medicare and Medicaid Services (CMS) began reducing Medicare payments to hospitals with higher than expected readmissions within 30 days of discharge. This led to focused efforts at many institutions for identifying patients at high risk for readmission and introducing programs to ameliorate this risk. Current estimates indicate that 25% of patients hospitalized in the US have diabetes as a primary or secondary diagnosis. Approximately 10–20% of patients with diabetes are readmitted within 30 days of hospital discharge, with higher rates of readmission observed in those with more comorbidities. Based on readmission data from the AHRQ Healthcare Cost and Utilization Project, diabetes accounts for 10% of readmission costs among patients with public and private insurance, highlighting the importance of identifying factors that can potentially be addressed to reduce these numbers. It is in this context that Rubin and colleagues developed a tool for predicting risk for readmission in patients with diabetes, the Diabetes Early Readmission Risk Indicator (DERRITM). The DERRI model was developedby examining46variables aspredictors of readmission risk in retrospective training and validation cohorts of 44,203 patient discharges from one urban academic medical center. Using multivariable logistic regression, ten variables were identified as statistically significant but modest predictors of readmission (Table 1). In the current report, Rubin et al. examine a subset (n = 8189) of the original parent DERRI population with CVD (DERRI-CVD), a group identified as being at high risk for readmission. Ten variables are again identified as being predictive of readmission, five of which overlap with the parent DERRI population (Table 1). What is most striking from both DERRI and DERRI-CVD is the observation that patients who were discharged from the hospital within 90 days of the index hospitalization were most likely to be readmitted within the next 30 days (odds ratio of approximately 2.0), indicating a cycle of hospital discharge and readmission. Using DERRI-CVD patients, the authors observed a progressive increase in readmissions according to quintiles of risk, reaching 38.6% of patients in the highest quintile. Other variables identified as predictive for readmission included education level, employment status, hospital proximity, and psychiatric disease all of which may serve as surrogate markers of socioeconomic status (SES), which has been identified in prior studies investigating readmission risk. These SES factors present complex challenges to a health care system that is currently in a state of flux in the US. Despite the large number of variables included in DERRI and DERRI-CVD, other variables that may have contributed to diabetes-related readmissions were not examined. These include type anddurationof diabetes,measuresof glycemic control prior to andduring the index hospitalization, type of diabetes medications prescribed at admission and discharge, the ability of patients to fill prescribed medications at discharge, and whether any form of diabetes education

DOI: 10.1016/j.jdiacomp.2017.06.006

1 Figure or Table

Cite this paper

@article{Swami2017TheFC, title={The futile cycle of hospital readmission in patients with diabetes.}, author={Janya Swami and Mary T Korytkowski}, journal={Journal of diabetes and its complications}, year={2017}, volume={31 8}, pages={1252-1253} }