MON-LB003 Risk Factors For 30-day Readmission After Diabetic Ketoacidosis Using Nationwide Readmissions Database

Abstract

Introduction: Nearly 1 in 10 individuals in the United States have Diabetes Mellitus [1]. One potential preventable complication is Diabetic Ketoacidosis (DKA). Better understanding of the risk factors for readmissions of DKA will allow the development and implementation of specific patient-centered interventions to decrease future readmissions. We sought out to determine the 30-day all-cause readmission rate for adults (age > 18) admitted with DKA and the associated predictors of readmissions. Methods: We utilized Agency of Healthcare Research and Quality’s (AHRQ) Health Care Utilization Project’s (HCUP) 2014 Nationwide Readmission Database which includes 14.9 Million discharges across 22 states accounting for 51.2% of the total U.S. population and 49.3% of all U.S. hospitalizations to identify admissions with a DKA related ICD-9 diagnosis (250.10, 250.11, 250.12, and 250.13) associated with both Type 1 and Type 2 Diabetes Mellitus. Applicable admissions were all adults (age > 18)with an index hospitalization between January 1 and November 30, 2014. Patients who died during index admission and those with missing covariates were excluded. All-cause readmission including DKA within 30-days of DKA were analyzed. Statistical analysis was completed with Stata 15 (StataCorp, College Station, TX) with p-values < 0.05 considered statistically significant. A univariate and multivariate analysis of data collected was completed using both odds ratio and chi square test for significance. Predictors for readmission were determined using a multivariate logistic regression model following sequential step-wise elimination of covariates including demographics, comorbidities, hospital characteristics, length of stay (LOS) for index admission, and the modified Elixhauser Comorbidity Index. Results: A total of 66,896 patients met criteria for DKA related index admission. Of which, there was 12,954 (19.36%) all-cause readmissions within 30-days including 7,167 were again for DKA accounting for 55.32% of all readmissions. Multivariate analysis showed that the predictors of 30-day readmission were younger age, (with adults age <35 the highest risk), female, disposition at discharge to short term hospital or home health or against medical advice), from a zip code with the lowest income quartile, Medicare as payer, lengthier LOS, presence of comorbidities, absence of obesity, and presence of renal failure. Conclusion: Almost 1 in 5 (19.36%) patients discharged after a DKA admission were readmitted within 30 days. Physician awareness and development of targeted interventions for individuals with risk factors and high-risk for readmissions may help decrease future morbidity and mortality.