Avoidable hospitalizations for diabetes: comorbidity risks

Abstract

This study examined the risk for avoidable diabetes hospitalizations associated with comorbid conditions and other risk variables. A retrospective analysis was conducted of hospitalizations with a primary diagnosis of diabetes in a 2004 sample of short stay general hospitals in the United States (N = 97,526.) Data were drawn from the Health Care Utilization Project National Inpatient Sample. Avoidable hospitalizations were defined using criteria from the Agency for Healthcare Research and Quality to analyze 2 types of ambulatory care sensitive conditions (ACSCs): short-term complications and uncontrolled diabetes. Maternal cases, patients younger than age 18, and transfers from other hospitals were excluded. Avoidable hospitalization was estimated using maximum likelihood logistic regression analysis, where independent variables included patient age, gender, comorbidities, uninsurance status, patient's rural-urban residence and income estimate, and hospital variables. Models were identified using multiple runs on 3 random quartiles and validated using the fourth quartile. Costs were estimated from charge data using cost-to-charge ratios. Results indicated that these 2 ACSCs accounted for 35,312 or 36% of all diabetes hospitalizations. Multiple types of comorbid conditions were related to risk for avoidable diabetes hospitalizations. Estimated costs and length of stay were lower among these types of avoidable hospitalizations compared to other diabetes hospitalizations; however, total estimated nationwide costs for 2004 short-term complications and uncontrolled diabetes hospitalizations totaled over $1.3 billion. Recommendations are made for how disease management programs for diabetes could incorporate treatment for comorbid conditions to reduce hospitalization risk.