Hospital readmission rate in US states: Is it higher where more chronically ill patients cluster?

Monday, June 23, 2014: 10:15 AM
LAW B2 (Musick Law Building)

Author(s): Jayasree Basu

Discussant: Lee R Mobley

Research Objective
Hospital readmission has received growing attention because of its implications for cost and quality of care. A 2011 report by Dartmouth Atlas revealed striking variations in 30-day readmissions rates between hospital regions. Previous studies reported readmission rates among patients to increase with number of chronic conditions. This study examines small area-variations in readmission rates to assess whether higher readmissions rate in an area is associated with higher clusters of chronically ill population.

Study Design
The study uses hospital discharge data of patients in 6 US states,  California, Florida, Moussouri, Nevada, New York, and Tennessee for  2009 in the adult age group (18 and above). The data is obtained from the Healthcare Cost and Utilization Project (HCUP) State Inpatient Databases (SID) of the Agency for Healthcare Research and Quality, linked to contextual and provider data from Health Resources and Services Administration. The states were selected based on the availability of confidential 2009 HCUP-SID data and encrypted patient identifier to calculate readmissions. The unit of analysis is the Primary Care Service Area (PCSA), validated as the smallest primary care market. The study uses a multivariate cross sectional design, regressing area-level  chronic condition indictor on area readmission rate for all causes, after controlling for other area-level characteristics reflecting socio-demographic compositions, income, insurance status, population density and resource availabilities. In a second level of analysis, additional covariates indicating patient severity,  hospital characteristics, (e.g. mortality rates, length of stay, hospital size) of the area, are also included. The study estimated Ordinary Least Squares method (OLS) model using fixed effects by state, after conducting appropriate tests for the model fitness and correcting for heteroscedasticity by the Hubert-White sandwich estimator.

Principal Findings
Controlling for covariates, the study found that readmissions are significantly greater in areas with a higher prevalence of multiple chronic condition discharges. The readmission rates were significantly higher in PCSAs having higher proportions of patients with 2-3 chronic conditions (coefficient=0.44, P<.01), as well as those with 4 + chronic conditions (coefficient=0.25, p<.01), compared to the areas with a higher concentration of 0-1 chronic condition discharges (the reference group). This statistically significant relationship prevailed even after adjusting for area characteristics that reflected average hospital size and quality of care.  In addition to differences across states, higher readmission rates were also found in areas with higher proportions of publicly insured discharges (primarily Medicare), racial ethnic minorities, older adults (85 and above), and those with larger hospitals and higher hospital length of stay.

Conclusions
Using small area analysis, the study shows that areas with higher concentration of more chronically ill population are more likely to have higher readmission rates. The findings indicate that chronic care management is an integral part of reducing spatial variations of health care utilization and costs.

Implications for Policy, Delivery or Practice

The study calls for policies that should lead to interventions identifying higher risk patients before discharge from the hospital, follow-up strategies after discharge, and sufficient payment for management of chronic diseases, in order to achieve lower rates of readmission across areas.