Abstract 12: Impact of Community Factors on Readmission Rates

Abstract

Objective Readmissions in the Medicare population are common and expensive, and have thus become a major target for intervention by federal and state policymakers. Starting in 2012, the Centers for Medicare and Medicaid Services will penalize hospitals with high readmission rates, in hopes of reducing readmissions nationally. However, many worry that variations in readmission rates are driven less by the quality of the hospital’s care and more by community-level factors such as socioeconomic status and availability of primary care. Empirical data here would be immensely helpful. Therefore, we sought to understand to what extent variations in readmission rates across communities are driven by the quality of hospital care in those communities versus socioeconomic factors and supply-side factors. Methods We used national Medicare data and Dartmouth Atlas data to calculate hospital referral region (HRR)-level readmission rates for congestive heart failure (CHF). We examined the impact of community poverty levels, racial makeup, and supply of physicians and hospital beds on readmission rates using linear regression models. We also examined the impact of patient case-mix and hospital quality performance on readmission rates. Finally, using partial r-square values from multivariable models, we examined the degree to which community-level variations in readmission rates are driven by variation in the sickness of the population, the quality of the hospital care, the supply of physicians and hospital beds, and the socioeconomic makeup of the population. Results Readmission rates across communities (as defined by HRRs) ranged from 10% to 32% for CHF. In univariate analysis, HRRs with higher poverty rates, a higher proportion of black patients, more physicians and hospital beds in the community, and a sicker population had higher readmission rates. In multivariate models that examined the degree to which variation in readmission rates for CHF between HRRs was explained by each of these different community-level factors, we found that supply-side variables (physician and bed supply in the community) were most important (explaining 17% of the variation) followed by socioeconomic characteristics of the community (poverty rate and racial makeup) at 9%. Differences in hospital quality explained 5% of the variation in readmission rates and differences in case mix explained 4%. Conclusions and Implications Community-level socioeconomic variables and supply-side variables play a much larger role in explaining variations in readmissions than the quality of hospital care or underlying sickness of the population. These findings suggest that the current federal initiatives to penalize hospitals may not be optimally targeted. Focusing on community-level factors such as the supply and mix of physicians and targeting efforts towards poor and minority communities may be more fruitful approaches to reducing readmissions. Otherwise, we risk targeting the providers that care for disenfranchised patients in ways that are likely to worsen existing disparities in care.