Hospital safety‐net status and postdischarge outcomes: The impact of socioeconomic status and Medicare post‐acute care types

Abstract

AbstractAimTo examine the impact of socioeconomic status (SES) and postacute care (PAC) locations on the association between hospital safety‐net status and 30‐day postdischarge outcomes (readmission, hospice use, or death).MethodMedicare Current Beneficiary Survey (MCBS) participants during 2006–2011 who were Medicare Fee‐for‐Service beneficiaries aged 65.5 years or older were included. The associations between hospital safety‐net status and 30‐day post‐discharge outcomes were evaluated by comparing the models with and without PAC and SES adjustments. Safety‐net hospital status was defined as being in the top 20% of hospitals ranked by hospital‐level percent of total Medicare patient days. SES was measured using individual‐level SES (dual eligibility, income, and education) and the Area Deprivation Index (ADI).ResultsThis study identified 13,173 index hospitalizations for 6,825 patients; 1,428 hospitalizations (11.8%) were in safety‐net hospitals. The average unadjusted 30‐day hospital readmission rate was 22.6% in safety‐net hospitals versus 18.8% in nonsafety‐net hospitals. Regardless of whether patient SES status was controlled or not, safety‐net hospitals had higher estimated probabilities of 30‐day readmission (ranging from 0.217 to 0.222 vs. 0.184 to 0.189), and lower probabilities for having neither readmission nor hospice/death (0.750–0.763 vs. 0.780–0.785); for models additionally adjusted for PAC types, safety‐net patients had lower rates of hospice use or death (0.019–0.027 vs. 0.030–0.031).ConclusionsThe results suggested that safety‐net hospitals had lower hospice/death rates but higher readmission rates relative to outcomes at nonsafety‐net hospitals. Readmission rate differences were similar regardless of patients' SES status. However, the rate of hospice referral or death rate was related to SES, which suggested that the outcomes were affected by SES and PAC types.