BACKGROUND
Safety‐net hospitals (SNHs) care for more patients of low socioeconomic status (SES) than non‐SNHs and are disproportionately punished under SES‐naive Medicare readmission risk‐adjustment models. This study was designed to develop a risk‐adjustment framework that incorporates SES and to assess the impact on readmission rates.
METHODS
California Office of Statewide Health Planning and Development data from 2007 to 2011 were used to identify patients undergoing radical cystectomy (RC) for bladder cancer (n = 3771) or partial nephrectomy (PN; n = 5556) or radical nephrectomy (RN; n = 13,136) for kidney cancer. Unadjusted hospital rankings and predicted rankings under models simulating the Medicare Hospital Readmissions Reduction Program were compared with predicted rankings under models incorporating SES and hospital factors. SES, derived from a multifactorial neighborhood score, was calculated from US Census data.
RESULTS
The 30‐day readmission rate was 26.1% for RC, 8.3% for RN, and 9.5% for PN. The addition of SES, geographic, and hospital factors changed hospital rankings significantly in comparison with the base model (P < .01) except for SES for RC (P = .07) and SES and rural factors for PN (P = .12). For RN and PN, the addition of SES predicted lower percentile ranks for SNHs and thus improved observed‐to‐expected rankings (P < .01). For RC, there were no changes in hospital rankings.
CONCLUSIONS
SES is important for risk adjustments for complex surgical procedures such as RC. Patient SES affects overall hospital rankings across cohorts, and critically, it differentially and punitively affects rankings for SNHs for some procedures. Cancer 2018. © 2018 American Cancer Society.