Beyond classical risk adjustment: Socioeconomic status and hospital performance in urologic oncology.

526 Background: Safety-net hospitals care for more patients of lower socioeconomic status (SES) than non-safety-net hospitals and may be disproportionately punished under readmission risk adjustment models that do not incorporate (SES). We developed a readmission risk adjustment framework incorporating SES to assess impact of SES on safety-net hospital rankings for patients undergoing major surgery for urologic malignancies. Methods: Quasi-experimental design using California Office of Statewide Health Planning and Development data from 2007-2011. Subjects included all patients undergoing radical cystectomy for bladder cancer (n = 3,771), partial nephrectomy (n = 5,556), and radical nephrectomy (n = 13,136) for kidney cancer. Unadjusted hospital rankings and predicted rankings under a base model, which simulated the Medicare Hospital Readmissions Reduction Program model, were compared with predicted rankings under models incorporating socioeconomic status. Socioeconomic status was derived from a multifactorial neighborhood score at the ZIP code level calculated from US Census data. The main outcome measures were hospital rankings based on 30-day all-cause readmission rate and differences between model predicted rankings. Results: For all procedures, the addition of socioeconomic status, geographic, and hospital factors changed the overall hospital rankings significantly compared with the base model (p < 0.01), with the exception of socioeconomic status in radical cystectomy (p = 0.07) and socioeconomic status and rural factors in partial nephrectomy (p = 0.12). For radical nephrectomy and partial nephrectomy, the addition of socioeconomic status and hospital factors significantly improved the mean ranking of safety-net hospitals and improved the ratio of observed relative to expected rankings (p < 0.01). For radical cystectomy there was no significant change in rankings with the addition of socioeconomic status, rural status, or hospital factors. Conclusions: Adding socioeconomic status to existing Medicare readmission risk adjustment models leads to significant changes in hospital rankings, with a differential impact on safety-net hospitals.