Does Favorable Selection Among Medicare Advantage Enrollees Affect Measurement of Hospital Readmission Rates?. (August 2014)

• Record Type: Journal Article
• Authors: Wong, Edwin S.; Hebert, Paul L.; Maciejewski, Matthew L.; Perkins, Mark; Bryson, Chris L.; Au, David H.; Liu, Chuan-Fen
• Is Part Of: Medical care research and review. Volume 71:Number 4(2014:Aug.)
• Page Start: 367
• Page End: 383
• Publication Date: 2014-08
• Subjects: Veterans -- Medicare -- readmission -- risk adjustment -- selection bias; Medical care -- United States -- Periodicals; Medical care -- Periodicals; Medical care -- Research -- United States -- Periodicals; Medical care -- Research -- Periodicals; Medical economics -- United States -- Periodicals; Medical economics -- Periodicals; 362.10973
• DOI: 10.1177/1077558714533823
• Languages: English
• ISSNs: 1077-5587

Abstract

Literature indicates favorable selection among Medicare Advantage (MA) enrollees compared with fee-for-service (FFS) enrollees. This study examined whether favorable selection into MA affected readmission rates among Medicare-eligible veterans following hospitalization for congestive heart failure in the Veterans Affairs Health System (VA). We measured total (VA + Medicare FFS) 30-day all-cause readmission rates across hospitals and all of VA. We used Heckman's correction to adjust readmission rates to be representative of all Medicare-eligible veterans, not just FFS-enrolled veterans. The adjusted all-cause readmission rate among FFS veterans was 27.1% (95% confidence interval [CI] = 26.5% to 27.7%), while the adjusted readmission rate among Medicare-eligible veterans was 25.3% (95% CI = 23.6% to 27.1%) after correcting for favorable selection. Readmission rate estimates among FFS veterans generalize to all Medicare-eligible veterans only after accounting for favorable selection into MA. Estimation of quality metrics should carefully consider sample selection to produce valid policy inferences.