Complication Rates, Hospital Size, and Bias in the CMS Hospital-Acquired Condition Reduction Program. (November 2017)

• Record Type: Journal Article
• Authors: Koenig, Lane; Soltoff, Samuel A.; Demiralp, Berna; Demehin, Akinluwa A.; Foster, Nancy E.; Steinberg, Caroline Rossi; Vaz, Christopher; Wetzel, Scott; Xu, Susan
• Is Part Of: American journal of medical quality. Volume 32:Number 6(2017:Nov./Dec.)
• Page Start: 611
• Page End: 616
• Publication Date: 2017-11
• Subjects: Medicare payment -- hospital-acquired conditions -- Patient Safety Indicators -- value-based purchasing; Medical care -- Quality control -- Periodicals; Medical care -- Evaluation -- Periodicals; Medical care -- Utilization review -- Periodicals; Medical care -- United States -- Periodicals; 362.1
• DOI: 10.1177/1062860616681840
• Languages: English
• ISSNs: 1062-8606

Abstract

In 2016, Medicare's Hospital-Acquired Condition Reduction Program (HAC-RP) will reduce hospital payments by $364 million. Although observers have questioned the validity of certain HAC-RP measures, less attention has been paid to the determination of low-performing hospitals (bottom quartile) and the assignment of penalties. This study investigated possible bias in the HAC-RP by simulating hospitals' likelihood of being in the worst-performing quartile for 8 patient safety measures, assuming identical expected complication rates across hospitals. Simulated likelihood of being a poor performer varied with hospital size. This relationship depended on the measure's complication rate. For 3 of 8 measures examined, the equal-quality simulation identified poor performers similarly to empirical data (c-statistic approximately 0.7 or higher) and explained most of the variation in empirical performance by size (Efron's R 2 > 0.85). The Centers for Medicare & Medicaid Services could address potential bias in the HAC-RP by stratifying by hospital size or using a broader "all-harm" measure.