Abstract 17032: Case Mix Predicts Hospital Performance on Measures of Patient Experience

• Published in: Circulation (New York, N.Y.) Vol. 124; no. suppl_21
• Main Authors: Joynt, Karen E, Orav, E. John, Jha, Ashish K
• Format: Journal Article
• Language: English
• Published: 22.11.2011
• ISSN: 0009-7322; 1524-4539
• DOI: 10.1161/circ.124.suppl_21.A17032

Abstract only The Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) is the national assessment of hospital performance on patient experience with care, with over 95% of acute-care U.S. hospitals participating. The Value-Based Purchasing (VBP) program ties Medicare payments to hospitals’ performance on this metric. Some have expressed concern that sicker patients may rate their hospital experience more poorly, potentially putting hospitals that care for the sickest patients at a disadvantage. However, we lack empirical data to support or refute this notion. We used the latest available national HCAHPS and Medicare data to assess whether hospitals’ performance on HCAHPS measures varied by case mix, defined by patients’ predicted mortality. We examined how variations based on case mix compared with variations based on structural characteristics such as size and teaching status. We found that hospitals that cared for sicker patients were more often urban, non-profit hospitals that were larger, located in the northeast or west and more likely to have a teaching affiliation. In bivariate analyses, we found that hospitals with the sickest patients (those in the highest quartile of predicted mortality) had substantially lower HCAHPS performance compared with hospitals with the least sick patients (lowest quartile of predicted mortality) (as measured by proportion of patients giving the hospital a 9 or 10, 62.6% versus 69.4%, p<0.001). In multivariable analyses adjusting for size, teaching status, and other hospital characteristics, the differences in HCAHPS performance across quartiles of predicted mortality narrowed somewhat (63.3% versus 68.5%, p<0.001). The differences based on case mix were larger than those based on size, teaching status, location, or any of the other covariates we examined. As Medicare implements VBP, our findings suggest that hospitals with the sickest patient population are at greater risk of being financially penalized.