The AAMC submitted comments March 28 to the Centers for Medicare and Medicaid Service (CMS) regarding potential future enhancements to the Hospital Compare Overall Hospital Quality Star Ratings methodology.
The AAMC made the following specific recommendations to CMS as it reexamines the methodology for the Star Ratings:
- Suspend the Star Ratings: CMS should remove the publication of the Star Ratings from the Hospital Compare website until CMS is able to address significant concerns with the methodology.
- Improve Underlying Measures: CMS should improve existing measures in use in the hospital quality reporting and performance programs, including the incorporation of sociodemographic factors in measure-level risk adjustment. CMS should remove PSI-90 from the Star Ratings.
- Overall Composite Ratings Add to Confusion About Hospital Confusion: A rating that combines all of the multiple dimensional aspects into a summary score may not provide a patient or consumers with the information that is truly important for an individual’s situation. The AAMC urges CMS to explore the template matching, or other approaches that directly compare patient groups, as a possible alternative model to use for rating hospitals.
- Measure Groupings: CMS should undertake further analysis on how to improve measure groupings before implementing any changes.
- Regrouping Measures: The AAMC does not support the proposed alternative grouping for the Safety of Care group or the use of PSI components in lieu of the PSI-90 composite measure. CMS should consider simpler alternative approaches before implementing any regrouping of measures.
- Incorporating Precision of Measures: CMS should continue to analyze and share more information on potential approaches to improving the incorporation of measure precision in the ratings and be transparent in advance of implementing changes.
- Period-to-Period Star Rating Shifts: The AAMC supports the proposal to move to an annual update cycle, until there are further improvements in the reliability and stability of the methodology, and further exploration of “partial-star” Overall Hospital Ratings. We believe the drawbacks outweigh the possible benefits of incorporating weighted averages to address period-to-period shifts.
- Peer Grouping: The AAMC remains supportive of peer grouping and believes stratified comparisons are useful to patients and consumers to best understand the different types of hospitals available to them, especially as a short-term solution to the broader need to develop more rigorous risk adjustment at the measure-level. We urge CMS to ensure that stratified comparisons of hospital performance are clear when published on Hospital Compare.
- Explicit Approach: The AAMC supports additional work around the development of an explicit approach to replace the Latent Variable Modeling.
- Clustering Alternative: CMS should consider alternatives to the current k-means clustering, with a focus on an approach that allows for predictable, fixed targets.
- User-Customized Star Rating: The AAMC is supportive of exploring user-customization to the Star Ratings, but believes additional investigation and analysis is needed to better understand patient and consumer interest and to ensure that customized ratings are reliable and valid.