Does Risk-Adjustment for Sociodemographic Status (SDS) Have an Impact on Hospital Performance?

Co-author: Dawn Sievert

 

Ongoing research examines the impact of applying sociodemographic status (SDS) risk-adjustment to the Centers for Medicare & Medicaid Services (CMS) quality reporting programs. The main driver for investigation is a community assertion that SDS risk-adjustment is necessary to accurately gauge the quality of care given to patients of diverse backgrounds. Contrary to popular belief, recent research produced evidence that SDS risk-adjustment has little to no impact on hospital performance ratings.

 

The National Quality Forum (NQF) convened an SDS Expert Panel to consider if, when, and how outcome performance measures should be adjusted to address the community’s concern. The NQF SDS trial period is a two-year project that tests all measures submitted after April 15, 2015, as well as those undergoing endorsement maintenance review during the trial period, using two models: one with SDS factors incorporated and one without.

 

Measure developers completed and presented results on nearly half of the 16 admissions/readmissions measures from this period. The available SDS variables, including race, ethnicity, gender, age, payer information, income, zip code, and the AHRQ-validated SES index score, were evaluated within the measures. Thus far, findings indicate that most of the measures perform well without the inclusion of SDS factors. The results comparing the models with and without SDS adjustment were highly correlated. The addition of SDS risk-adjustment variables in measure models often had little or no effect on hospital performance.

 

In fact, measure developers acknowledged that where there is an apparent impact on performance, hospitals themselves may be contributing to the disparities in outcomes for socioeconomically disadvantaged groups. Given these findings, and the complex pathways that could explain any relationship between SDS factors with readmission, adjusting for these factors could obscure true signals of care quality. For all of these reasons, measure developers do not intend to incorporate SDS variables in most of the approved measures.

 

These findings stress the importance of evaluating all solutions, and possibly developing new measures, to reflect all factors that influence readmission risk for socioeconomically disadvantaged patients. The measure developers will share the NQF SDS trial period recommendations with the NQF Standing Committee and CMS for consideration. In addition, the Office of the Assistant Secretary for Planning and Evaluation (ASPE) is researching the impact of SDS on quality measures, resource use, and other measures under the Medicare programs, as directed by the Improving Medicare Post-Acute Care Transformation (IMPACT) Act.

 

CMS received strong pushback from policy makers, providers, and hospital organizations over perceived unfairness of the missing SDS-adjustment in the current measures. As was widely reported, the pushback played a role in CMS’ decision to delay the public reporting of the Overall Hospital Quality Star Rating until a future Hospital Compare release. CMS will continue to monitor the work done by NQF and ASPE and ensure the most accurate representation of hospital performance. Based on the results to-date, however, it is not clear how, if at all, SDS risk-adjustment will change the reported outcomes.