Jennifer Gaudet Hefele Leads Research on Care Disparity at Nursing Homes

Reports examining racial and ethnic disparities in the quality of nursing home care typically focus on the very big picture – the collective experience of thousands of residents living in many facilities. Should the same kind of information about individual nursing homes be publicly reported as well?

It’s commonly acknowledged that in the U.S., older Black and Hispanic adults often reside in lower-quality nursing homes, which focuses debate on access to high-quality facilities rather than the care of individual residents. That begs another question: Do disparities still exist between residents of different races and ethnicities within the same nursing homes?

Assistant professor Jennifer Gaudet Hefele of the UMass Boston Gerontology Department, along with six co-authors, recently published the first study to examine within-facility differences in nursing home quality by race and ethnicity for a large set of publicly reported quality measures nationwide.

The results: Overall differences were small and sometimes indicated slightly poorer care of white residents, on average. But the range of quality scores at some individual homes, based on race and ethnicity, was substantial. The authors concluded that their results supported calls favoring public reporting of quality measures based on race and ethnicity at the individual nursing-home provider level.

Although the mean differences were small, “the range of differences suggest that care for residents of different races and ethnicities can be alarmingly different in the same nursing home,” Hefele and her co-authors wrote in an article describing their work in the Joint Commission Journal on Quality and Patient Safety.

“Some facilities had scores for residents of different races and ethnicities that differed by as much as 30 percentage points,” they wrote.

The study examined data from more than 15,000 nursing homes and 3 million residents. It tracked eight nursing home quality measures used by the Centers for Medicare and Medicaid Services. Those measures are published publicly on Nursing Home Compare report cards for individual homes.

The study focused on nursing homes that cared for at least 30 white residents and a minimum of 30 residents who were black or 30 who were Hispanic. Although a majority of the measures showed statistically significant differences across racial and ethnical groups, most of those differences were less than 2 percentage points. But the authors point to the sometimes wide range of quality-measure results as an important factor.

Beyond showing whether and how quality varied by race and ethnicity within the same nursing homes, the study also verified that such stratified, nursing-home level measures were statistically feasible.

Stratifying measures creates smaller sample sizes and can result in statistically unstable quality measures, an often-cited obstacle in discussions about measure stratification. The study by Hefele and her colleagues compared the stability of the stratified measures over four years to that of the overall, publicly reported measures and found similar levels of stability.

They concluded that the stratified measures pass the same levels of acceptability as the CMS-reported measures, and are thus statistically feasible to construct and report.

The authors said the results of their research showed the value of publicly reporting of stratified quality measures, as it could provide a basis for action for nursing homes, consumers and policymakers if necessary. But they cautioned that further enquiry into measure validity and causal factors was needed before such public reporting takes place.

“Our work provides support for the call to explore the possibility of publicly reporting stratified nursing home measures,” the authors wrote. “Although the chances of selecting a nursing home that delivers wildly different care to residents of different racial/ethnic groups is low, it is clearly a chance that many would rather not take.”

Leave a Reply

Your email address will not be published. Required fields are marked *