The Metrics Matter: Improving Comparisons of COVID-19 Outbreaks in Nursing Homes

Katherine E.M. Miller, Rebecca J. Gorges, R. Tamara Konetzka, Courtney H. Van Houtven

Research output: Contribution to journalArticlepeer-review

Abstract

Objectives: In the United States, nursing facility residents comprise fewer than 1% of the population but more than 40% of deaths due to Coronavirus Disease 2019 (COVID-19). Mitigating the enormous risk of COVID-19 to nursing home residents requires adequate data. The widely used Centers for Medicare & Medicaid Services (CMS) COVID-19 Nursing Home Dataset contains 2 derived statistics: Total Resident Confirmed COVID-19 Cases per 1000 Residents and Total Resident COVID-19 Deaths per 1000 Residents. These metrics provide a misleading picture, as facilities report cumulative counts of cases and deaths over different time periods but use a point-in-time measure as proxy for number of residents (number of occupied beds in a week), resulting in inflated statistics. We propose an alternative statistic to better illustrate the burden of COVID-19 cases and deaths across nursing facilities. Design: Retrospective cohort study. Setting and Participants: Using the CMS Nursing Home Compare and COVID-19 Nursing Home Datasets, we examined facilities with star ratings and COVID-19 data passing quality assurance checks for each reporting period from May 31 to August 16, 2020 (n = 11,115). Methods: We derived an alternative measure of the number of COVID-19 cases per 1000 residents using the net change in weekly census. For each measure, we compared predicted number of cases/deaths by overall star rating using negative binomial regression with constant dispersion, controlling for county-level cases per capita and nursing home characteristics. Results: The average number of cases per 1000 estimated residents using our method is lower compared with the metric using occupied beds as proxy for number of residents (44.8 compared with 66.6). We find similar results when examining number of COVID-19 deaths per 1000 residents. Conclusions and Implications: Future research should estimate the number of residents served in nursing facilities when comparing COVID-19 cases/deaths in nursing facilities. Identifying appropriate metrics for facility-level comparisons is critical to protecting nursing home residents as the pandemic continues.

Original languageEnglish (US)
Pages (from-to)955-959.e3
JournalJournal of the American Medical Directors Association
Volume22
Issue number5
DOIs
StatePublished - May 2021
Externally publishedYes

Keywords

  • Nursing homes
  • coronavirus
  • measurement

ASJC Scopus subject areas

  • General Nursing
  • Health Policy
  • Geriatrics and Gerontology

Fingerprint

Dive into the research topics of 'The Metrics Matter: Improving Comparisons of COVID-19 Outbreaks in Nursing Homes'. Together they form a unique fingerprint.

Cite this