Skip to main content
Open Access Publications from the University of California


UC San Francisco Electronic Theses and Dissertations bannerUCSF

Nursing Home Residents Stuck in the Revolving Door of the Emergency Room: Does Cognitive Impairment Make a Difference?


Significance– Nursing home (NH) residents use the ED at higher rates than community-dwelling older adults and nearly 2/3 of NH residents cared for in the ED have cognitive impairment (CI). Studies indicate hospitalization of nursing home residents is frequent and often preventable, however little is known about ED use by residents with CI.

Methods– Utilizing a 5% random sample of Medicare beneficiaries residing in NHs in the U.S. (n=132,753, of whom 76,961 had CI), this secondary analysis of 2006 MDS and Medicare claims data examined the prevalence of and factors associated with ED use with and without hospitalization, and preventable ED use, and how such utilization varied by the presence and severity of CI. Descriptive statistics were used to describe the study sample. Multinomial logistic regression was used to estimate the dichotomous dependent variables; Poisson regression was used to estimate the continuous dependent variables, controlling for sociodemographic/health characteristics.

Results– Sixty two percent (n=82,335) of the sample had an ED visit with or without hospital admission. Residents with CI had fewer total number of ED visits/year compared to those without CI. Controlling for acuity and other factors, the odds and rates of ED use with and without hospitalization varied by both presence and severity of CI. Odds of any ED visit were highest for residents with very mild CI and risk declined as CI worsened. There was a similar trend for ED visits resulting in hospitalization, with 13% higher odds for residents with very mild CI compared to residents without CI; risk also declined as CI worsened. Risk of ED use without hospitalization was greatest for those with mild CI and moderate-severe CI, with 7-12% higher odds of an ED visit without hospitalization/year than those without CI. Almost 24% (n=31,566) of the sample had a potentially preventable ED visit. Residents with very mild-mild CI were at highest risk, with 10-15% higher odds of a preventable ED visit compared to those without CI.

Implications– Identifying and modifying factors responsible for ED use during an illness episode may result in more appropriate use of the ED, reduced costs and better health outcomes. Using preventable ED use as a NH quality indicator may be warranted.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View