Emergency Department Use by Nursing Home Residents: Effect of Severity of Cognitive Impairment

• Published in: The Gerontologist 2012-06, Vol.52 (3), p.383-393
• Main Authors: Stephens, Caroline E, Newcomer, Robert, Blegen, Mary, Miller, Bruce, Harrington, Charlene
• Format: Article
• Language: English
• Subjects: Accountability; Aged; Aged, 80 and over; Aging; Ambulatory care; Centers for Medicare and Medicaid Services (U.S.) - statistics & numerical data; Cognition; Cognition Disorders - diagnosis; Cognition Disorders - epidemiology; Cognitive ability; Cross-Sectional Studies; Dementia; Dementia disorders; Emergency medical care; Emergency Medical Services; Emergency Service, Hospital - statistics & numerical data; Emergency Service, Hospital - utilization; Family; Female; Health Facilities; Homes for the Aged - statistics & numerical data; Hospitalization - statistics & numerical data; Hospitals; Humans; Identification; Incidence; Logistic Models; Male; Marriage and Family Education; Nursing; Nursing Homes; Nursing Homes - statistics & numerical data; Older Adults; Outpatient Clinics, Hospital - statistics & numerical data; Outpatient Clinics, Hospital - utilization; Patients; Prevalence; Probability; Quality of Health Care; Regression (Statistics); Risk; Risk Factors; Senility; Severity (of Disability); Severity of Illness Index; Socioeconomic Factors; United States; Utilization Review - statistics & numerical data
• ISSN: 0016-9013; 1758-5341
• DOI: 10.1093/geront/gnr109

Purpose: To examine the 1-year prevalence and risk of emergency department (ED) use and ambulatory care-sensitive (ACS) ED use by nursing home (NH) residents with different levels of severity of cognitive impairment (CI). Design and Methods: We used multinomial logistic regression to estimate the effect of CI severity on the odds of any ED visit and any ACS ED visit in a 2006 national random sample of NH residents, controlling for predisposing, enabling, and need characteristics. Results: Of 132,753 NH residents, 62% had at least one ED visit and approximately 24% had at least one ACS ED visit in 2006. The probability of any ED visit or any ACS ED visit varied with the severity of resident CI. Residents with mild CI had up to 15% higher odds of any ED or any ACS ED visit and those with more moderate CI had 9% higher odds of an ACS ED visit compared with those without CI. The probability of any ED visit was negatively associated with advanced dementia (adjusted odds ratio = 0.60; 95% CI = 0.55-0.65). Implications: Earlier identification of persons with mild CI may facilitate patient, family, and staff education, as well as advanced care planning to reduce ACS ED visits. Both ACS ED use and hospitalizations, adjusted for case mix, should be used as quality indicators to help ensure greater accountability for high-quality NH care and more appropriate utilization of ED resources.