Construction and Validation of Risk-adjusted Rates of Emergency Department Visits for Long-stay Nursing Home Residents

Abstract

Background: Long-stay nursing home (NH) residents are at high risk of having emergency department (ED) visits, but current knowledge regarding risk-adjusted ED rates is limited.

Objectives: To construct and validate 3 quarterly risk-adjusted rates of long-stay residents' ED use: any ED visit, ED visits without hospitalization or observation stay (outpatient ED), and potentially avoidable ED visits (PAED).

Research design: The authors calculated quarterly NH risk-adjusted ED rates from 2011 Q2 to 2013 Q3 national Medicare claims and Minimum Data Set data. Using random-effect linear regressions, the authors validated these rates against Nursing Home Compare overall 5-star quality ratings and examined their associations with hospitalization rates to provide a quality context.

Subjects: Resident-quarter observations (7.3 million) from 15,235 unique NHs.

Results: Risk-adjusted rates of any ED, outpatient ED, and PAED averaged 9.7%, 3.4%, and 3.2%, respectively. Compared with NHs with 1 or 2 stars overall rating, NHs with ≥3 stars were significantly associated with lower rates of any ED visit, outpatient ED, and PAED (β, -0.23%, -0.16%, and -0.11%, respectively; all P<0.01). Pearson Correlation coefficients between hospitalization rates and rates of any ED visit, outpatient ED, and PAED were 0.74, 0.31, and 0.46, respectively.

Conclusions: The moderately negative associations of 5-star ratings with ED rates provide supportive evidence to their validity. Outpatient ED and PAED were moderately correlated to hospitalizations suggesting they provided more information about quality than any ED.