Derivation and validation of a feasible emergency department specific frailty index to predict adverse outcomes

TitleDerivation and validation of a feasible emergency department specific frailty index to predict adverse outcomes
Publication TypeConference Paper
Year of Publication2017
AuthorsBrousseau A., Dent E., Hubbard R., Melady D., Emond M., Mercier E., Costa A.
Conference NameJournal of the American Geriatrics Society
IssueSupplement 1
PublisherBlackwell Publishing Inc.
Conference LocationNetherlands
Keywords*adverse outcome, *emergency ward, *Frailty, *validation process, Aged, Death, Female, Geriatric Assessment, Hospitalization, Human, human tissue, logistic regression analysis, long term care, major clinical study, Male, model, Odds Ratio, prospective study

Background Frailty is an overarching concept in geriatric medicine. However its utility in the emergency department (ED) was not well understood. Objectives were to derive and validate an ED specific frailty index (FI-ED), using a cumulative deficits model; and to evaluate its ability to predict adverse outcomes. Method This was a large multinational prospective cohort study using data from: The Management of Older Persons in Emergency Departments (MOPED) and the interRAI study. The FI-ED was derived from the Canadian sample and validated in the multinational sample. Inclusion criteria were all patients >= 75 years old presenting to an ED. The FI-ED used 24 variables identified in the interRAI ED-Contact Assessment tool, a brief focussed geriatric assessment. Its ability to predict adverse outcomes were analysed by logistic regression with odds ratio (OR). Results There were 3903 participants: 2153 in the derivation sample and 1750 in the validation sample. In the derivation sample, increasing FI-ED was significantly associated with admission (OR 1.43 [95% CI 1.34-1.52]), death in hospital (OR 1.55 [1.38-1.73]), prolonged hospital stay (OR 1.37 [1.22-1.54]), needs for Comprehensive Geriatric Assessment (OR 1.51 [1.41-1.60]) and discharge to long-term care (OR 1.30 [1.16-1.47]). In the validation sample, results were similar except for long-term care disposition (OR 0.84 [0.75 0.85]). Conclusion The FI-ED conformed to characteristics previously reported in other geriatric populations. It was accurately derived and validated from a brief geriatric assessment feasible in the ED and can be used in the ED to predict adverse outcomes.

Reseach Notes