Background: Older people in the Emergency Department (ED) are clinically heterogenous and some presentations may be better suited to alternative out-of-hospital pathways. A new interdisciplinary comprehensive geriatric assessment (CGA) team (Home FIRsT) was embedded in our acute hospital's ED in 2017.

Aim: To evaluate if routinely collected CGA metrics were associated with ED disposition outcomes.

Design: Retrospective observational study.

Methods: We included all first patients seen by Home FIRsT between 7(th) May and 19(th) October 2018. Collected measures were sociodemographic, baseline frailty (Clinical Frailty Scale), major diagnostic categories, illness acuity (Manchester Triage Score) and cognitive impairment/delirium (4AT). Multivariate binary logistic regression models were computed to predict ED disposition

Outcomes: hospital admission; discharge to GP and/or community services; discharge to specialist geriatric outpatients; discharge to the Geriatric Day Hospital.

Results: In the study period, there were 1,045 Home FIRsT assessments (mean age 80.1 years). For hospital admission, strong independent predictors were acute illness severity (OR 2.01, 95% CI 1.50-2.70, P<0.001) and 4AT (OR 1.26, 95% CI 1.13 - 1.42, P<0.001). Discharge to specialist outpatients (e.g. falls/bone health) was predicted by musculoskeletal/injuries/trauma presentations (OR 6.45, 95% CI 1.52 - 27.32, P=0.011). Discharge to the Geriatric Day Hospital was only predicted by frailty (OR 1.52, 95% CI 1.17 - 1.97, P=0.002). Age and sex were not predictive in any of the models.

Conclusions: Routinely collected CGA metrics are useful to predict ED disposition. The ability of baseline frailty to predict ED outcomes needs to be considered together with acute illness severity and delirium.