Background: tools are required to identify high-risk older people in acute emergency settings so that appropriate services can be directed towards them.

Objective: to evaluate whether the Identification of Seniors At Risk (ISAR) predicts the clinical outcomes and health and social services costs of older people discharged from acute medical units.

Design: an observational cohort study using receiver–operator curve analysis to compare baseline ISAR to an adverse clinical outcome at 90 days (where an adverse outcome was any of death, institutionalisation, hospital readmission, increased dependency in activities of daily living (decrease of 2 or more points on the Barthel ADL Index), reduced mental well-being (increase of 2 or more points on the 12-point General Health Questionnaire) or reduced quality of life (reduction in the EuroQol-5D) and high health and social services costs over 90 days estimated from routine electronic service records.

Setting: two acute medical units in the East Midlands, UK.

Participants: a total of 667 patients aged ≥70 discharged from acute medical units.

Results: an adverse outcome at 90 days was observed in 76% of participants. The ISAR was poor at predicting adverse outcomes (AUC: 0.60, 95% CI: 0.54–0.65) and fair for health and social care costs (AUC: 0.70, 95% CI: 0.59–0.81).

Conclusions: adverse outcomes are common in older people discharged from acute medical units in the UK; the poor predictive ability of the ISAR in older people discharged from acute medical units makes it unsuitable as a sole tool in clinical decision-making.