Development and evaluation of the GEMS (gait, eyesight, mental state, sedation) tool as an aid to predict outcome after hospitalization


Dr Becky Jupp MRCP, Department of Medicine for the Elderly, Royal Bournemouth Hospital, Castle Lane East, Bournemouth, Dorset BH7 7DW, UK. Email:


Aim:  Discharge planning can be a lengthy process. Prediction of a patient's rehabilitation potential and likely discharge destination, early on in their admission, could be a useful guide for medics, therapists, patients and their carers. Reliable prediction could be used to improve efficiency of discharge planning. The aim of the present study was to identify factors linked to discharge to a residential or nursing home placement and to develop a tool to guide rehabilitation requirements.

Methods:  This was a three-phase prospective observational study with blinded end-point evaluation in two non-acute rehabilitation hospitals. The study recruited 1174 patients admitted for rehabilitation, over 65 years of age. Phase 1 evaluated 200 patients to identify factors predisposing to institutional discharge. The GEMS (gait, eyesight, mental state, sedation) tool was formulated and validated on the ward in which it was developed (phase 2a) and on two other wards at the same rehabilitation hospital (phase 2b). In phase 3, the tool was evaluated remotely.

Results:  Patients discharged to a nursing or residential home placement were significantly more likely to have abnormal vision (P = 0.01, 95% confidence interval [CI] = 0.18–0.81), impaired cognitive function (P = 0.012, 95% CI = 0.19–0.81), gait abnormalities (P = 0.01, 95% CI = 0.18–0.79), and more likely to be taking tranquillizers (sedation) (P = 0.0001, 95% CI = 0.01–0.45). Over the three phases, the GEMS tool had a sensitivity of 61.4–88.1% and a specificity of 28.9–61.0%.

Conclusion:  A GEMS score of 2 or more is significantly associated with discharge to a residential or nursing home placement. This could be used to aid discharge planning and direct rehabilitation service provision. Geriatr Gerontol Int 2011; 11: 8–15.