Cognitive function tests are often used to predict rehabilitation outcomes. We aimed to determine how predictive the MMSE, CLOX and a short Frontal Lobe Assessment (sFLA) were in determining likely improvement in activities of daily living and discharge home.

In a prospective observational study, we evaluated a cohort of 241 patients [97 Male mean (SD) median age: 84.4 (7.27) 85 years]. Functional ability was assessed using the Barthel Activities of Daily Living (BADL) scale. Outcomes were an improvement in one domain on the BADL and discharge home.

Whatever the tool, abnormal cognition was an independent factor for lack of improvement in BADL [MMSE – P=0.000 (B=1.11; 95%CI: 1.05–1.17); CLOX – P=0.007 (B=1.13; 95% CI: 1.06–1.22) and sFLA – P=0.0001 (B=1.19; 95% CI: 1.09–1.31)] and for failure to discharge home [MMSE – P=0.0001 (B=1.13; 95%CI: 1.06–1.19); CLOX – P=0.007 (B=1.12; 95%CI: 1.03–1.21) and sFLA – P=0.002 (B=1.18; 95%CI: 1.06–1.31)]. The MMSE correlated positively with the CLOX and sFLA (r=0.54: P=0.000 and r=0.7: P=0.000 respectively) and a weaker positive correlation between the CLOX and sFLA (r=0.43: P=0.000). The Receiver Operative Characteristic (ROC) Curves for all tests mirrored each other across the range of scores with similar and modest areas under the curves for the prediction of improvement in BADL and discharge home (BADL: range 0.65–0.68 and discharge home: range 0.70–0.77).

Although the MMSE, CLOX and sFLA assess different aspects of cognition, there seems little benefit of one test over another. Over reliance on these tests alone, to determine the likely outcome of rehabilitation is unjustified and patients should not be denied rehabilitation just because they may be abnormal.