Cognition, even in the normal range, predicts disability: cross-sectional and prospective analyses of a population-based sample.

OBJECTIVE: To determine if the modified mini-mental state examination (3MS) predicts functional status and if any effect on function is observed within the normal range of cognition.

DESIGN: Cohort study.

SETTING: Community-dwelling older adults in the Canadian province of Manitoba sampled in 1991 and followed in 1996.

PARTICIPANTS: Baseline sample of 1751 adults aged 65+ from a representative registry. Five years later, 1028 participants remained in the community and had no missing data.

MEASUREMENTS: The 3MS, age, gender, education, living arrangements, self-rated health, and depressive symptoms were self-reported. Functional status was assessed using the Older Americans Resource Survey, which was dichotomized into no/mild disability versus moderate/severe disability.

RESULTS: Baseline 3MS score predicted baseline functional status. This effect was a gradient across the entire 3MS score, extending into the normal range with no apparent threshold. In logistic regression models, the unadjusted odds ratio (OR, 95% confidence interval) for the association of 3MS score with disability was 0.94 (0.93, 0.95); the adjusted OR was 0.96 (0.95, 0.98) in models including age, gender, education, and other covariates. Baseline 3MS score also predicted functional status 5 years later: The unadjusted OR for disability was 0.94 (0.92, 0.95); the adjusted OR was 0.97 (0.95, 0.99). Again, the risk of functional impairment at time 2 was a gradient effect, extending into the normal range of baseline 3MS score.

CONCLUSIONS: The 3MS predicts functional decline, and this effect is a gradient effect. These results support the hypothesis that cognition is a continuum in risk.