The Clinical Utility and Ecological Validity of the Medication Management Ability Assessment in Older Adults with and without Dementia.

Abstract

Older adults are susceptible to medication nonadherence, which may signify functional decline. Thus, performance-based proxies of medication-taking behavior may help diagnose dementia. We assessed the Medication Management Ability Assessment's (MMAA) clinical utility and ecological validity. This was a retrospective chart review of 180 outpatients (age=72±8years) who completed the MMAA during clinical evaluations. Forty-seven were cognitively normal (CN), 103 had mild cognitive impairment (MCI), and 30 had dementia. Most (136) were independent in medication management, whereas 28 were assisted and 16 were dependent. Kruskal-Wallis tests assessed whether MMAA scores differed by diagnosis and independence. Receiver operating characteristic (ROC) analyses identified diagnostic cut-offs. Classification accuracy estimates were derived. MMAA performance differed across diagnosis as expected (p's<.001). Those who were independent in medication management outperformed assisted and dependent counterparts (p's<.001). Assisted and dependent cases were no different. At a cut-off=23, the MMAA was good-to-strong in distinguishing dementia from CN cases (Sn=0.96, Sp=0.83), dementia from MCI (Sn=0.70, Sp=0.83), and dementia from functionally unimpaired cases (Sn=0.78, Sp=0.83). At a cut-off=27, it had good sensitivity but weaker specificity when distinguishing both MCI and all cognitively impaired patients (MCI and dementia) from CN cases (Sn=0.81, Sp=0.66 and Sn=0.81, Sp=0.72, respectively). The MMAA has ecological validity and clinical utility in identifying dementia. Its inclusion in neuropsychological practice may be especially useful when medication mismanagement is suspected.