Identifying patients at risk for medication mismanagement: using cognitive screens to predict a patient's accuracy in filling a pillbox.

Objective: To study the relationship between four cognitive screens (Mini-Mental State Examination [MMSE], Mini-Cog, Medication Transfer Screen [MTS], and Medi-Cog [Mini-Cog + MTS]) and pillbox concordance.

Design: Prospective cross-sectional pilot study.

Setting: Primary care federal health care system.

Participants: English literate inpatients not previously diagnosed or treated for dementia and without physical handicap preventing use of a pillbox.

Main Outcome Measure: Correlation between cognitive screens and prospective pill-count scores (PPCS).

Interventions: Mini-Cog, MTS, and MMSE screening was followed by a 28-compartment pillbox skills assessment. A passing PPCS was defined as correctly loading 80% of the medications in the pillbox. A PPCS of <80% identified patients for pillbox-organization education or supportive intervention. Variables associated with total and passing PPCS were analyzed by multivariate linear and logistic regression, respectively.

Results: Fifty-three patients discharged on >1 medication completed all screenings. Other than cognitive screening, only age was associated with total and passing PPCS. After adjustment for age, Medi-Cog had the highest correlation with total PPCS [r2=0.53; P<0.001), whereas Mini-Cog was the single cognitive assessment that remained significantly associated with a passing PPCS (r2=0.23; P=0.023). Age-adjusted models, including MMSE, had relatively poor association with total PPCS (r2=0.23; P=0.046) and no association with passing PPCS (r2=0.15; P=0.46). The Medi-Cog exhibited modest highest overall sensitivity (72%) and specificity (61%) to detect a passing PPCS.

Conclusion: MMSE is a relatively poor measure of the ability of patients to fill a pillbox. The Medi-Cog and Mini-Cog may have value for assessing pillbox concordance for patients who load their own pillboxes.