RESUMO
OBJECTIVE: To compare patient cognition measured by Medi-Cog, a tool to assess cognitive literacy and pillbox skills, with pillbox concordance using two scoring methods, Pillbox Fill (PBF) and Prospective Pill Count (PPC). DESIGN: Prospective, descriptive, cross-sectional study. SETTING: Primary care. PARTICIPANTS: Multiethnic participants with type 2 diabetes with sufficient vision and dexterity to load a pillbox. INTERVENTION: Medi-Cog scores were correlated with ability to fill a pillbox based on both the PPC and the PBF scoring methods. Variables were analyzed by multivariate linear and logistic regression. MAIN OUTCOME MEASURES: To determine whether there is a difference between PBF and PPC scoring methods relative to Medi-Cog prediction of pillbox concordance. RESULTS: Sixty-four participants loaded an average of 5.2 medications. Mean Medi-Cog score for five patients who failed PBF but passed PPC were lower than the entire cohort (5.6 compared with 6.2). Correlation between PBF and PPC methods was 0.978; P = 0.01. Regression values for Medi-Cog's ability to predict PBF and PPC scores were r = 0.668 and r2 = 0.446, and r = 0.660 and r2 = 0.436; P < 0.01 for all. CONCLUSION: Compared with PPC, PBF proved to be a more conservative scoring method and captured an additional five patients who scored less-well on the Medi-Cog. Future studies are needed to explore the value of using pillbox assessments as well as cognitive screening prior to recommending pillbox use.
