JMIR Publications

ABSTRACT

Background:

Assessing medication literacy in children is essential for evaluating school-based health education initiatives, yet traditional measurement tools often fail to engage young respondents, potentially compromising data quality. Gamified assessments offer a promising alternative, but their psychometric properties, comparability to conventional formats, and responsiveness to intervention effects require systematic evaluation.

Objective:

This study developed and evaluated a gamified assessment tool for measuring medication literacy across four domains (Knowledge, Attitude, Perceived Behavioral Control, and Behavioral Intention) in elementary school children through three sequential phases: psychometric validation, methodological comparison with a traditional paper-based questionnaire, and application in an intervention context.

Methods:

In Phase 1, content validity was first established through a two-round expert review. The refined tool was then administered to 81 students to assess construct validity (Rasch analysis) and test-retest reliability (one-month interval). Phase 2 employed a randomized crossover design with 85 students to assess equivalence (ICC, Bland-Altman, decision consistency) and superiority (data quality, enjoyment, preference) compared to paper-based administration. We adapted the CHERRIES framework to report on user recruitment, engagement metrics, and data completeness, as these principles remain applicable to digital survey modalities. Phase 3 involved 85 students in a pre-post intervention design, with responsiveness evaluated using generalized linear mixed models.

Results:

The tool demonstrated strong content validity (S-CVI/Ave = 1.00), adequate construct validity (unidimensionality supported, item fit within 0.5–1.5, point-measure correlations ≥0.67), and good test-retest reliability (ICC: 0.76–0.92). Equivalence with paper-based administration was confirmed across all domains (ICC: 0.94–0.99; mean bias near zero; κ: 0.87–0.97). While data quality was comparable (validity rates: 100% vs 90.7%, P = .13), the gamified tool yielded substantially higher enjoyment and preference ratings (Cohen's d = 2.00 and 1.62, respectively). The tool detected significant intervention effects across all four domains (β: 1.74–2.86; IRR = 1.52, all P < .001), with moderate to large effect sizes and consistent improvements across classrooms despite baseline differences.

Conclusions:

The gamified assessment is a psychometrically sound, methodologically robust alternative to traditional questionnaires. It achieves measurement equivalence while enhancing child engagement—a dual advantage that supports both data quality and feasibility in pediatric research. By demonstrating consistent intervention effects across diverse classrooms, the tool further holds potential to advance health equity in school-based medication literacy education.