Assessing Geriatric Competencies in Residents: Validating the 5Ms Dimensions

Background Despite undergraduate training in geriatric care, gaps persist throughout residency, highlighting limitations of current assessment methods in evaluating medical expertise across geriatric dimensions. Objective We developed a case-based assessment using the geriatric 5Ms framework (Mind, Mobility, Medications, Multicomplexity, Matters Most), aligned with undergraduate objectives and North American internal medicine milestones. We present feasibility data and preliminary validity evidence of using the geriatric 5Ms framework to evaluate residents' geriatric medical expertise. Methods During a 2023 mandatory academic session at a single site, 68 first- to third-year internal medicine residents were randomly assigned to complete assessment and management plans for 3 of 6 geriatric cases within 1 hour. Two blinded educators rated performance on 5Ms dimensions and non-geriatric medical expertise using a 3-level rating scale (0 to 2). We collected feasibility data (logistical integration, participation rates, time to design cases, rate responses) and validity evidence, based on Messick's framework, in part through a post-assessment questionnaire. Results Sixty-five residents completed 3 cases each, and 3 residents completed 2 cases each, resulting in 201 total cases, each integrating all 5Ms dimensions. Scores across the 5Ms dimensions ranged from 0.8 to 1.3, indicating partial assessment and management. All 5Ms dimensions (mean=1.1, SD=0.3) scored significantly lower than non-geriatric medical expertise (mean=1.5; SD=0.3; t(64)=9.58; P<.001). Interrater reliability was moderate to strong (ICC=0.67-0.85, P<.001). Most residents rated the cases (59 of 67, 88%; mean=4.4; SD=0.7) and the assessment (56 of 67, 84%; mean=4.1; SD=0.7) as representative of clinical practice. Conclusions A case-based assessment using the geriatric 5Ms framework demonstrated feasibility and preliminary validity for evaluating residents' geriatric medical expertise.