Self-Directed Learning in Health Professions Education: A Systematic Review and Meta-Analysis

This systematic review and meta-analysis evaluates the effectiveness of self-directed learning (SDL) in health professions education (HPE), examining its impact on knowledge, clinical performance, and behavioral outcomes. It investigates whether core SDL components influence educational outcomes, updating and extending the foundational work of Murad et al. (2010).

We searched CINAHL, Embase, OVID Medline, PsycINFO, and Web of Science (2009-2023) for comparative studies evaluating SDL interventions in HPE. From 6,786 screened articles, 125 studies met inclusion criteria, with 48 eligible for meta-analysis. We conducted a three-level random-effects meta-analysis and moderator analyses on profession, outcome type, SDL modality, and facilitator role. Five independent reviewers conducted screening and extraction, resolving discrepancies via consensus.

The meta-analysis incorporated 74 effect sizes from 48 studies, revealing a small-to-moderate overall effect (Cohen's d = 0.34, 95% CI 0.04, 0.64) with significant heterogeneity (I² = 87%). SDL as intervention showed larger effects (d = 0.54 vs. d = -0.27, p = 0.004). Most studies involved Kirkpatrick Level 2 outcomes (knowledge/skills, 78%), with some Level 3 outcomes (skills/behaviors, 22%) and no Level 4 outcomes (patient/system) reported. Most teachers were absent or acted as facilitators, while learners were less likely to be involved in choosing resources (21%) or in assessments (25%).

This updated meta-analysis reaffirms that SDL reliably enhances knowledge acquisition but suggests that it may yield only modest gains in clinical skills and behaviors. The wide variability in how SDL is defined and reported underscores the need for a consensus definition of SDL.