Portfolios in medical education: why do they meet with mixed success? A systematic review

Abstract

Context The move towards competence‐based medical education has created a need for instruments that support and assess competence development. Portfolios seem suitable but mixed reports of their success are emerging. Methods To examine the effectiveness of portfolios, we searched PubMed and EMBASE using the keyword ‘portfol*’, PsychInfo and ERIC using the keywords ‘portfol*’ and ‘medical education’ and references of retrieved papers for empirical studies on portfolios in all phases of medical education. Thirty of 1939 retrieved papers met the inclusion criteria and were analysed. Data were collated against the research question, number of subjects, design, setting, findings and limitations, purpose and content, mentoring and assessment. We analysed impact using a modified version of Kirkpatrick’s hierarchy. Results Because differences across studies precluded statistical meta‐analysis, the data were analysed by context, goals and procedure. Positive effects were strongest in undergraduate education. Important factors for success were: clearly communicated goals and procedures; integration with curriculum and assessment; flexible structure; support through mentoring, and measures to heighten feasibility and reduce required time. Moderately good inter‐rater reliability was reported and global criteria and discussions among raters were beneficial. Formative and summative assessment could be combined. Without assessment, portfolios were vulnerable to competition from other summative assessment instruments. Conclusions For portfolios to be effective in supporting and assessing competence development, robust integration into the curriculum and tutor support are essential. Further studies should focus on the effectiveness and user‐friendliness of portfolios, the merits of holistic assessment procedures, and the competences of an effective portfolio mentor.