Portfolio programs in medical education: what kind of assessments should be used to objectively assess learning success?

Background:
Literature suggests portfolio is a powerful tool to promote learning success of medical students over time. However, student learning processes are mostly evaluated by questionnaires or interviews. The interpretation of such rather subjective assessment tools may lead to biased conclusions. Therefore, other, more objective assessment instruments seem to be needed.
Summary of work:
To answer the question “Which assessment instruments were implemented in studies using portfolio programs to objectively evaluate learning process over time in medical students?”, a systematic review was conducted in the Cinahl, Cochrane, Embase and PubMed databases. The Cochrane risk of bias tool was used to assess procedural validity.
Summary of Results:
Out of 254 eligible articles, four studies (medicine=3; midwifery=1) could be included. All portfolio studies showed a high risk of bias. Assessment tools used, were written tests (standard multiple-choice questions (MCQs)) and a clinical skills test (Objective Structured Clinical Examination (OSCE)).
Discussion:
MCQs and OSCEs are considered valid instruments to assess the learning process of medicine and midwifery students, in line with the learning criteria: “knows”, “knows how” and “shows how” (Miller’s learning pyramid). However, are these tools applicable and should they be mandatory for medical and health profession education research?
Conclusion:
Valid assessment tools in portfolio are, but rarely, used. To increase the objectivity of portfolio it is suggested that MCQs and OSCEs are included in addition to questionnaires or interviews. Future high-quality studies (RCTs) are recommended to test the external validity in other health care education settings (e.g. physiotherapy).
Take home message:
The use of valid assessment tools in portfolio studies is scarce. It is suggested that future portfolio studies should implement MCQs and OSCEs to assess learning progress in health care students. To reduce bias, future studies should be registered a priori and implemented using a randomized controlled trial design.

Jaar van publicatie:2018