What are the effects of teaching Evidence-Based Health Care

Evidence-based healthcare (EBHC) knowledge and skills are recognised as core competencies of healthcare professionals worldwide, and teaching EBHC has been widely recommended as an integral part of their training. The objective of this overview of systematic reviews (SR) was to update evidence and assess the effects of various approaches for teaching evidence-based health care (EBHC) at undergraduate (UG) and postgraduate (PG) medical education (ME) level on changes in knowledge, skills, attitudes and behaviour. This is an update of an overview that was published in 2014. The process followed standard procedures specified for the previous version of the overview, with a modified search. Searches were conducted in Epistemonikos for SRs published from 1 January 2013 to 27 October 2020 with no language restrictions. We checked additional sources for ongoing and unpublished SRs. Eligibility criteria included: SRs which evaluated educational interventions for teaching EBHC compared to no intervention or a different strategy were eligible. Two reviewers independently selected SRs, extracted data and evaluated quality using standardised instrument (AMSTAR2). The effects of strategies to teach EBHC were synthesized using a narrative approach. Previously published version of this overview included 16 SR, while the updated search identified six additional SRs. We therefore included a total of 22 SRs (with a total of 141 primary studies) in this updated overview. The SRs evaluated different educational interventions of varying duration, frequency, and format to teach various components of EBHC at different levels of ME (UG, PG, mixed). Most SRs assessed a range of EBHC related outcomes using a variety of assessment tools. Two SRs included randomised controlled trials (RCTs) only, while 20 reviews included RCTs and various types of non-RCTs. Diversity of study designs and teaching activities as well as aggregated findings at the SR level prevented comparisons of the effects of different techniques. In general, knowledge was improved across all ME levels for interventions compared to no intervention or pre-test scores. Skills improved in UGs, but less so in PGs and were less consistent in mixed populations. There were positive changes in behaviour among UGs and PGs, but not in mixed populations, with no consistent improvement in attitudes in any of the studied groups. One SR showed improved patient outcomes (based on non-randomised studies). Main limitations included: