A randomized study of iterative hypothesis testing in undergraduate psychiatric education

Cross-sectional study of the quality of randomized control trials on problem-based learning in medical education

Problem-based learning (PBL) is increasingly being used in medical education globally, but its effectiveness in teaching remains controversial. A randomized controlled trial (RCT) is the method of choice for evaluating its effectiveness. The quality of an RCT has a significant effect on this evaluation, but to date we have not seen an assessment of the quality of RCTs for PBL. Two researchers searched MEDLINE and EMBASE for RCTs addressing PBL in medical education. The overall quality of each report was measured on a 28-point overall quality score (OQS) based on the 2010 revised Comprehensive Standards for Reporting and Testing (CONSORT) Statement. Furthermore, to study the key factors affecting OQS more effectively, a linear regression model of those factors was established using SPSS. After literature screening, 30 RCTs were eventually included and analyzed. The median OQS was 15 (range, 7-20), which meant that half of the items in the revised 2010 CONSORT statement were poorly reported in at least 40% of the RCTs analyzed. The regression model showed that the year of publication of RCTs and the impact factors of the journals in which they were published were the main factors affecting OQS. The overall quality of reporting of RCTs on PBL teaching in medical education was not satisfactory. Some RCTs were subjectively selective in reporting certain items, leading to heterogeneity in quality. It is expected that statisticians will develop new standards more suitable for evaluating RCTs related to teaching research and that editors and peer reviewers will be required to review the relevant RCTs more strictly.

The effectiveness of using virtual patient educational tools to improve medical students' clinical reasoning skills: a systematic review

BACKGROUND

Use of virtual patient educational tools could fill the current gap in the teaching of clinical reasoning skills. However, there is a limited understanding of their effectiveness. The aim of this study was to synthesise the evidence to understand the effectiveness of virtual patient tools aimed at improving undergraduate medical students' clinical reasoning skills.

METHODS

We searched MEDLINE, EMBASE, CINAHL, ERIC, Scopus, Web of Science and PsycINFO from 1990 to January 2022, to identify all experimental articles testing the effectiveness of virtual patient educational tools on medical students' clinical reasoning skills. Quality of the articles was assessed using an adapted form of the MERSQI and the Newcastle-Ottawa Scale. A narrative synthesis summarised intervention features, how virtual patient tools were evaluated and reported effectiveness.

RESULTS

The search revealed 8,186 articles, with 19 articles meeting the inclusion criteria. Average study quality was moderate (M = 6.5, SD = 2.7), with nearly half not reporting any measurement of validity or reliability for their clinical reasoning outcome measure (8/19, 42%). Eleven articles found a positive effect of virtual patient tools on reasoning (11/19, 58%). Four reported no significant effect and four reported mixed effects (4/19, 21%). Several domains of clinical reasoning were evaluated. Data gathering, ideas about diagnosis and patient management were more often found to improve after virtual patient use (34/47 analyses, 72%) than application of knowledge, flexibility in thinking and problem-solving (3/7 analyses, 43%).

CONCLUSIONS

Using virtual patient tools could effectively complement current teaching especially if opportunities for face-to-face teaching or other methods are limited, as there was some evidence that virtual patient educational tools can improve undergraduate medical students' clinical reasoning skills. Evaluations that measured more case specific clinical reasoning domains, such as data gathering, showed more consistent improvement than general measures like problem-solving. Case specific measures might be more sensitive to change given the context dependent nature of clinical reasoning. Consistent use of validated clinical reasoning measures is needed to enable a meta-analysis to estimate effectiveness.