Interventions to improve diagnostic decision making: A systematic review and meta-analysis on reflective strategies.

ATTITUDE (Psychology); COGNITION; CONFIDENCE intervals; DECISION making; DIAGNOSIS; ERIC (Information retrieval system); INFORMATION storage & retrieval systems; MEDICAL databases; MEDICAL information storage & retrieval systems; PSYCHOLOGY information storage & retrieval systems; MEDICAL personnel; MEDLINE; META-analysis; ONLINE information services; REFLECTION (Philosophy); ADULT education workshops; SYSTEMATIC reviews; DECISION making in clinical medicine; EFFECT sizes (Statistics)

Aims: To identify, appraise and describe studies of cognitive interventions to improve diagnostic decision making (DDM) amongst medical professionals, assess their effectiveness and identify methodological limitations in existing studies. Methods: We systematically searched for studies (publication date 2000–2016) in multiple databases including Cochrane Controlled Trials, EMBASE, ERIC, Medline, PubMed and PsycINFO, and used additional strategies such as hand searching and snowballing. Included studies evaluated cognitive interventions to enhance DDM amongst medical professionals, using defined outcomes such as diagnostic accuracy. A meta-analysis assessed the impact of "reflection". Results: Forty-four studies out of 10,114 screened citations, involving 4380 medical professionals, were included. Studies evaluated reasoning workshops/curricula, de-biasing workshops, checklists, reflection, feedback, and instructions to induce analytical thinking. Guided reflection was demonstrated to improve DDM [effect size 0.38(95%CI 0.23–0.52), p < 0.001]. Immediate feedback and modeling reflection using contrasting examples also appeared to improve diagnostic accuracy, however underlying methodological issues prevented a quantitative assessment of any strategies other than reflection. Conclusions: Educational interventions incorporating practising deliberate reflection on a formulated diagnosis, modeled reflection on contrasting examples and immediate feedback are promising strategies for improving DDM. The effectiveness of other strategies is unknown, with more methodological refinements required in future research. [ABSTRACT FROM AUTHOR]