Background: Communication is a common element in all medical consultations, affecting a range of outcomes for doctors and patients. The increasing demand for medical students to be trained to communicate effectively has seen the emergence of interpersonal communication skills as core graduate competencies in medical training around the world. Medical schools have adopted a range of approaches to develop and evaluate these competencies.

Objectives: To assess the effects of interventions for medical students that aim to improve interpersonal communication in medical consultations.

Search methods: We searched five electronic databases: Cochrane Central Register of Controlled Trials, MEDLINE, Embase, PsycINFO, and ERIC (Educational Resource Information Centre) in September 2020, with no language, date, or publication status restrictions. We also screened reference lists of relevant articles and contacted authors of included studies.

Selection criteria: We included randomised controlled trials (RCTs), cluster-RCTs (C-RCTs), and non-randomised controlled trials (quasi-RCTs) evaluating the effectiveness of interventions delivered to students in undergraduate or graduate-entry medical programmes. We included studies of interventions aiming to improve medical students' interpersonal communication during medical consultations. Included interventions targeted communication skills associated with empathy, relationship building, gathering information, and explanation and planning, as well as specific communication tasks such as listening, appropriate structure, and question style.

Data collection and analysis: We used standard methodological procedures expected by Cochrane. Two review authors independently reviewed all search results, extracted data, assessed the risk of bias of included studies, and rated the quality of evidence using GRADE.

Main results: We found 91 publications relating to 76 separate studies (involving 10,124 students): 55 RCTs, 9 quasi-RCTs, 7 C-RCTs, and 5 quasi-C-RCTs. We performed meta-analysis according to comparison and outcome. Among both effectiveness and comparative effectiveness analyses, we separated outcomes reporting on overall communication skills, empathy, rapport or relationship building, patient perceptions/satisfaction, information gathering, and explanation and planning. Overall communication skills and empathy were further divided as examiner- or simulated patient-assessed. The overall quality of evidence ranged from moderate to very low, and there was high, unexplained heterogeneity. Overall, interventions had positive effects on most outcomes, but generally small effect sizes and evidence quality limit the conclusions that can be drawn. Communication skills interventions in comparison to usual curricula or control may improve both overall communication skills (standardised mean difference (SMD) 0.92, 95% confidence interval (CI) 0.53 to 1.31; 18 studies, 1356 participants; I² = 90%; low-quality evidence) and empathy (SMD 0.64, 95% CI 0.23 to 1.05; 6 studies, 831 participants; I² = 86%; low-quality evidence) when assessed by experts, but not by simulated patients. Students' skills in information gathering probably also improve with educational intervention (SMD 1.07, 95% CI 0.61 to 1.54; 5 studies, 405 participants; I² = 78%; moderate-quality evidence), but there may be little to no effect on students' rapport (SMD 0.18, 95% CI -0.15 to 0.51; 9 studies, 834 participants; I² = 81%; low-quality evidence), and effects on information giving skills are uncertain (very low-quality evidence). We are uncertain whether experiential interventions improve overall communication skills in comparison to didactic approaches (SMD 0.08, 95% CI -0.02 to 0.19; 4 studies, 1578 participants; I² = 4%; very low-quality evidence). Electronic learning approaches may have little to no effect on students' empathy scores (SMD -0.13, 95% CI -0.68 to 0.43; 3 studies, 421 participants; I² = 82%; low-quality evidence) or on rapport (SMD 0.02, 95% CI -0.33 to 0.38; 3 studies, 176 participants; I² = 19%; moderate-quality evidence) compared to face-to-face approaches. There may be small negative effects of electronic interventions on information giving skills (low-quality evidence), and effects on information gathering skills are uncertain (very low-quality evidence).  Personalised/specific feedback probably improves overall communication skills to a small degree in comparison to generic or no feedback (SMD 0.58, 95% CI 0.29 to 0.87; 6 studies, 502 participants; I² = 56%; moderate-quality evidence). There may be small positive effects of personalised feedback on empathy and information gathering skills (low quality), but effects on rapport are uncertain (very low quality), and we found no evidence on information giving skills. We are uncertain whether role-play with simulated patients outperforms peer role-play in improving students' overall communication skills (SMD 0.17, 95% CI -0.33 to 0.67; 4 studies, 637 participants; I² = 87%; very low-quality evidence). There may be little to no difference between effects of simulated patient and peer role-play on students' empathy (low-quality evidence) with no evidence on other outcomes for this comparison. Descriptive syntheses of results that could not be included in meta-analyses across outcomes and comparisons were mixed, as were effects of different interventions and comparisons on specific communication skills assessed by the included trials. Quality of evidence was downgraded due to methodological limitations across several risk of bias domains, high unexplained heterogeneity, and imprecision of results. In general, results remain consistent in sensitivity analysis based on risk of bias and adjustment for clustering. No adverse effects were reported. 

Authors' conclusions: This review represents a substantial body of evidence from which to draw, but further research is needed to strengthen the quality of the evidence base, to consider the long-term effects of interventions on students' behaviour as they progress through training and into practice, and to assess effects of interventions on patient outcomes. Efforts to standardise assessment and evaluation of interpersonal skills will strengthen future research efforts.