Answers are grounded in this paper and cite their sources.
Background Effective doctor–patient communication is fundamental to patient-centered, safe and ethical clinical practice. Although embedded in India’s undergraduate curriculum through the Attitude, Ethics and Communication (AETCOM) framework, scalable methods for observing, assessing and providing feedback on student communication remain limited. We assessed AETCOM-aligned communication competency among medical undergraduates using AI-enabled simulated consultations, and examined the feasibility of AI-assisted transcript capture and faculty-reviewed communication assessment. Methods An IEC-approved cross-sectional pilot study was conducted among Phase III Part I MBBS students at a tertiary care teaching medical college. Students voluntarily completed audio-only simulated consultations using ConversationAIly, an AI-enabled conversation analysis platform. A 20-item pilot-developed pre-participation questionnaire (Supplementary file) assessed five determinant domains. Raw platform exports were deduplicated and quality screened. Stratified transcript samples were independently scored by two faculty raters using a seven-domain rubric; inter-rater reliability was assessed using ICC (2,1). Post hoc AI rescoring of analyzable transcripts used transcript text and the rubric, with faculty ratings merged only after AI scores were frozen. AI–faculty agreement was evaluated using Lin’s concordance correlation coefficient (CCC), mean difference, mean absolute error, and Bland–Altman analysis. Results Of 123 registered students; 89 students completed at least one consultation. Deduplication yielded 383 unique sessions; 127 transcripts passed quality screening. Faculty inter-rater reliability was excellent (mean ICC = 0.996). Mean Faculty-rated communication score was 22.55/35 (SD 7.35). Avoidance of stigmatizing language scored highest (D6: 4.59/5), while teach-back and safety-netting (D7: 2.14/5) and concern/belief elicitation (D2: 2.60/5) scored lowest. Female students had higher attitude scores than male students (4.36 vs 4.04, p = 0.003), and regular prior clinical communication exposure was associated with higher readiness (p = 0.033). Exploratory AI–faculty agreement for the total score was high (Lin’s CCC = 0.960; mean difference − 1.00/35; 95% limits of agreement − 4.43 to + 2.43). Conclusions ConversationAIly enabled scalable capture and assessment of simulated undergraduate consultations. Baseline performance was weaker in elicitation, rapport building, and teach-back/safety-netting than in explanation and stigma-free language. The rubric demonstrated high faculty reliability. High exploratory AI–faculty concordance warrants prospective evaluation of automated assessment but should not be interpreted as formal validation of AI scoring.
No takes yet. Share an insight, caveat, or question.
Kar et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: