Related Experiment Video
Updated: Sep 27, 2026

Virtual Agent for Real-Time Motivational Interviewing by Integrating Adaptive Nonverbal Behavior and Language Models
Published on: December 23, 2025
Evaluation of a Motivational Interviewing-Based Training Programme in Vaccination Counselling for Family Physicians:
Roxana Surugiu1, Virginia-Maria Rădulescu2, Anca Deleanu3
1Biochemistry Department, University of Medicine and Pharmacy Craiova, 200349 Craiova, Romania.
Abstract:
Background/Objectives: Motivational interviewing (MI) is an established approach to addressing vaccine hesitancy, yet immediate changes following dedicated training for family physicians remain insufficiently documented in Romania. This study evaluated pre-post changes in physicians' MI knowledge, self-reported communication behaviour, and self-perceived ability to counsel patients about vaccination, together with patient-level change following consultations delivered by trained physicians. Methods: This uncontrolled pre-post study used a voluntary, self-selected non-probability sample of 28 family physicians recruited through announcements distributed via county family medicine associations; participants completed paired assessments immediately before and after structured MI-based training. The three main physician-level outcomes were MI knowledge (MISI), corrected self-reported MI communication behaviour (MIBI), and self-perceived MI ability (SPMI). The linked component comprised 151 patients nested within 15 physicians. Patient changes were averaged within physician, and the physician-level mean patient composite change was used as the principal outcome of this component. Paired physician changes and physician-aggregated patient change were assessed using Wilcoxon tests, rank-biserial effect sizes, and bootstrap 95% confidence intervals (CIs), with Holm adjustment across the three physician outcomes. Results: Mean MISI increased by 39.80 percentage points (95% CI: 29.59-50.51), though the limited internal consistency of this seven-item composite requires item-level interpretation, and corrected MIBI by 0.44 points (95% CI: 0.20-0.68; rrb = 0.75) and SPMI by 1.22 points (95% CI: 0.89-1.61; rrb = 0.99); all Holm-adjusted p-values were <0.001. Across the 15 linked physicians, the mean patient composite change was 0.55 points (95% CI: 0.38-0.75; rrb = 1.00; p < 0.001). No consistent association was observed between physician-level competency changes and mean patient change. Conclusions: Immediate post-training physician scores and post-consultation patient composite scores were higher. Controlled studies incorporating observed consultations and follow-up are required before causal or sustained effects can be inferred.
