Impact of Virtual Standardized Patient Training on Diagnostic Exam Performance in Medical Students: A Cluster Randomized Controlled Trial

Title: Cluster Randomized Controlled Trial Evaluating the Impact of Dual-Teacher and Virtual Standardized Patient Collaborative Training on Diagnostic Performance Among Clinical Medicine Undergraduates Background and Purpose: Background: To address the challenges of disconnection between basic and clinical knowledge and insufficient resources for clinical reasoning training in clinical medical education, this project developed a Dual-Teacher and Virtual Standardized Patient collaborative teaching model. Currently, this model lacks high-quality empirical evidence. Purpose: Through a rigorous randomized controlled trial, to empirically demonstrate the immediate additive effect of this model in Diagnostics teaching and explore its potential impact on long-term clinical competency. Research Methods: Design: A cluster randomized controlled trial based on natural classes, using a superiority test. Subjects and Setting: All 8 natural classes of the 2023 clinical medicine major, implemented in the Diagnostics course on the main campus. Randomization and Allocation Concealment: This study employs a dual-executor, two-stage fully concealed lottery method. The specific standard operating procedure is as follows: Stage 1 (Code Assignment): Independent Executor A organizes student representatives from each class to randomly draw confidential codes (A, B, C, D), forming a class-code correspondence table. After being signed by the representatives, it is sealed on the spot and handed to the supervisor for safekeeping. Executor A is unaware of the subsequent grouping rules. Stage 2 (Lottery Grouping): Another independent Executor B (unaware of the code-class correspondence) draws lots for the codes under video recording. The first two codes drawn are designated as the intervention group. Stage 3 (Revelation and Announcement): After the draw, the supervisor publicly opens the sealed envelope, announces the classes corresponding to the codes, and completes the group announcement. The entire process is video-recorded and archived. This procedure ensures that no one can predict the specific grouping of any class before the revelation. Blinding: Course instructors, outcome assessors (graders), and data analysts are blinded; students and intervention implementers are not blinded. Intervention Measures: Intervention Group: Standard teaching plus after-class independent use of the VSP system for training. Control Group: Receives only the identical standard teaching. The teaching syllabus, class hours, and instructors are completely identical for both groups. The only variable is whether VSP is provided. Observation Indicators and Measurement Time Points: Primary Outcome: Comprehensive Diagnostics course score (theory plus practice) measured at course end. Secondary Outcomes: Diagnostic theory score, practical skills score, VSP system usage adherence data, student satisfaction measured at course end or during. Exploratory Outcome: National Clinical Medicine (Undergraduate) Proficiency Test score measured approximately 1 year after the Diagnostics course ends. Sample Size: Using classes as the analysis unit, based on all available teaching units, totaling 8 classes (4 intervention clusters, 4 control clusters). Statistical Analysis: Primary analysis follows the intention-to-treat principle. A multilevel model will be used to compare score differences between groups, controlling for baseline scores and other covariates. Sensitivity analyses will be conducted for missing data (e.g., worst-case scenario analysis). Ethics and Registration: This study has been approved by the Medical Ethics Committee of Chengde Medical College (Approval No: 2025068).