Evaluating real-patient learning in medical education - Hungarian validation of the Manchester Clinical Placement Index.

Introduction: The Manchester Clinical Placement Index (MCPI) is an instrument to measure medical undergraduates' real-patient learning in communities of practice both in hospital and in GP placements. Its suitability to evaluate the quality of placement learning environments has been validated in an English-language context; however, there is a lack of evidence for its applicability in other languages. Our aim was to thoroughly explore the factor structure and the key psychometric properties of the Hungarian language version. Methods: MCPI is an 8-item, mixed-method instrument which evaluates the quality of clinical placements as represented by the leadership, reception, supportiveness, facilities and organization of the placement (learning environment) as well as instruction, observation and feedback (training) on 7-point Likert scales with options for free-text comments on the strengths and weaknesses of the given placement on any of the items. We collected data online from medical students in their preclinical (1st, 2nd) as well as clinical years (4th, 5th) in a cross-sectional design in the academic years 2019-2020 and 2021-2022, by the end of their clinical placements. Our sample comprises data from 748 medical students. Exploratory and confirmatory factor analyses were performed, and higher-order factors were tested. Results: Although a bifactor model gave the best model fit (RMSEA = 0.024, CFI = 0.999, and TLI = 0.998), a high explained common variance (ECV = 0.82) and reliability coefficients (ωH = 0.87) for the general factor suggested that the Hungarian version of the MCPI could be considered unidimensional. Individual application of either of the subscales was not supported statistically due to their low reliabilities. Discussion: The Hungarian language version of MCPI proved to be a valid unidimensional instrument to measure the quality of undergraduate medical placements. The previously reported subscales were not robust enough, in the Hungarian context, to distinguish, statistically, the quality of learning environments from the training provided within those environments. This does not, however, preclude formative use of the subscales for quality improvement purposes.