Background Despite a major research focus on clinical reasoning over the last several decades, a method of evaluating the clinical reasoning process that is both objective and comprehensive is yet to be developed. The aim of this study was to test whether a dual approach, using two measures of clinical reasoning, the Clinical Reasoning Problem (CRP) and the Script Concordance Test (SCT), provides a valid, reliable and targeted analysis of clinical reasoning characteristics to facilitate the development of diagnostic thinking in medical students. Methods Three groups of participants, general practitioners, and third and fourth (final) year medical students completed 20 on-line clinical scenarios -10 in CRP and 10 in SCT format. Scores for each format were analysed for reliability, correlation between the two formats and differences between subject-groups. Results Cronbach’s alpha coefficient ranged from 0.36 for SCT 1 to 0.61 for CRP 2, Statistically significant correlations were found between the mean f-score of the CRP 2 and total SCT 2 score (0.69); and between the mean f-score for all CRPs and all mean SCT scores (0.57 and 0.47 respectively). The pass/fail rates of the SCT and CRP f-score are in keeping with the findings from the correlation analysis (i.e. 31% of students (11/35) passed both, 26% failed both, and 43% (15/35) of students passed one but not the other test), and suggest that the two formats measure overlapping but not identical characteristics. One-way ANOVA showed consistent differences in scores between levels of expertise with these differences being significant or approaching significance for the CRPs. Conclusion SCTs and CRPs are overlapping and complementary measures of clinical reasoning. Whilst SCTs are more efficient to administer, the use of both measures provides a more comprehensive appraisal of clinical skills than either single measure alone, and as such could potentially facilitate the customised teaching of clinical reasoning for individuals. The modest reliability of SCTs and CRPs in this study suggests the need for an increased number of items for testing. Further work is needed to determine the suitability of a combined approach for assessment purposes.
展开▼
