Contribution of examiner subjectivity and patient heterogeneity on long case examination

To evaluate the effect of examiner subjectivity and heterogeneity among the cases on scores from the Royal College of Physicians (RCPT) long case examination.

Data from internal medicine candidates who performed clinical part of RCPT board certifying examination in academic year 2008 were collected. For each candidate, scores from pair of examiners for each of the long case was stratified based on disease category according to the course syllabus into 3 groups; very common, common and uncommon diseases. The scores also categorized according to difficulty level subjectively and rated by the examiners into 3 levels; easy, moderate and difficult. Mean scores in each group of encounters were compared using ANOVA.

There were 21 examination centers involved with 1,840 number of encounters by 232 candidates. Among 437 patients that have been used for the long case, common scenarios (27.6% of the total) were cirrhosis, hyperthyroidism, cerebral thrombosis, bronchogenic carcinoma, rheumatic heart disease and thalassemia. Mean and SD of scores from the very common, common and uncommon diseases were 75.5 +/- 11.6, 75.6 +/- 10.6 and 74.7 +/- 11.3 respectively, with no statistical significant difference between the groups. Mean and SD of scores from the easy, moderate and difficult cases were 76.1 +/- 10.5, 74.8 +/- 11.0, 75.5 +/- 10.9 respectively. The moderate group has the lowest score with a statistical significant difference from other groups (p = 0.042).

In current RCPT long case examination, difficulty of the case appears to contribute to variation in scores derived from the examiners. Measures for score adjustment and examiner calibration should be implemented in the future.