The participants were the 13 Internal Medicine/Pediatrics residents, and represented all four years of training (2, 4th-year; 3, 3rd-year; 4, 2nd year; and 4, 1st-year residents). A third year resident failed to complete the 0-month tests and surveys, and a first year resident failed to complete the 12-month tests and surveys. This left 12 responses for both sets (0- and 12-month) tests and surveys. The average attendance for the sessions was 12, with a range of 10–13 attendees.
The results from the knowledge test are given in Figure . As a group, the residents' mean score was 74% (95% CI, 68%–80%) for the 0-month survey and 91% (95% CI, 85–96%) for the 12-month survey. These confidence intervals do not overlap. This suggests that, if hypothesis testing were done, the results would probably reach statistical significance for the knowledge test.
Practice knowledge test results (mean and 95% CI): before (0-month) and after (12-month) the course
On follow-up, we performed two post-hoc analyses. First, we were interested if these knowledge scores would decline over time. Therefore, we compared the knowledge test scores on the first six months topics to the scores on the last six months topics. Both sets of scores were derived from the 12-month knowledge test. We found that the mean scores appeared similar (first 6 months mean scores: 92% correct; the last 6 months mean scores: 90% correct).
Second, we were interested if the missing data on the 0-month and 12-month data could have impacted the results. Since one third-year resident completed the 12-month but not the 0-month test, we were interested in exploring if his responses on the 12-month test could have caused a larger difference between these two tests. After censoring his data, there appeared to be little impact on the 12-month results (censored mean score = 0.91, censored 95% CI, 85–96%). Additionally, the first year resident who failed to complete the 12-month test may have also impacted the results. Due to loss of identity links, we could identify her data to censor from the 0-month test. However, we censored the lowest score on the 0-month test as representing hers (this assumes that her score lowered the 0-month data the most, and, therefore, had the largest impact on 0-month mean score by skewing it away from the 12-month mean score). After censoring this data, we found no significant change in the 0-month results (censored mean score 0.75, censored 95% CI, 70–80%).
The results for the self-assessed skill survey are given in Table . The mean scores on the 12-month survey (3.65) were higher than in the mean scores for the 0-month survey (2.62). The confidence intervals from this data do not overlap. This suggests that, if hypothesis testing were done, the results would probably reach statistical significance for the self-assessed practice management skills survey.
The results of the curriculum evaluation survey are given in Table . All statements had a mean rating of greater than 3.90. The two statements assessing the residents' views towards practice management education in general (value of practice management education and the need to expand their education) both had mean scores of 4.67.