85 Studying physician-related attributes of SDM reference the scoring category; for example, SDML for SDM case vignette scores in the lowest Category, L. The supplemental file, the Appendix, provides a more comprehensive description of the statistical methods used. Results Overall, 114 potential participants were contacted from five different hospitals. Thirty-eight (33%) participated. Participants consisted of nine men and twenty-nine women. Fifteen were trainees. Twenty-three participants ranked vignettes for job satisfaction, and fifteen ranked for complexity. Table 1 provides an overview of case vignette-type contributions per preference category. Results from comparisons and statistical analysis are represented in the following two paragraphs. Most observed results are compared to expected results based on our null hypothesis. Job satisfaction When comparing observed vignette counts with expected sums, as displayed in Table 1, it was clear that sum scores S deviated significantly from the chance level. SDM was assigned to Category L more often than expected and less often than expected to Category H. The same conclusion holds for EMOL and EMOH. The opposite conclusion holds for TECH, which was assigned significantly less often than expected to Category L and more often to Category H. After the Bonferroni correction was applied, all tests except for EMOH were still significant. Previous results were consistent with pairwise comparisons. Results from these comparisons are partially displayed in Table 2. For Category L, the differences between TECHL and SDML (S = -32, E = 6.9, p < 0.001) and between TECHL and EMOL (S = -28, E = 6.9, p < 0.001) were significantly smaller than expected, with TECH vignettes being picked far less often in Category L. The difference between SDML and EMOL was not significant (S = 4, E = 0, p = 0.428). For Category H, the difference between TECHH and SDMH was significantly larger than expected (S = 48, E = 6.9, p < 0.001), with TECH being picked far more often. The difference between TECHH and EMOH was more extensive than expected as well (S =- 39, E = 6.9, p < 0.001), but the difference between SDMH and EMOH was not significant (S = -9, E = 0, p = 0.094). These patterns remained intact after the Bonferroni correction. These results are reasons to reject the null hypothesis of indifference in the low (L) and high (H) rating categories of ‘job satisfaction’, as handling TECH cases seems to lead to higher job satisfaction and dealing with SDM and EMO cases leads to lower job satisfaction. Complexity The significance tests showed that EMO was assigned to L less frequently than expected from the chance level, and TECH was assigned to L more frequently than expected. However, only the first deviation was significant after the Bonferroni correction. The other frequencies, displayed in Table 1, did not deviate significantly from the chance level. 5
RkJQdWJsaXNoZXIy MTk4NDMw