Chapter 2 54 Because the results of the interviews suggested that variance in fidelity is related to the differences in the trainers’ professional skills, this gives input for further education and supervision. We have decided to add the IT-IS scale (50) to the RCT, as it focuses on measuring clinicians’ skills related to fidelity. One of our feasibility criteria for a good infrastructure for education and supervision was to give all 12 trainers two days’ training, and to supervise them for two hours per week. A two days’ training was indeed achieved for all 12 trainers, and the trainers had eight hours of supplemental training in one year. These 24 hours were more than the length of training in five studies in McGuire’s review, but less than in three other studies (one of which had involved only 50% of the trainers) (see above). Although the institute had intended to provide supervision once weekly, productivity requirements meant that trainers could attend supervision only once every two weeks for two hours throughout the duration of the pilot study and thereafter. This was more intensive than in one RCT, which reported monthly supervision (6), but less intensive than in another, which reported weekly supervision (8). Because, in one year, the institute successfully achieved a good average total fidelity score of 4.0, we suggest that fidelity may be further improved by continuing such education and supervision, especially if the focus lies on aspects of fidelity that most need improvement (see above). Although our results on dropout from IMR were worse than those reported in the review on IMR (5), and also than those in one earlier RCT on IMR (6), they do not appear to be worse than the numbers in two other RCTs (7, 9), even though the overall results are not entirely comparable: in our study, dropout is defined as < 70% participation. But with regard to the planned RCT, we conclude that if no measures are taken to reduce it, allowance should be made for substantial dropout from IMR. McGuire et al (5) report a median completion rate of 63% with a range of 15%-86%. If we adopt his definition of program completion—having received all IMR modules, which in our study would mean a completion rate of 44%— our result falls well within this range. But as McGuire et al said with regard to the studies they reviewed, we feel that the dropout and completion rates found in our study “leave much room for improvement.” For the RCT, we plan to recommend clinicians and managers to pay attention to this aspect and to take various measures including the use of a good dropout protocol. Our baseline analysis of completers’ and non-completers’ characteristics also suggest that, to reduce dropout from IMR, special attention is required by male participants, people who receive social security benefit, and people who score lower on the IMR scales at baseline. And although “intention to treat” will be used in the RCT, we have decided that a ratio of 3:2 will be used for
RkJQdWJsaXNoZXIy MjY0ODMw