520 chapter 13 scan (i.e., other than the patients who contacted their physician because had difficulties coping with the blinding, as mentioned in the previous paragraph on patient blinding). It has potentially also resulted in several protocol deviations, i.e., patients who declined the recommended surgery, thereby disregarding their a posteriori increased risk of malignancy. I hypothesize that discontent regarding the received treatment may have influenced the reported HRQoL in a negative way. Any diagnostic misestimation-related discontent has likely not resulted in statistically significantly lower response rates of the HRQoL questionnaires, which could have caused data missing at random and subsequent bias: 21 of 23 (91%) patients undergoing active surveillance following an [18F]FDG negative nodule completed all HRQoL questionnaires, as compared to 83 of 103 (81%) patients who underwent diagnostic surgery for a benign or malignant nodule (p=0.22, Pearson Chi-squared test). Altogether, even though we feel that physicians and researchers have put a lot of time and effort in patient counselling during the EfFECTS trial, there may have been room for improvement and further prevention of diagnostic misconception and misestimation. A further simplification of the patient information, an explanatory video on the trial website, and/or a dedicated, independent researcher for all patient counselling should be considered in future diagnostic RCTs with a design similar to the EfFECTS trial. Diagnostic confidence and treatment compliance As shared decision-making is increasingly important to providing good health care, I would like to advocate that diagnostic trials should evaluate patient-bound implementability-related factors, too. This includes measuring the effect of the index test on diagnostic confidence (i.e., the degree of confidence in the established diagnosis) and subsequent treatment compliance to the recommended management [501]. Following the design of the EfFECTS trial, we were able to include these parameters in our assessment of the trial’s main results as well as the cost-utility analysis [28, 501]. As described in Chapter 4, six of twenty patients underwent diagnostic surgery (three during and three after the 1-year protocolled follow-up) despite recommend active surveillance after a negative [18F]FDG-PET/CT scan. Only one of them was not fully reassured by the negative [18F]FDG-PET/CT result. The main reason for surgery in all six patients was not fear of cancer, but (increasing) compressive symptoms that caused globus sensation and discomfort. On the one hand, these numbers show that the diagnostic confidence was very high following a negative [18F]FDG-PET/CT scan for a Bethesda III/IV thyroid nodule. On the other hand, it shows that treatment compliance could be improved, even though the rate of noncompliance and its reasons in the EfFECTS trial are in line with literature and the univariate sensitivity analysis of our cost-utility study showed that [18F]FDG-PET/CT is likely still cost-effective with such figures (Chapter 6, table 4 and figure 4) [28, 529, 530, 532]. Treatment compliance may be improved by considerate patient selection for additional diagnostics and only offer additional diagnostics when the test results are likely to change rather than postpone patient
RkJQdWJsaXNoZXIy MTk4NDMw