382 chapter 6 following any first Bethesda III or IV result), but was likely the cost-effective strategy due to the high costs of molecular testing. Their results were most sensitive to the costs of molecular testing [579]. The number of cost-effectiveness studies from a European perspective is limited. A recent study from a Dutch perspective estimated that molecular testing may save a considerable number of repeat FNAC procedures and diagnostic surgeries in Bethesda III and V nodules, resulting in a net saving of €100 and €4,100 for these cytological categories, respectively. Unfortunately, the study excluded Bethesda IV nodules from their analysis [525]. To the best of our knowledge, the current study is the first cost-utility analysis on additional diagnostics in indeterminate thyroid nodules to be performed alongside a randomised controlled clinical trial. This contrasts our study with previous cost-utility analyses and provides a unique perspective. Our observed first-year healthcare consumption data and quality of life assessments are unparalleled, especially in patients with indeterminate thyroid nodules. By incorporating these data into a comprehensive lifelong cost-utility model, we presented a scenario that most accurately reflects real-world clinical practice. In contrast, most previous cost-utility studies used a theoretical base case, a more simplified model, somewhat idealized parameters, and/or a limited time horizon. Any lifelong HRQoL effects and (lifelong) costs other than the direct medical costs (i.e., costs for other health care consumption, patient costs, and productivity losses) were often disregarded in these studies [29, 176, 178, 495, 579]. In previous studies, the possibility of patient crossover between management strategies over time was also seldom taken into account [29, 176, 178, 495, 579]. We previously recognized that the therapeutic yield of [18F]FDG-PET/CT is influenced by patient preference and treatment compliance. This directly reflects on health-care consumption volumes and costs. Shared decision-making is crucial to carefully determine the most suitable management strategy for individual patients and prevent noncompliance, as well as to optimize the use of valuable diagnostic resources [501]. This is a dynamic process, in which preferences and interests may change as time passes. In studies on the natural course of cytologically benign nodules, up to 24% of nodules were surgically resected as time passed, primarily due to compressive symptoms [530-532]. It is important to acknowledge the dynamics of clinical practice in a cost-effectiveness model, too, as this may prevent overestimation of an effect of any given strategy. To account for this, our model included a yearly probability of surgery despite a negative [18F]FDG-PET/CT, a probability that surveillance of an [18F]FDG-negative nodule would end, and a probability to re-enter active surveillance after it had previously ended (Table 2, Figure 2). For the Markov model, we used triangular distributions for probabilities (Table 2) and utilities (Table 3) when uncertainty about these parameters was asymmetric. The base-case parameter value was the mode of the triangular distribution. Due to the asymmetry, the mean parameter value in the analysis was typically higher than the base-case value (by on average 18%, at most 67%). For the utilities, the higher mean of some parameters could be in favour of the [18F]FDG-PET/CT-driven group
RkJQdWJsaXNoZXIy MTk4NDMw