219 Discussion understand post-treatment changes, it is important to understand the context in which these changes occur. Specifically, it is important to determine whether the observed changes in patients differ from what would be expected in healthy individuals. For assessing change in cognitive test score a reliable change index (RCI) is commonly used.46,47 The RCI provides a statistical framework to evaluate the significance of changes in individual scores over time, distinguishing reliable changes from those due to chance, practice effects or measurement error alone. Especially practice effects are a common problem in neuropsychological research, which occur when individuals perform better on a test the second time than individuals who perform the test for the first time. By using normative test-retest data from a control population, an RCI eliminates the need for a control group. However, the RCI calculation is also dependent on the availability of normative data. As I observed calculating RCI’s in Chapter 4, this normative data is often limited or completely lacking, which consequently limits the possibility of using more sophisticated RCI calculations. Additionally, finding a norm population that matches one’s own sample and research setup (e.g., age, test-retest interval) can be challenging. The investigation of physiological brain changes is confronted with similar challenges. Previous research in healthy subjects indicated that assessing cerebrovascular reactivity (CVR) using a computer-controlled gas system during Blood Oxygenation Level Dependent (BOLD) imaging exhibits good reproducibility.48 However, to enhance the ability to detect subtle changes specific to treatment effects (e.g. radiotherapy), such as those observed in Chapter 7, it is important to account for natural variation and, again, control for confounding factors (e.g. age). Unfortunately, due to practical constraints, ethical considerations and limited resources a control population is often not included. This while the absence of a control group can limit the ability to draw definitive conclusions. For specific research questions large data repositories like the UK biobank can provide solutions, but ideally research, especially that evaluating treatment effects, should include a matched control population to be sensitive to and better understand the treatment-specific changes. In research grant proposals, allocating budget for control measurements can pose a financial challenge, especially when these costs need to be accommodated in addition to the expenses already incurred for patient data collection and researcher salaries. Consequently, the inclusion of a control population is frequently neglected due to limited available research funding, underscoring the undervaluation of control participants in research. However, a potential solution to this challenge may lie in the combination of control groups across multiple studies, enabling the simultaneous collection of diverse data types. This approach not only offers costsaving benefits but also presents ethical advantages. 9
RkJQdWJsaXNoZXIy MTk4NDMw