Decisional Conflict and Regret | 65 2 Increased attention to SDM has led to the development of patient decision aids. Such aids might overcome most of the issues leading to DC as mentioned by participants in the qualitative studies. Evidence suggests that patient decision aids can lead to a clinically meaningful reduction of DC57,58,67,68. To date, only a few decision aids have been developed for patients with HNC, and impact evaluations of these tools are not yet available69,70. Decision Regret Compared to DC, DR has been studied more extensively. The prevalence of DR in the available studies differs vastly, ranging from 0%-86%. The type of measurement instrument clearly influences observed prevalence: study-specific questionnaires all showed a low prevalence (5-36%) whereas the validated questionnaires show evidently higher prevalence (SSS 72-86% and DRS 14-86%), suggesting underestimation of DR when using unvalidated questionnaires. DRS results were highly heterogeneous. Removing the outlier52 on nasopharyngeal cancer did not improve heterogeneity and had limited impact on results. The six other studies using the DRS consisted of a variety of HN tumors, stages, and treatments. Although we hypothesize that tumor location, stage, received treatments, timing of measurement, and the occurrence of complications or poor outcomes influence the level of DR, the available studies prohibit comparisons of clinical subgroups due to the inclusion of mixed populations and limited reporting on these variables. Therefore, we were unable to explain the observed heterogeneity, and the pooled estimate of DR prevalence should be interpreted with caution. A recent systematic review investigating the extent and risk factors of DR after a variety of healthcare decisions showed that DR prevalence is higher in more complex or life-threatening diseases16. The authors reported that risk factors related to the decisionmaking process (e.g., DC, satisfaction with information provided, role in decision-making) were most important, followed by treatment related factors (e.g., complications, adverse outcomes) and rarely sociodemographic characteristics (e.g., age, education). However, the review did not include studies on HNC and future research is needed to identify risk factors for DR in the HNC population. Limitations of the review process We used a highly sensitive search to identify all possible studies. To manage the vast number of records retrieved, we used an AI-supported inclusion process. As a result, we did not manually screen all identified records. However, we used a very conservative approach, using a much higher than minimally recommended number of manual inclusion decisions22, and screening a substantial amount of records with a 2.0 star rating or less. At a 2.5 star rating as threshold for exclusion, sensitivity of Rayyan’s AI method is reported to be 100%. We therefore believe it is unlikely that we have missed relevant publications.
RkJQdWJsaXNoZXIy MTk4NDMw