Stefan Elbers

54 Chapter 2 these programmes. Despite a common heritage, the results of this assessment indicate that the included interventions do not share an equal underlying effect. Rather, interventions generally consist of a unique collection of multiple different action mechanisms that are generally not explicitly described. Strengths and Limitations We encountered several problems regarding interpretation of the study data. First, for the majority of the cohorts, we identified a risk of bias, which negatively influences the validity or our results. Especially the study attrition rates, indicating that a substantial minority dropped out of the programme or discontinued participation, introduce significant non- response bias. Furthermore, incomplete reporting of the intervention and its outcomes remains an issue. To increase accuracy of reporting as well as improved understanding of how a particular IMPT programme may benefit patients, we restate the recommendation by Williams and colleagues (2012) to provide a clear rationale for that particular set of treatment components and to test this by including process measures (e.g. Nicholas et al., 2014), instead of generally referring to a biopsychosocial approach. A practical tool that supports clear reporting is the TIDieR checklist, which includes clear guidance for reporting study rationale and action mechanisms (Hoffmann et al., 2014). Beyond the investigation of treatment benefit and harms of IMPT a standardized level of comparison, e.g. a core outcome set (Williamson et al., 2017), is required for harmonization of outcome assessment and supporting detailed meta analyses. Heterogeneous outcome assessment in the context of IMPT has been consistently reported (Deckert et al., 2016; Kamper et al., 2014; Waterschoot et al., 2014). Involving the patient perspective in defining helpful treatment approaches is generally recommended for such actions (Williamson et al., 2017). For IMPT an international initiative has developed a core outcome set for a domain set of outcomes, comprising the biopsychosocial impact of chronic pain in patients undergoing IMPT (Kaiser et al., 2018). Implementing such recommendations would help also to reduce reporting bias and enhance reporting quality of clinical trials (Williamson et al., 2017). On a smaller scale, successful implementation of similar initiatives has resulted in improved collaboration between healthcare services and a homogeneous dataset (Tardif et al., 2017). It should be noted that such initiatives are either shaped by national requirements and resources, commonly organized in national registries, or aim for international application in clinical trials (Kaiser et al., 2018). An important future challenge is to harmonize these approaches in order to achieve results in both objectives. Second, the categorization of treatmentmodalities is likely to contain erroneous interpretations, either due to incomplete reporting or to misinterpretation during the data extraction. Moreover, it is important to realize that the categories for the treatment modalities, the classification of tailoring and what constitutes as an IMPT programme remain arbitrary and leave room for discussion