Stefan Elbers

208 Chapter 8 treatment programs. For example, Papies (2016) introduced a framework specifically designed to take unintentional effects on behaviour into account (e.g. an automatically triggered response in an associated environmental context). A more stepwise and methodical approach to optimize behaviour change has been developed by Michie and colleagues (2011). This ‘Behaviour ChangeWheel’, helps to determine what needs to change for pre-specified outcomes to occur and to identify specific intervention components that support this change. This model has already been applied in other rehabilitation contexts to further refine intervention components (e.g. Connell et al., 2016). These ideas also fit the recently developed classification system to characterize rehabilitation treatments – the Rehabilitation Treatment Specification System (RTSS). This framework combines general behaviour change principles, such as habit formation, with rehabilitation treatment components, and aims to open the black box of rehabilitation interventions by clarifying targets for treatment, active ingredients and dosing parameters (Hart et al., 2019; Whyte et al., 2019). The second suggestion for clinical practice is to take stock of the existing heterogeneity within treatment programmes. Although we were able to discern to a certain degree different treatment modalities from the study reports in Chapter 2, the specific treatment content as well as the underlying mechanisms were often not provided. We believe that this is problematic for evidence-based practice, because higher levels of detail are crucial to distinguish homogeneous clusters of studies for meaningful meta-analyses. A suggestion for improvement for IMPT programmes would be to more clearly describe and classify the intervention procedures, assisted by such frameworks as the RTSS and the TIDieR checklist (Dijkers, 2019; Hart et al., 2014; Hoffmann et al., 2014). Sharing this information between facilities would be an opportunity to learn from the existing heterogeneity and to promote refinement of procedures. It would also help to better distinguish certain approaches from others and check to what extent a programme conforms to the biopsychosocial model. This would further help working towards a common understanding of IMPT programmes (Kaiser et al., 2017). Third, the large variation in patient characteristics between and within studies (identified in Chapters 2 and 4) hinders the meaningful interpretation of summary statistics. This aligns with the observations of Smeets (2021) and Eccleston & Crombez (2017) who acknowledge the idiosyncratic nature of pain management and recommend using more single-case studies to better understand this process. For rehabilitation practice, an appropriate example would be the use of ecological momentary assessment (EMA) to monitor changes in key outcomes over time in individual patients. This type of methodology is characterized by multiple assessments over time, capturing the current state of a particular outcome for a subject in a real-world context (Shiffman et al., 2008). EMA not only enhances the