Which trial design is best for the study of physical, psychological, and self-management interventions for people with persistent pain? developing best-practice trial methods across the pragmatic-explanatory continuum

Abstract

This PhD project aimed to address major challenges of designing and conducting randomised controlled trials (RCTs) for physical, psychological, and self-management (PPS) interventions for people with pain. Such challenges include designing adequate ‘sham’ control interventions and ensuring participant and provider blinding to control for context-dependent effects. Studies are also commonly criticised for limited generalisability of results beyond the immediate study population and setting; high generalisability being a typical objective of so-called pragmatic trials. The primary objective of this PhD was to provide evidence-based guidance for trial design that is specific to PPS interventions and covers best-practice methods across the pragmatic-explanatory spectrum. A secondary objective was to explore the developed recommendations on the example of a feasibility trial protocol, studying a multimodal intervention for people with painful diabetic neuropathy.

The project involved two systematic reviews of trial methods. One describing self-declared ‘pragmatic’ and ‘comparative effectiveness’ pain trials, and another PPS trials employing ‘sham’ or ‘placebo controls’. The latter involved a meta-regression analysis, which demonstrated that the level of similarity between experimental and control interventions can bias trial outcomes. Informed by these reviews, two independent consensus processes were conducted. One process led to two guidance statements for pragmatic trials in pain research. This was based on expert-led discussions as part of the dedicated 2020 Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) meeting. The second consensus process involved a modified Delphi survey with trial designers and placebo effect researchers, interviews with patient partners, and live consensus discussions. This process resulted in a comprehensive guideline for control intervention design, conduct, and reporting in PPS trials. The so-called ‘CoPPS Statement’ was published in the BMJ and adopted by the EQUATOR network library. Finally, a multimodal osteopathy-based intervention was created, and a feasibility trial protocol written to test the usability of the guidance frameworks. The systematic reviews, consensus processes, and feasibility trial protocol development are described in detail in this thesis.

As a result of this PhD, trial designers now have access to two comprehensive evidence-based frameworks to guide their decision-making in the design, conduct, and analysis phases of PPS trials. End-users of trials can employ these frameworks to guide their interpretation of such studies. Overall, these novel frameworks address major methodological criticisms of earlier trials in a manner that is specific to PPS interventions and may allow for more rigorous and transparent future research.

Furthermore, this project highlighted that while nonpharmacological interventions are distinct from drugs, RCTs of PPS interventions can still be of high quality. The developed methodological guidance statements facilitate this by accounting for differences to drug research, providing rigorously developed and innovative resources for researchers designing and conducting RCTs of PPS interventions for people with pain and other conditions. The usability evaluation as part of the feasibility trial protocol development showed the CoPPS Statement to be comprehensive and clear for the design of control interventions in an efficacy-focused trial. The IMMPACT statements were less prescriptive but were found relevant for trials across pragmatic-explanatory spectrum. However, the distinctiveness of complex nonpharmacological intervention research will remain a challenge for the involved professions, for evidence synthesis, and for policy decision-making.

Project limitations include a lack of practical testing of the developed recommendation frameworks prior to their publication and insufficient involvement of people from regions other than Europe, the US, and Australasia. Future research needs to determine the applicability of the developed frameworks to other therapeutic areas and settings. Their long-term impact on trial quality and practical usefulness will also have to be evaluated to identify potential needs for refinement.