Establishing consensus on domains and items for Reporting intervention Fidelity in Non-Drug, non-surgical trials: the ReFiND Delphi study

Objectives: To develop consensus on core domains and items to inform a reporting guideline for intervention fidelity in non-drug, non-surgical trials (ReFiND). Design: International, online three-round Delphi process. Methods: We invited researchers with expertise in intervention fidelity, non-drug, non-surgical trials, trial reporting, and process evaluation methods to take part in a three-round Delphi study. They rated the importance of 7 intervention fidelity domains and 79 items using a 9-point scale (1 to 3= not important, 4 to 6= important but not essential, 7 to 9= very important). A comment box was provided along with each domain and item for panellists to provide the rationale for their ratings. Consensus was defined as more than 70% of panellists rating 1-3 (exclusion) or 7-9 (inclusion) after two voting rounds. A summary of ratings and panellists rationale from Rounds One and Two were provided to panellists in Round Two and Three, respectively. We conducted a thematic analysis to synthesise qualitative data. Results: Sixty-one panellists completed Round One, 55 completed Round Two and 56 completed Round Three. Panellists had diverse backgrounds and were from 21 countries, including six low- and middle-income countries. After three rounds, five domains (Intervention Design, Intervention Fidelity Measurement Methods, Provider Training, Intervention Delivery by Providers, Intervention Receipt by Participants) and 20 items reached consensus for inclusion in the reporting guideline. Panellists rationale for their ratings was based on how specifically each domain and item related to intervention fidelity, how feasible it would be to implement in practice, and how broadly it could be applied across different contexts. Conclusions: Using systematic methods, this study reached consensus on a core set of domains and items for inclusion in ReFiND. This core set provides a foundation for consistent and transparent reporting of intervention fidelity and helps researchers and clinicians better interpret and replicate interventions tested in non-drug, non-surgical trials.