Improving retrospective intervention descriptions: Lessons learned from research on type 2 diabetes programmes in the United Kingdom and the Republic of Ireland.

Abstract

In recent years, multiple countries worldwide have implemented behavioural interventions within national healthcare systems. Describing the content of these interventions is critical to improve their implementation, replication, and effectiveness, as well as to advance behavioural science. Tools, such as the Behaviour Change Technique Taxonomy, can enhance the quality of intervention description and reporting. As interventions are frequently developed without the use of such tools, retrospective coding of existing interventions to accurately characterise their content is becoming more common. However, the use of these tools for retrospective coding poses various challenges, the discussion of which has been neglected to date. This commentary discusses the challenges encountered when retrospectively describing the content of five nationally implemented programmes for type 2 diabetes in the United Kingdom and the Republic of Ireland and suggests recommendations to tackle these challenges. We present important methodological, practical, and ethical considerations for researchers to reflect on, relevant to the retrospective description of existing interventions. Specifically, we discuss (i) the importance of positive relationships and collaboration with intervention stakeholders, (ii) the practical and ethical considerations when analysing the content of implemented interventions, (iii) the independence of research teams and the potential for misclassification of intervention content, and (iv) the challenges associated with the analysis of intervention content using behavioural science tools. There is a growing demand for more robust approaches to address the methodological, practical, and ethical challenges associated with such studies. The present commentary describes key issues to be considered by research teams, as well as concrete recommendations to improve the retrospective characterisation of intervention content.