Developing a reporting guideline for social and psychological intervention trials

Social and psychological interventions are often complex. Understanding randomised controlled trials (RCTs) of these complex interventions requires a detailed description of the interventions tested and the methods used to evaluate them. However, RCT reports often omit, or inadequately report, this information. Incomplete and inaccurate reporting hinders the optimal use of research, wastes resources and fails to meet ethical obligations to research participants and consumers. In this paper, we explain how reporting guidelines have improved the quality of reports in medicine, and describe the ongoing development of a new reporting guideline for RCTs: CONSORT-SPI (an extension for social and psychological interventions). We invite readers to participate in the project by visiting our website, in order to help us reach the best-informed consensus on these guidelines (http://tinyurl.com/CONSORT-study).

Social and psychological interventions aim to improve physical health, mental health, and associated social outcomes. They are often complex and typically involve multiple, interacting intervention components (for example several behaviour change techniques) that may act and target outcomes at several levels (for example individual, family, community).1 Moreover, these interventions may be contextually dependent on the hard-to-control environments in which they are delivered (for example healthcare settings, correctional facilities).2 The functions and processes of these interventions may be designed to accommodate particular individuals or contexts, taking on different forms, while still aiming to achieve the same objective.3

Complex interventions are common in public health, psychology, education, social work, criminology and related disciplines. For example, multisystemic therapy is an intensive intervention for juvenile offenders.4 Based on social ecological and family systems theories, multisystemic therapy providers target a variety of individual, family, school, peer, neighbourhood and community influences on psychosocial and behavioural problems. Treatment teams of professional therapists and case-workers work with individuals, their families and their peer groups to provide tailored services. These services may be delivered in homes, social care and community settings. Other examples of social and psychological interventions may be found in reviews by the Cochrane Collaboration (for example the Developmental, Psychosocial and Learning Problems Group; the Cochrane Public Health Group) and the Campbell Collaboration.

To understand their effects and to keep services up to date, academics, policy makers, journalists, clinicians and consumers rely on research reports of intervention studies in scientific journals. Such reports should explain the methods, including the design, delivery, uptake and context of interventions, as well as subsequent results. Accurate, complete and transparent reporting is essential for readers to make best use of new evidence, to achieve returns on research investment, to meet ethical obligations to research participants and consumers of interventions, and to minimise waste in research. However, randomised controlled trials (RCTs) are often poorly reported within and across disciplines including criminology,5 social work,6 education,7 psychology8 and public health.9 Biomedical researchers have developed guidelines to improve the reporting of RCTs of health-related interventions.10 However, many social and behavioural scientists have not fully adopted these guidelines, which may not be wholly adequate for social and psychological interventions in their current form.5,8 Because of the unique features of these interventions, updated reporting guidance is needed.

This article describes the development of a reporting guideline that aims to improve the quality of reports of RCTs of social and psychological interventions. We explain how reporting guidelines have improved the quality of reports in medicine, and why guidelines have not yet improved the quality of reports in other disciplines. We then introduce a plan to develop a new reporting guideline for RCTs - CONSORT-SPI (an extension for social and psychological interventions) - that will be written using best practices for guideline development and dissemination.11 Wide stakeholder involvement and consensus are needed to create a useful, acceptable and evidence-based guideline, so we hope to recruit stakeholders from multiple disciplines and professions. Randomised trials are not the only rigorous method for evaluating interventions; many alternatives exist when RCTs are not possible or appropriate because of scientific, practical and ethical concerns.12 Nonetheless, RCTs are important to policy makers, practitioners, scientists and service users, as they are generally considered the most valid and reliable research method for estimating the effectiveness of interventions.13 Moreover, many of the issues faced in reporting RCTs also relate to other evaluation designs. As a result, this project will focus on standards for RCTs, which could then also inform the development of future guidelines for other evaluation designs.