Developing the agenda for core outcome set development

Abstract

Introduction and aims A core outcome set (COS) is defined as an agreed standardised set of outcomes that should be measured and reported, as a minimum, in all clinical trials in specific areas of health or health care. Their use allows research to be compared and combined as appropriate, and may ensure that all studies provide usable information. There is currently no accepted gold standard method for COS development and further work was necessary to explore choices about methods, and what the priorities are for guidance and further research in this area. This thesis aimed to investigate what is currently known about COS development, and explore developers’ experiences of developing COS. Methods A systematic review of studies reporting the development of a COS was undertaken, and the methodological techniques used in these studies was described. A mixed methods approach was undertaken to explore COS development, drawing on qualitative interviews with, and an online web-based survey of, COS developers. This thesis used a Triangulation Design to obtain different but complementary data on the same topic for comprehensiveness. Results The systematic review identified 198 published studies that described the development of COS for clinical trials. The systematic review demonstrated variability in the ways that COS had been developed, particularly the methods used and the stakeholders included as participants in the process. Patient participants had infrequently been included in the development of COS (18%). Key aspects of the process were frequently not reported. Eighty-one (48%) developers completed the survey. The majority of survey respondents (73%) felt that there is a need for methodological guidance or research to inform future activity to develop COS. Areas for future guidance or research included: stakeholder involvement, patient involvement in particular; choice of methodology, and consensus formation. 32 interviews were conducted with COS developers (18 with published, and 14 with ongoing, COS projects). Developers found the process of COS development to be a challenging process, in part due to the nature of COS development being an emerging field of research, but also in part to not always considering important methodological details from the outset, for example their choice of methods and stakeholders. There was a variety of influences on developers’ choice of methods, which included the previous literature on COS development, expert advice, developers’ own experience with methods and the resources available to developers. The absence of guidance in COS development, and the prominence of uncertainties, dominated developers’ accounts. Conclusions The work in this thesis has brought COS together in one place for the first time, summarises key characteristics of COS and their development, and provides the first comprehensive account of COS development. It will inform the development of much needed guidance in this area and help to improve COS development methodology. Guidance needs to determine commonalities across different disease areas, and promote awareness of important issues; encourage COS developers to think about their own contexts and circumstances, and enable COS developers to make decisions about methods that best suit their needs and resources. Guidance seems to be needed for all aspects of COS development, but it was particularly felt that guidance around the systematic review process, conduct of Delphi, and conduct of consensus meetings, are high priority.