Healthcare workers want to listen more to patients and their carers in all sorts of areas of healthcare. This can include choosing topics for medical research. We looked at how patients and carers have helped to choose topics for research about type I diabetes. We aimed to find out if, and why, researchers often rejected their choices. We looked at a project which brought together patients, carers and healthcare workers to choose topics for research about type 1 diabetes. The group first asked patients, carers and healthcare workers to suggest ideas for research questions. But the group had to follow rules about what counted as a good research question. Some people's ideas did not count as good research questions, and they were rejected at the start. We looked at who were most likely to have their ideas rejected at the start. We found that patients and carers were most likely to have a suggestion rejected. Then we looked at the rejected questions in detail. They were mostly about curing diabetes, preventing diabetes and understanding how diabetes works. There were also some questions about access to medicines and the quality of care. Researchers should ask patients and carers for help deciding what counts as a good research question from the start of projects like these. We should also think about what might be getting in the way of patients and carers making more of a difference in research.
Background Patients and carers are increasingly involved in deciding on topics for medical research. However, so far, it has been difficult to gain an accurate picture of the impact of such involvement because of poor reporting and evaluation in published studies to date. This study aimed to explore how a partnership of patients, carers, healthcare professionals and organisations identified questions for future research and why patients and carers had a limited impact on this process. Methods In the first stage of the partnership process, relevant service users and providers (including patients, carers, healthcare professionals and voluntary organisations) were invited to submit suggested research questions about the treatment of type 1 diabetes, via a national online and paper survey. The partnership followed formal protocols that defined a researchable question. This meant that many respondents' suggested research questions were rejected at the start of the process. We analysed survey submissions to find out which groups of respondents were most likely to have their suggestions rejected and what these suggestions were about. Results Five hundred eighty-three respondents submitted 1143 suggested research questions, of which 249 (21.8 %) were rejected at the first stage. Respondents with lived experience of this long-term condition (patients and carers) were more likely than those without lived experience to submit a research question that would be rejected (35.6 vs. 16.5 %; p < 0.0005). Among the rejected questions submitted by patients and carers, there were several key themes: questions about cure, cause and prevention, understanding the disease, healthcare policy and economics. Conclusions In this case study, early decisions about what constituted a researchable question restricted patients' and carers' contributions to priority setting. When discussions about a project's remit take place before service users are involved, researchers risk distorting the potential impact of involvement. Impact assessments should consider not only the differences patients and carers make to research but also the differences they could have made in the absence of systemic barriers. We recommend that initiatives aimed at involving patients and carers in identifying research questions involve them as early as possible, including in decisions about how and why suggested research questions are selected or rejected.