Background Public involvement (PI) in research, promoted by the UK Department of Health, has gained considerable momentum. Despite this, examination of the actual impacts of this involvement or how such impacts might be measured effectively remains limited. Evidence about the impact of PI is essential to ensure integrity, avoid harmful effects and to justify resources. In light of this, the overall aim of this MRC multi-phase study was to advance understanding of the impacts of PI, to contribute to more robust assessment of these impacts and to improve standards of PI practice. The Delphi process described here formed part of the multi-phase study.

Methods A modified Delphi approach integrated qualitative and quantitative research methods. Three workshops were conducted to generate qualitative data around PI values and normative debates; and to generate items for a subsequent two-round Delphi survey. Critical consensus and clear consensus criterion were pre-determined at 70% and 60% respectively. Through open questions participants were able to make additional comments and to generate potential solutions to identified problems. Written responses were analysed thematically. Workshop and Delphi study participants were drawn from diverse settings including universities, health and charitable organisations. Participants included members of the public, academic, clinical and user-researchers, research funders and managers, self-selected as having some clear previous experience of PI in research; 42 participated in the workshops and 231 completed both rounds of the Delphi survey.

Results Critical consensus was reached across the different stakeholder groups that the public: have unique knowledge and expertise complimentary to that of researchers which should be valued equally (90%); can influence how research is used (85%); should be involved in publicly funded research, and are entitled to say what and how research is undertaken (80%); and that the public perspective, when considered at all stages, ensures more ethical research (70%). However, it was also agreed that assessing how PI influences a research project is methodologically problematic. Qualitative analysis highlighted the multi-dimensional nature of ‘PI process’ considerations. Researchers were sometimes seen as lacking understanding of the ‘added value’ of PI. The current PI Policy imperative was perceived by many as encouraging ‘tokenism’. Participants felt that a ‘fit for purpose’ approach, reflecting study and social contexts, was a more appropriate way of assuring effective PI than dependence on pre-conceived models.

Conclusion Host organisations, including academic institutions, need to strengthen infrastructure to ensure that systems promote and facilitate PI better.