Supporting participation in clinical research: decision aids for trial recruitment?

Much work has been done in recent decades to ensure that participation in clinical research is informed and voluntary, but assessments of the quality of participant information sheets, recruitment consultations and participants’ understanding of research studies continue to suggest there is scope for improvement.

In this issue of Health Expectations, Juraskova et al. ¹ report on a preliminary assessment of a decision aid for women invited to participate in a randomized controlled trial comparing two drugs for the prevention of breast cancer following treatment for ductal carcinoma in situ. We think this is one of the first published reports of a resource developed explicitly as a decision aid for people facing decisions about trial participation (some participant information sheets might ‘count’ as decision aids on some definitions, but they have generally not emphasized the comparison and choice between participation and non‐participation as options).

In Juraskova et al.’s study, the decision aid was judged favourably and deemed helpful by the women who reviewed it. It seemed to aid understanding without provoking anxiety. This suggests that decision aids for research participation decisions – we might call them ‘research recruitment decision aids’– merit further investigation.

Many of the issues that need to be addressed during the development and evaluation of research recruitment decision aids are similar to those that have arisen during the development and evaluation of treatment decision aids (aids for people facing decisions about treatment or risk management in non‐research contexts). However, research recruitment decision aids will require some additional considerations.

Developers of research recruitment decision aids will need to make judgements about how best to structure what may be a complex set of options: an invitation to participate in clinical research usually expands the option set available to people who may already be facing decisions about the treatment or health risk management. Juraskova et al. presented women with two main options for managing their risk of future breast cancer: standard care and a clinical trial. They noted that standard care could be risk management without or with tamoxifen, and that the clinical trial would be risk management on the IBIS‐II DCIS study with random assignment to tamoxifen or anastrozole (Appendix example a). They might also have considered alternatives such as (i) presenting risk management with and without pharmaceuticals as the two main options and noting that risk management with pharmaceuticals could be either tamoxifen without trial participation or random allocation to either tamoxifen or anastrozole within trial participation, or (ii) presenting three main options: risk management without tamoxifen, risk management with tamoxifen, and trial participation with random allocation to risk management with tamoxifen or anastrozole. With more elaborate trial designs, and in situations in which there are multiple treatment/management interventions and combinations in use, there may be many ways of presenting the option set. The choice between these is important because they give different relative priority to research participation as an option and because they are likely to influence people’s choices in different ways – something that will need to be investigated.

Developers of research recruitment decision aids will also need to consider whether and how to present the possible implications of the research (and of individuals’ choices about participation in it) in terms of knowledge production, progress in health care and the treatment of future patients. Many people consider the potential to contribute to knowledge and help improve health care for others to be important and legitimate reasons to participate in research (see the paper by Bakos et al. in this issue), ² and a case can be made for encouraging people to consider the social consequences of their participation choices. ³ However, the question of how this should be done has been somewhat neglected to date. There is usually significant uncertainty about whether, how and to what extent future generations will benefit from a proposed or ongoing study because the findings and the use that will be made of them are both to some extent unpredictable. The appropriateness of different ways of encouraging people to consider the potential to benefit others as they face research participation decisions thus warrants further study and debate. Questions also need to be asked about whether and how people should be given information and encouraged to consider the potential implications for science and others of their choosing not to participate in research. ⁴

Developers of research recruitment decision aids will also need to grapple with the ongoing debate about whether and how to help people to combine their own values and preferences with information about their various options in order to make a selection. The merits of decision analytic‐based calculations that generate preference‐personalized recommendations, and of values clarification exercises that encourage people to express, consider and integrate their views about issues pertinent to the decision, are both contested. ⁵ , ⁶ , ⁷ Their effects in health‐related contexts are still poorly understood, and the extent to which they are valued by decision makers is unclear (see the review of parental decision support needs by Jackson et al. in this issue). ⁸ Developers who plan to include either of these features in research recruitment decision aids will need to consider which aspects of the interventions under investigation, the research processes (e.g. additional consultations and clinical and questionnaire‐based follow‐up) and the intended outcomes of the research they should encourage people to focus on in the decision process. They need to accommodate the possibility that potential research participants might vary significantly in terms of the relative importance they attach to each of these.

As with treatment decision aids, the question of which criteria should be used to evaluate research recruitment decision aids is likely to be challenging. In terms of formative evaluation, Juraskova et al.’s work illustrates how even questions about how to present probabilistic information might be answered differently depending on whether the focus is on providing the format that is most liked or the format that is most readily understood. (Future research might usefully investigate whether people’s preferences change when they are given information about the proportion of people who understand information correctly with different formats.) Once research recruitment decision aids are ready to evaluate in larger‐scale trials, decisions will have to be made about whether to assess – and how to value – their impact on rates of participation, informed participation, informed participation that is somehow additionally assessed as being congruent with individuals’ values and other criteria such as health state outcomes.

If research recruitment decision aids become popular with researchers and research ethics committees, the numbers of people exposed to decision aids could rise quickly and significantly. If trial protocols start to specify that decision aids should be offered to potential participants, either as supplements to or replacements for standard participant information sheets, the kinds of barriers that have been identified to the routine use of treatment decision aids ⁹ , ¹⁰ may be quickly overcome. If this scenario is likely, then the issues outlined above should perhaps be considered with some urgency.