This study has detailed, for the first time, how recruiters deal with patients’ treatment preferences in three different RCTs, including one where recruiters were trained and supported to elicit and explore treatment preferences. The means by which preferences were elicited did not differ greatly between the three trials; there was either a simple indirect open styled question that revealed the patient’s preference, or a preference was voiced unprompted after the introduction of a treatment. There were, however, very clear differences in how recruiters responded to the expressed preferences. In the ProtecT RCT (trial 3), trained recruiters explored the preferences and enabled participants to become clearer about whether their views were robust enough to be sustained or were sufficiently weak that participation in the RCT became possible . Most were open to randomization following exploration of their preferences. In the other two trials, preferences were often accepted at face value without further discussion and exploration of the underlying rationale for the treatment choice. No participants who expressed a preference in these trials were randomized. This study has illustrated how key techniques, provided as part of a training package , have been implemented by trial recruiters to elicit and explore patients’ treatment preferences in recruitment discussions to improve levels of informed consent and consideration of trial participation. Such techniques could be further evaluated in a comparative study and form the basis for training and empowering recruiters to explore preferences in future trials.
A recent systematic review of the recruitment activity of clinicians across a variety of RCTs highlighted the need for training, concluding that understanding and communicating RCT methods was a priority for future interventions to improve recruitment . Findings from a survey and workshop of UK Clinical Research Collaboration registered clinical trial unit directors confirmed these findings, identifying methods for improving recruitment and in particular training for site staff as the highest priority for trials methodology research . Although trial recruiters perceive patients’ treatment preferences as a barrier to recruitment [6–9], a recent synthesis of the perspective of 72 recruiters from six RCTs revealed clear obstacles and hidden challenges relating to their dual roles as clinicians and recruiters that led to their discomfort in approaching and discussing trials with patients [9, 34]. The study found that recruiters were more likely to accept patients’ treatment preferences at face value without further exploration rather than offer recruitment to the RCT if that preference accorded with their own views, and highlighted the need for recruiter training and support, particularly in the management of treatment preferences . Findings from the current study support these conclusions and offer suggestions in how to better manage treatment preferences.
In the past, it has been assumed that preferences make recruitment difficult , and that ‘challenging’ preferences may be considered to be coercive . This research shows that recruiters can be trained to elicit and address patients’ treatment preferences during RCT recruitment appointments, and that they do this more often than those without this specific training, and that this can lead to an increase in the numbers of patients who then consider and accept recruitment to the RCT [14, 15, 19]. The justification for exploring patients’ treatment preferences at trial recruitment should be to gauge their level of understanding of all treatment options to ensure that their decision whether to participate or not in the trial is well-informed. This is the ethical basis for preference exploration and this gauging of understanding should occur even if the patient enters the recruitment appointment in a position of equipoise, as they may have a weak or misinformed basis for their position.
The process of recruitment is a socially delicate, interactional undertaking, with subtle moments in appointments influencing participants’ understandings and willingness to participate . Despite this, very little research has focused on the content of the interactions that lead to informed consent and RCT participation. A small number of studies have shown that doctors can struggle to explain RCTs clearly [38, 39], that shared decision-making is not always practiced , that some participants take part because of altruism , and some seem unaware that they are involved in research . Recordings of recruitment appointments can permit an open debate about whether the exploration of patients’ treatment preferences should be considered coercive or an informative part of gaining fully informed consent, and where a line might be drawn. We did not assess the impact of preference exploration from the patients’ perspective because the focus of our study was on recruitment appointments, and so it will be important to include patients’ perspectives of recruitment and preference exploration in future research. These issues are complex and will need to balance formal ethical guidance with increasing demands by patients to be fully informed.
One of the challenges with training recruiters to communicate the concept of clinical equipoise and to spend time understanding patients’ preferences is the impact on the time taken in consultations in which consent to participate is sought. In this study, appointments conducted by research nurses in trial 3 (ProtecT), who were supported to explore patients’ preferences, were longer (56 minutes on average) than those conducted by surgeons and the research nurse in trial 1 (16 minutes), although similar to the combined appointments with oncologists and surgeons in trial 2 (64 minutes). The difference in time between trials 1 and 3 may be explained by the simpler nature of trial 1’s interventions for a non-life threatening condition, but may also be due to the extra time taken to understand patients’ treatment preferences in trial 3. Whilst longer consultations may lead to increased costs, the costs of trial extensions, closures, or trials never being undertaken because it is thought to be too difficult to recruit need to be considered and compared with the costs of training clinicians to recruit effectively, a skill which is likely to be transferable between RCTs. Consideration should also be given to who is best placed to recruit trial participants. A nested RCT in the ProtecT feasibility study showed that trained research nurses were as effective as doctors at recruiting to the trial and more cost-effective , but nurses can also experience difficulty with recruitment [9, 37]. It has yet to be formally assessed if doctors can be trained to explore treatment preferences as consistently as the trained research nurses in the ProtecT study.
Although the ways recruiters responded to patients' preferences were different between trial 3 (ProtecT) and the other trials, it is possible that the observed differences were related to factors other than the specific training. There were, for example, differences across the trials in terms of the seriousness of the condition, the risk profile of the treatments, and the backgrounds and numbers of recruiters. Trial 1 offered the least invasive treatments with relatively low-risk profiles for a non-life threatening condition, with one research nurse doing most of the actual recruiting. In this trial, preferences were readily accepted without exploration which may be a reflection of the relative ‘mildness’ of the condition and low-risk profile of both treatment options. However, it may also reflect the previously identified discomfort and difficulties that nurses may have with trial recruitment (and in particular treatment preferences) in relation to their perceived roles as a caring clinical nurse . Trial 2, in contrast, offered recruitment by surgeons and oncologists to a trial with more complex treatments for a potentially life-threatening condition, with diverse and potentially debilitating side effects. Despite these differences with trial 1, a sizeable proportion of recruiters also showed a similar ready acceptance of preferences, and those that did not accept immediately did not explore the underlying rationale for the treatment choice, missing opportunities to promote more informed decision-making. These findings were in contrast to the research nurses who were trained in trial 3 to sensitively explore treatment preferences.
It must also be acknowledged that the RCTs compared here might not be representative of RCTs in general and that they were a ‘convenience’ sample. Trial 1 sought external help with recruitment after experiencing difficulties, and trial 2 included an integrated qualitative study of recruitment from the outset, but the data presented here were collected before any training on patients’ treatment preferences had been given in these trials. The issue of generalization or transferability of findings should be considered in light of the number of audio-recordings analyzed, particularly in the comparator trials. The number of available audio-recordings in trial 1 was relatively small. This trial also relied on data from only one study centre and with the same research nurse following up on patients to determine their willingness to be randomized. These issues suggest that caution in interpretation of findings may be warranted, however, similar findings were observed in the other comparator trial (trial 2) where a larger number of consultations were analyzed from a variety of clinicians across two study centers. As most trials do not routinely audio-record their trial recruitment consultations, access to such data is limited.
A strength of the study was the range of qualitative analyses undertaken on ProtecT trial data [9, 15, 17, 34, 37] and the availability of audio-recorded data of recruitment appointments across three different trials. These permitted the identification of key aspects that can lead to improved recruitment (more patients sufficiently informed to consider participation and higher recruitment rates). Methods for eliciting preferences, ascertaining and acknowledging their rationales, balancing participants’ views about treatment, empathy with the difficult situation of clinical equipoise, and retaining an open mind to all treatments, could be developed into training programs to be used in other trial contexts. The strategies need to be combined with simple communication techniques, such as using open questions, long pauses, and readily ceding the floor , along with detailed considerations of how to avoid coercion. Previous research suggests that training oncologists in effective communication styles and behaviors may positively impact upon RCT recruitment . The question remains of how much training recruiters should receive to effectively explore treatment preferences to facilitate recruitment and informed consent - whether simple encouragement is sufficient or whether more detailed training, monitoring, and feedback on communication skills and implementation of key strategies is required.
Findings here support suggestions now being made in usual clinical practice to diagnose patients’ treatment preferences and engage patients in discussion to improve informed decision-making . However, doctors are not routinely trained to address patients’ treatment preferences. Communication skills required by clinicians needed to explore preferences are not often the focus of consultations in everyday practice, therefore it is not surprising that clinicians are not routinely addressing preferences within the context of trial recruitment. The key techniques identified in this study to address treatment preferences could be applied outside of the trial context to usual clinical care.