Providing trial results to participants in phase III pragmatic effectiveness RCTs: a scoping review

Background There is an ethical imperative to offer the results of trials to those who participated. Existing research highlights that less than a third of trials do so, despite the desire of participants to receive the results of the trials they participated in. This scoping review aimed to identify, collate, and describe the available evidence relating to any aspect of disseminating trial results to participants. Methods A scoping review was conducted employing a search of key databases (MEDLINE, EMBASE, PsycINFO, and the Cumulative Index to Nursing & Allied Health Literature (CINAHL) from January 2008 to August 2019) to identify studies that had explored any aspect of disseminating results to trial participants. The search strategy was based on that of a linked existing review. The evidence identified describes the characteristics of included studies using narrative description informed by analysis of relevant data using descriptive statistics. Results Thirty-three eligible studies, including 12,700 participants (which included patients, health care professionals, trial teams), were identified and included. Reporting of participant characteristics (age, gender, ethnicity) across the studies was poor. The majority of studies investigated dissemination of aggregate trial results. The most frequently reported mode of disseminating of results was postal. Overall, the results report that participants evaluated receipt of trial results positively, with reported benefits including improved communication, demonstration of appreciation, improved retention, and engagement in future research. However, there were also some concerns about how well the dissemination was resourced and done, worries about emotional effects on participants especially when reporting unfavourable results, and frustration about the delay between the end of the trial and receipt of results. Conclusions This scoping review has highlighted that few high-quality evaluative studies have been conducted that can provide evidence on the best ways to deliver results to trial participants. There have been relatively few qualitative studies that explore perspectives from diverse populations, and those that have been conducted are limited to a handful of clinical areas. The learning from these studies can be used as a platform for further research and to consider some core guiding principles of the opportunities and challenges when disseminating trial results to those who participated. Supplementary Information The online version contains supplementary material available at 10.1186/s13063-021-05300-x.


Background
Reporting of trial results to those who participated is a fundamental requirement of ethical trials, underscored in the 2013 version of the Declaration of Helsinki which states 'Researchers, authors, sponsors, editors and publishers all have ethical obligations with regard to the publication and dissemination of the results of research.' and 'All medical research subjects should be given the option of being informed about the general outcome and results of the study.' [1]. In a recent survey of authors of clinical trials, 27% reported having disseminated results to participants, 13% planned to do so, but 33% had no intention of communicating the results to their participants (with a further 10% stating they were unsure, and the final 17% indicating 'other' or not answering) [2]. Other research indicates that, of those who do intend to share results, often this is operationalised passively, puts the onus on participants to request or access information and provides information in forms that may be difficult to access or understand (e.g. scientific publication) [3]. Clinical trial registers, such as Clinical-Trials.gov, indicate that there are hundreds of thousands of trials currently registered (https:// clinicaltrials.gov/ct2/resources/trends). If only 27% of these registered trials are returning results to participants, this is a significant problem. Most research participants want to be informed about the results of the study they participated in [4]. These figures suggest poor practice that requires urgent attention.
It is clear that the research community need to know how to do this better. A range of barriers have been reported to impact on trial teams' abilities to disseminate results to participants [2]. These include concerns that patients do not want or will not understand results, uncertainty about what results to share and with whom and when, difficulty reaching patients, lack of early planning and support, lack of academic expectation or incentive and concerns related to researchers lack of experience [2]. Several changes have been proposed in order to improve rates of sharing of trial findings with participants and/or improve how the findings are shared but a multi-factorial multistakeholder approach will be required to make this common practice [2]. One of the key considerations to facilitate dissemination is the provision of evidence based practical guidance on when, what and how to share results with trial participants for which there have been several calls [2,5].
A previous review of how results have been disseminated to research participants has been published, but is now several years out of date [4]. Furthermore, the review considered research more generally and did not focus on particular study designs and how these might influence the what, when and how of dissemination [4]. It is likely that different research designs may vary in the challenges and opportunities for providing results to those who participated for example, in some trials, people do not know what treatment they received, which raises questions about whether to reveal that information or not. Clinical trials, and in particular phase III pragmatic effectiveness randomised controlled trials (RCTs) are an important place to start because they usually enrol the largest number of trial participants (due to them being confirmatory trials), and therefore, any recommendations developed will have the widest potential reach. Detailed understanding of one major study design used in a diverse research context will provide a baseline platform on which to generate good practice that can be applied to other settings. In addition, given the current spotlight on research transparency, it is critical that we provide up to date evidence summaries that can inform the development of dissemination of trial results to participants [6].
As a first step to generate evidence-based recommendations for triallists to implement the dissemination of results to trial participants, there is a need to collate the evidence on what research has been done to date. The purpose of the evidence review outlined in this manuscript was to identify the breadth of research, not determine which method is 'best'; thus, a scoping review was conducted. Scoping reviews aim to 'systematically map the literature available on a topic, identifying key concepts, theories, sources of evidence and gaps in the research' [7]. They are useful to explore breadth and depth across heterogeneous literature [7,8].
This scoping review aimed to identify, collate and describe the available evidence relating to any aspect of disseminating trial results to participants both in terms of study characteristics and key features of the dissemination activity (namely what information to communicate, how to communicate it and reported advantages/disadvantages). The review had two objectives: 1. To develop an overview that identifies and characterises published research studies that have investigated any aspect of disseminating results to participants of phase III RCTs. 2. To identify evidence gaps where replication of evaluations or initiation of new research could be of value and to provide recommendations to directly inform research linked to this review as part of a programme of funded work (RECAP: Reporting Clinical trial results Appropriately to Participants [9]) to develop evidence-based recommendations that are attentive to diverse participants' experiences and preferences.

Methods
The work reported in this review relates to phase 1 of RECAP and specifically the identification of methods used to disseminate results to trial participants and reporting the key features of such methods. This scoping review was conducted and reported in accordance with the relevant items for scoping reviews specified in the Preferred Reporting Items for Systematic Reviews and Meta-analysis extension for Scoping Reviews (PRISMA -ScR) checklist (See Supplementary Table 1).

Search strategy
The search strategy was developed in consultation with a Senior Information Scientist and KG and informed by from a previous review in this area [4].  Shalowitz and Miller aimed to report on the trends in practice with regard to sharing results in terms of content, and stakeholder attitudes and so was deemed similar enough in scope to warrant it being used as a platform to identify studies published pre-2008 [4]. Studies identified in this previously published review that were relevant to phase III trials were identified from the reference list and included for data extraction. A full-search strategy is available in Appendix. Conference abstracts were included in the search and citation searching of systematic reviews identified was also conducted. We also searched the Studies Within Trials (SWAT) repository for ongoing studies. We did not contact relevant authors.

Inclusion and exclusion criteria
Reports eligible for inclusion included protocols, systematic reviews with/without meta-analysis, reports of RCTs, quantitative or qualitative studies and reports describing the process of results provision (both aggregate and individual) to trial participants. Reports known to the authors but not identified by the search (n=10, including some of the papers identified in the previous pre-2008 review which were relevant for trials [4]) were also included in the pool of potentially eligible studies and assessed using the inclusion and exclusion criteria. Included studies had to report data from, or information about, trials that recruited adult participants (aged 16 years and older), but these participants could be any trial stakeholder (e.g. REC members, trialists, funders, sponsors, representatives from industry, members of the public and/or current or previous participants in trials). Studies had to meet minimum eligibility criteria for inclusion. These criteria were that the study had to be about provision of results within the context of phase III pragmatic effectiveness trials in non-emergency settings. If the phase of trial was not clear the studies were still included, however, if studies included multiple phases of trial and it was not clear which data related to which phase the study was excluded. The decision to exclude results in emergency settings related to the role of proxies in the consent process and potentially the same individuals receiving results for trials in this setting. We believed this setting may influence what information is/ has been shared and decided to focus on trials in nonemergency settings as a starting point. Any interventions relating to dissemination of results were considered eligible as were any outcome assessments of effectiveness or indeed qualitative findings relevant to results dissemination. Studies were excluded based on the following criteria: papers reporting on provision of results in nontrial research; reports relating to phase I, II or IV trials specifically; reports using hypothetical trial scenarios; and non-empirical articles (e.g. commentaries).

Eligibility of studies
Titles and abstracts identified in the search were independently assessed by one reviewer (HB) with other reviewers (AS, SC, KG) double screening the search output. Any disagreement between abstract screeners on the eligibility of included papers was resolved through discussion. Full-text papers were obtained where applicable for those studies that on initial screening were considered potentially relevant and were further assessed for inclusion by one reviewer (HB) with queries resolved through discussion (KG). Posters were sought for relevant conference abstracts, etc. Papers pre-2008 were identified from reference [4] and the eligibility criteria applied and assessed as above.

Data extraction
A data extraction form was developed by the study team in advance of data collection. It captured information on study characteristics (e.g. population, setting, study design, participant characteristics) and data directly relevant to provision of trial results: description of the intervention disseminating results in relevant studies, description of results provided (i.e. aggregate vs individual results), mode, content, how content was developed/decided, description of PPI involvement, timing (e.g. before or after publication of results), who delivered trial results, any outcomes reported (i.e. of the evaluation of the results intervention), and reported advantages and disadvantages of results provision. The data extraction form was piloted on three papers. Data from all included studies were extracted by one reviewer (EC) with a random 40% of studies double data extracted (HB, KI, SC and

Data analysis
Data from the included studies were analysed using descriptive statistics with overall findings presented using narrative summary. As this was a scoping review, no formal critical appraisal of the quality of included studies was conducted, which is in line with published guidance for scoping reviews [7,8].

Search results
After removing duplicates, the searches run from January 2008 to August 2019 identified a total of 2085 studies as potentially eligible and included for further assessment and inclusion in the review. The majority of these reports (n=2005) were excluded due to not meeting the minimum eligibility criteria, providing 80 papers for a full-text review. Of these, a further 47 were excluded due to using hypothetical trial scenarios (n=2), not recruiting adults (n=5) not being phase III trials (n= 11), not being empirical studies (n=15), or not being about results dissemination (n=14) (Fig. 1). The remaining 33 reports were deemed eligible for inclusion in the review and progressed to data extraction. The 33 included papers reported studies from 27 trials, with the majority of studies (n=24) linked to separate trials . See further detials of characteristics of included studies in Table 1.

Characteristics of included studies
The 33 studies we identified for inclusion were published between 1987 and 2019 with 70% (n=23) published between 2010 and 2019and five pre-2008 studies being identified from the previous review [4] (Fig. 2).
Most studies (n=27, 82%) were conducted in single countries, namely the UK (n=11). The trial settings were situated across a broad range of clinical specialties and health areas but tended to focus more in secondary care settings and most frequently within oncology (n=11, 33%) (Fig. 3).
The median number of participants included in the eligible studies was 123 (range 10-3516) with a cumulative total of 12,700 participants across the 33 included studies. Of those that reported gender, ten included women only, two included men only, and eight included both men and women. Variability in reporting was also true for age which (where available) was most often reported as a median or mean, with the age of participants across reporting studies spanning from 18 to 80. Of the 33 included studies, nine (27%) reported on the ethnicity of study participants. The majority of participants in these studies were white, ranging from 83 to 98% (median 92%). Participants were from several trial stakeholder groups and included patients, trial participants (current or previous), health care professionals, sponsors, clinical trial unit directors, community partners, caregivers, and trial investigators. The review was very broad in its inclusion criteria with regard to study design and this is reflected in the types of studies included, some of which included mixed methods studies with various components. The methods each included study used to investigate aspects of dissemination of trial results varied, but of note only two were RCTs (Table 2). Only four out of the 33 studies reported any patient or public involvement (PPI) in the development of results materials and related procedures (Fig. 4). Involvement tended to take the form of PPI partners contributing to the writing of, or approving, the summary leaflet of results (see Table 2).

Key features of the trial result dissemination activity
The majority of studies (n=19) were investigating the dissemination of aggregate (whole trial) level results with two studies considering provision of individual results and a further one considering both individual and aggregate (Fig. 5).
With regard to the content of the results shared with participants, some studies gave no or little detail, stating 'lay summary of clinical trial results' whereas others provided detailed section headings for the content (such as the aim of the trial, how it was designed, and key design features of placebo and randomisation) and some reports provided direct examples of the feedback provided. Several study reports indicated that the trial abstract or trial registry entry had been rewritten in 'lay' language but there were also examples of how published trial findings had been adapted for participants (e.g. absolute frequencies presented rather than hazard ratios). There was also variability across the included studies with regard to the description of how the information provided to trial participants was developed, with 11 studies reporting directly on the development of dissemination information ( Table 2). For example, some studies reported working with a medical journalist to write the results leaflet, others reported adapting the technical summary into lay language (with and without patient input) and other studies conducted empirical research (using focus groups or surveys) to determine agreement on language used. Most included studies reported timing of provision of results as after peer-reviewed publication of trial results (n=7), or at publication (n=5), with other studies reporting after 'first public disclosure' (n=1), and several not specifying ( Table 2). Only a handful of studies reported on what the response (i.e. if an evaluation questionnaire was issued about the results) or uptake to the offer of results (i.e. requested to have access to results) was, and the uptake or response ranged from 5 to 78% (Table 2). Several modes of dissemination were reported across the included studies (Fig. 6). The most frequently used mode was postal (n=10), followed by online (n=3), faceto-face (n=1) and press release (n=1). Six studies reporting multiple methods (Fig. 6).

Summary of main results or findings from included studies
Only seven studies explicitly reported on an intervention to improve dissemination being investigated or evaluated in their study. In two studies, participants were randomly allocated to an intervention (e.g. letter detailing the results or letter providing access to study results on website) versus a control (e.g. study results press release or no results letter). In the remaining five, participants received the trial results (through media release, phone call, thank you card, weblink) or were asked to assess a results 'prototype before assessments of 'effectiveness' were made.
Outcomes reported included satisfaction with communication, preferences for information (content, presentation and length), understanding, whether participants discussed with others, trust in medical researchers, anxiety, guilt, anger, relief and whether results were helpful. The two RCTs comparing different methods for disseminating results measured different outcomes, trust or understanding. No studies explicitly assessed the acceptability to trial participants of receiving results from the trial they had participated in. However, there were reports of studies asking participants their preferences for receiving trial results (i.e. whether or not results were provided, or what mode of delivery).
Many positive features were reported by included studies with regard to dissemination of trial results to participants (see Table 3). Improvements in communication between health care teams and patients, and improvements in overall quality of care and satisfaction for patients were reported as benefits of dissemination of results to trial participants [2,12,22]. Demonstrating appreciation to participants for their contribution and increased accountability of researchers were cited as advantages [2,12]. One study mentioned that dissemination of trial results may facilitate trial retention [26] and another cited raising public awareness of the importance of research [36]. The potential to motivate others to participate in future research was also identified as a benefit of disseminating trial results [2,25,27].
On the whole, when participants had received results this had been largely viewed as positive [14-16, 21, 25, 39]. Participants in some of the included studies identified preferences for receipt of results such as a preferred mode and length but recognised that a one-size-fits-all approach may not be appropriate (e.g. if population has a range of age groups) [16,20,22,23,25]. Some studies also reported improvements in understanding of trial results (based on pre and post test) once participants had been provided with the trial results [22,39]. Participants also reported a feeling of 'pleasure' in what they viewed as contributing to a successful trial and that their contribution had been worthwhile and valued [2,22,25,26,36,38]. This extended into one study that noted that participants felt it important for their contribution to be explicitly recognised through receipt of a thank you         [22]. Participants (in this case in a randomised evaluation of trial results in an ongoing trial) reported that results were easier to discuss if they were perceived as 'positive' and whilst 'negative' results were discussed this was to obtain reassurance [23]. Unblinding also did not seem to be a concern for some participants (in a trial of nutritional supplement compared to placebo), and they were not discouraged to continue if they were in the placebo arm [27]. Some studies identified a range of disadvantages, or cautionary considerations, for trial participants and/or health care professionals when disseminating results to trial participants (see Table 3). Health care professionals or researchers discussed concerns such as the potential extra costs both in terms of resource and time (theirs and others) to ensure the provision of results was done well [11,13,25]. They also expressed concerns about not wanting to share unfavourable results with participants as this would require further explanation about research being for the benefit of future patients rather than for those who participate [12]. Linked to this there were worries about the emotional effect of results, and  Eleven studies were not included as they do not report dissemination of results but rather explore attitudes towards, wishes for, or practice of dissemination of results. Aggregate, refers to whole trial level results. Individual, refers to individual participant level results participants' understanding (both in terms of comprehension but also language barriers) [2,12,19,28]. Health care professionals in one study stated that an obligation to provide results to participants would make them less likely to enrol patients [12]. Some researchers stated that participants did not want to receive results, and assumed if they did, they would ask for them [2].
Some of the disadvantages reported by trial participants were similar to health care professionals with impacts on their emotions either directly linked to trial results or linked to their initial recruitment into the trial due to a difficult time in their lives [14,36]. When considering specific aspects of dissemination activity, some trial participants felt frustration at receiving aggregate results but not knowing individual results. In one case, some participants involved in a trial in pregnancy questioned their decision to consent to the trial and felt betrayed and angry towards trial and clinical staff when they had received results which they felt ill-informed to consider and which suggested mismatched expectations [33,35]. Trial participants also raised frustrations at the delay between end of the trial and contact with the results. 'Negative results' or receiving bad news were cited as concerns by some [24]. Whilst not raised as a direct concern by participants, some of the included studies did identify areas where confusion remained amongst participants after receiving results [22,25]. Logistical challenges linked to timelines of reporting but also ensuring participants were still alive and contact details were current were also cited as challenges [21,22,27].

Discussion
This review is the first systematic mapping of the evidence to describe the characteristics of studies that have evaluated or explored dissemination of trial results to participants in phase III pragmatic effectiveness trials. We identified eligible studies published over the past 33 years, including a range of study designs, and set across a range of clinical areas and specialities. Overall, the scoping review has identified the largely ad hoc approach to research in this area, focussing on discrete activities required for particular trials; however, more recent studies are taking a broader approach to assessing the problem and considering solutions such as the recent survey by Schroter et al. [2].
There are several gaps highlighted by the evidence map both in terms of areas for replication and those for initiation of new research. Firstly, with regard to replication, we identified only two trials (or SWATs) of results dissemination methods, both assessing different interventions and both measuring different outcomes. A programmatic approach to replicating interventions (that assess aspects of content and mode of delivery and are specified using TidIER) and measuring the same outcomes (identified as important to a range of stakeholders) is required if these types of studies are to contribute meaningfully to the evidence base [42]. With regard to new research, identifying what aspects of information should generally be considered core for inclusion in trial result summaries for participants would be helpful. Similar studies identifying what information potential trial participants and research nurses consider core for participant information leaflets have provided helpful summaries for consideration when designing patientfacing trial information; this could be extended to trial results [43]. In addition, better understanding of the barriers and enablers for trial teams in their ability and intention to disseminate results could help to provide  Receiving a results leaflet through the mail satisfactory and preferable to personal contact to enable study at length and in private. Most of the comments on the content and format of the leaflet were positive. Half expressed feelings of pleasure on receiving the leaflet, particularly at what they saw as the success of the trial, or felt that taking part had been worthwhile.
One negative consequence of receiving the results was that for some women it revived memories of a difficult time.
Avins, 2008 [15] Responses to provision of results largely positive and all respondents thought that future studies should include in person meetings.
Not specified Dorsey, 2008 [16] Media release from the investigators within a day after a sponsor-issued press release and a subsequent telephone call from the site staff to the participants; and conference call for research participants 2 weeks after the results were released.
Source and timing for learning study results and satisfaction with their communication.
Study participants reported high satisfaction with the telephone call and conference call but relatively low levels with the sponsor's press release.
Not specified Johnson, 2008 [17] N/A N/A N/A N/A Darbyshire, 2009 [18] Not specified Concern that participants were unwilling to discuss their diabetes and treatment The majority of participants were happy with the method by which they received their results and the same proportion were pleased that they were informed.
A small proportion of patients indicated that they were upset by the results.   Women were able to reconcile their original decision to participate in the trial but there were several concerns.
Return of results led participants to question the basis of their decision to consent to the trial. Some were shocked at the outcomes of the research. They were distressed by the discovery that by taking part in the trial they had exposed their child to a possible risk of harm. This was associated with guilt, anger and a sense of betrayal by the maternity staff and researchers involved in the trial. Others experienced a profound breach of trust. They questioned the motives and actions of those involved in the research, feeling  [2,3]. Attention also needs to be given to determining what constitutes an appropriate approach to dissemination as this may vary across trials involving different populations, assessing different kinds of intervention for different kinds of health problem, and perhaps with different types of findings. When considering who the results are for, the lack of ethnic diversity from those studies that reported their sample characteristics should be noted. Also, the lack of reporting of other characteristics, such as religion or sexuality, for which the reporting of results might need to consider and be sensitive to. Irrespective, ensuring trialists communicate results in culturally sensitive ways is also a key consideration going forwards. Another important area for investigation is how dissemination of results may need to be different in low-and middleincome countries, none of which were identified in our review, and adapted to a range of cultural contexts. There are some examples of this happening in African countries, with dissemination of findings through community events with song and dance [44,45]. This has N/A N/A Benefits of disseminating results to patients included supporting the spread of knowledge in the patient community, increased accountability for researchers, and an opportunity to empower patients. The potential to motivate people to participate in future research studies was noted by many and some suggested dissemination might encourage patients to consider interventions which could lead to better outcomes for them. It was also suggested that it might improve the doctor-patient relationship through building confidence and trust. Respondents further suggested that the impact of dissemination could be extended by giving patients the opportunity to share results within their own communities; there were mixed views on whether dissemination should be mandatory.
Some researchers said they do not think patients would be interested in receiving study results; others assumed not asking for them represented a lack of interest in getting them. Many were concerned that patients would lack the ability to understand the results and their implications.
N/A Not applicable as included study did not disseminate results to participants as part of study design particular salience given a 2018 research prioritisation exercise on methodological research for global health trials identified 'Methods of dissemination of findings' as number 7 (out of 27) and far above recruitment and retention (at 18 and 17, respectively) [46]. The results of this review also identify that only a minority of studies conducted in this area explicitly stated that they actively included patients or the public as partners in the development of the dissemination of results. Previous research has shown that amongst trial teams who intend to disseminate results to trial participants, most (60%) intend to involve patients and the public in the dissemination activity [47]. It is likely that this involvement happens in practice much more than is reported in the literature on results dissemination. However, a failure to share that knowledge on what worked well and what could be improved contributes to research waste.
Whilst the majority of studies reported that study participants viewed disseminating results as a positive endeavour that should be encouraged, the notion that it can be done poorly, and have poor consequences, including some that should be recognised as harms was also identified across several studies. Concerns were raised, for example about the difficult experiences of finding out that a trial had 'negative' findings, or that as a participant they had received an inferior treatment, or that providing results reminded them of a difficult time in their lives, or further, that they regretted their decision to participate in the trial in the first place. Ensuring trial teams have safeguards in place to protect against these unintended consequences of disseminating results is critical to ensure that the action has the intended effect. Considering early on with patient/public partners what to share, when and how, could help to address some of these concerns.

Strengths and limitations
This scoping review benefited from systematic methods which were guided by the general principles of the Cochrane handbook for systematic reviews intended to increase rigour and reduce bias [48]. As such, one of the key strengths of this review is the methods employed through each stage of study identification and extraction. In addition, the review identified a range of study designs and perspectives on dissemination of results and allows some level of aggregative summary to be synthesised. We also consulted with a range of relevant stakeholders (patients, methodologists, bioethicists, policy makers) via our advisory group who were key in helping ensure the scope of the review was relevant for current practice.
One of the limitations of the search was that we were limited to English language studies only and the search could have been enhanced through contacting known authors, etc. Therefore, there could be studies published in languages other than English that we have missed. Other limitations of our review are linked directly to the lack of complete reporting by the included studies. For example, it was not clear in all studies to explore perceptions of receiving results whether participants who were involved were only those who had opted to receive them in the first place. It would also be important in future research to explore the views of those who opted not to receive trial results. Finally, the review is likely subject to reporting bias in that many trials will report results back to participants and may have investigated that process but have not published and as such cannot be included in the review.

Conclusions
This scoping review has identified several studies that have explored the 'what' and 'how' of disseminating results of trials to those who participated. However, few high-quality evaluative studies have been conducted that can provide evidence on the best ways to deliver this activity. Indeed, the literature also shows heterogeneity around what outcomes are measured and reported and further still, are important to, relevant stakeholders. Identifying these and conducting further replications of existing evaluations would add significantly to the evidence base. Whilst there are some very in depth qualitative studies focussing on dissemination of results, these have tended to be conducted in a few discrete clinical areas and as such, opportunities to extend this work into other areas and to consider how findings and recommendations do and do not generalise across different trial contexts should also be considered. The learning from these studies can be used as a platform for further research and to consider some core guiding principles of the opportunities and challenges when disseminating trial results to those who participated.