Communicating Clinical Trial Results to Research Participants FREE

Despite exposing themselves to risk, volunteers in clinical trials are not routinely informed of trial results.¹ Systematic efforts to communicate trial results to participants have been few,² and their absence has led to high-profile public concerns about trial results,³ especially those showing a lack of benefit.^{4 - 5} In general, federal regulations⁶ and institutional review boards⁷ require disclosure only of information that may affect ongoing participation in a research study but not disclosure of research results at a trial's conclusion.⁶ In addition, although sponsors are usually not required to inform research participants of results, commercial sponsors often must inform public investors,⁸ in part to limit insider trading concerns.⁹

In the absence of a legal mandate, investigators have recommended that providing both positive and negative clinical trial results be the “ethical norm.”¹ Despite this recommendation, efforts to communicate results to research participants remain quite limited.^{10 - 11} Some participants never learn study results,¹² although many would like to have that information.¹³

We recently completed an industry-sponsored, multicenter, double-blind, randomized, parallel-group, placebo-controlled trial of an ω-3 fatty acid (ethyl-eicosapentaenoic acid) in subjects with mild to moderate Huntington disease (TREND-HD) that consisted of a 6-month placebo-controlled phase followed by a 6-month open-label phase.¹⁴ During the study, the investigators and sponsor developed a communication plan to release the results of the placebo-controlled portion of the study to research participants in a timely and personalized manner. We then surveyed research participants to evaluate the plan's effectiveness based on the timing and method of learning the results and their satisfaction with different parts of the plan.

The communication plan, developed during the course of the trial and approved by the University of Rochester research subjects review board, had 3 principal elements. The first was a media release from the principal investigators posted on the Huntington Study Group's public Web site¹⁵ and e-mailed to members of the Huntington disease community. The second was a telephone call from site staff (the site coordinator or investigator) to research participants providing results and next steps for their participation. The third was a joint telephone conference call for the investigators, sponsor, and study participants to communicate study results. Participants in the clinical trial were informed in advance by postal mail of the approximate time (month) when study results would be released by the sponsor and were told that investigators would try to inform them of results as soon as they were publicly released.

To evaluate the plan's effectiveness, we developed a 13-question postal survey asking study participants to indicate when they first learned the study results, their information sources, and their satisfaction with those sources (on a 5-point Likert scale, where 1 indicates completely dissatisfied and 5, completely satisfied). They were also asked to rate how well they understood the study drug's reported risks and benefits and their next steps for study participation (on a separate 5-point Likert scale, where 1 indicates strongly disagree [with a statement that they understood] and 5, strongly agree). Finally, they were asked to indicate which statement better described their preference: “It is important to me that results of clinical studies undergo a thorough review period before I am informed, even though it may delay my learning the results of the study” or “It is important to me to learn the results of a clinical trial as soon as possible, even though it may prevent a thorough review.” Sites that submitted the survey to and received approval from their respective institutional review boards in a timely manner were included in the results.

We summarized survey responses using descriptive statistics. We compared baseline characteristics using Fisher exact tests for categorical outcomes and ttests for continuous outcomes for research participants who were and were not surveyed and for participants who did and did not respond to the survey. We used an exact binomial test to examine no preference in the timing of participants learning study results. Satisfaction with the communication of study results was analyzed by dichotomizing the 5-point Likert scale and modeling the probability of being satisfied or completely satisfied as a function of communication source. Given that some participants rated multiple communication sources, a logistic regression model was fitted using generalized estimating equations, assuming a compound symmetry correlation structure.¹⁶ All statistical tests were performed at the 2-sided significance level of 5%, and no correction was made for multiple testing.

From April 20 to 23, 2007, the sponsor and the Huntington Study Group's chief biostatistician and principal investigator (I.S.) performed a planned analysis of the 6-month placebo-controlled phase of the study, which did not show safety concerns or clinical benefit from the experimental treatment. At 8 AM(eastern time) on April 24, 2007, the study sponsor issued a press release making public these results. Later that same day, investigators received via e-mail a summary of the study's results that included results of the primary and secondary measures and principal safety findings. The investigators were asked to telephone study participants to inform them of the results and their next steps. The next steps for trial participants were to stop taking study medication and to schedule a closeout visit with their site within approximately 1 month. On April 25, 2007, the Huntington Study Group issued a separate media release indicating the preliminary results of the study. On May 8, 2007, the study's sponsors and investigators held a 90-minute joint telephone conference call and invited all study participants and their caregivers to take part. Site staff notified study participants of the conference call and provided dial-in instructions. Fifty-seven telephone lines were in use during the teleconference, which was principally aimed at study participants and their caregivers. Approximately half of the call was devoted to opening remarks from the principal investigator and sponsor and half to answering the questions of study participants (2 study participants asked questions) and their caregivers (5 caregivers asked questions).

Of the 41 sites (6 in Canada and 35 in the United States) participating in this clinical trial, 28 (5 in Canada and 23 in the United States) participated in the survey, which was completed between June and November 2007. These 28 sites accounted for 217 of the total trial population (68.7%). The baseline characteristics of the population surveyed were generally similar to the overall study population, with the exception of a greater occurrence of depression in the surveyed population (comparisons are given in the Table). Of the 217 study participants surveyed, 114 (52.5%) responded. Respondents were older, received a diagnosis at an older age, and had a shorter trinucleotide repeat length, less suicidal ideation, and better total functional capacity (Table).

Nearly half (46.3%) of the study participants reported learning the study results within 1 day of the sponsor's public release (Figure 1). The telephone call from site investigators was the most common initial (73.1%) (Figure 2) and overall (82.4%) (data not shown) source of information about study results. Traditional media outlets (eg, newspaper, television/radio, and general Web sites) were uncommon sources of information for this study (none was used as the initial source of information; newspaper and television/radio were each used by 10.0% of participants as a later source of information).

Study participants reported high levels of satisfaction or complete satisfaction with the telephone call (89.3%) and conference call (82.1%), but relatively low levels with the sponsor's press release (50.0%) (Figure 3) (P < .001 for the telephone call vs press release; P = .03 for the conference call vs press release). Most respondents reported agreement or strong agreement with having a good understanding of the study drug's risks (84.5%), the study drug's benefits (83.0%), and their next steps for participation in the study (78.0%). Most study participants (59.4%; P = .06 for testing equal preference) tended to prefer that results undergo a thorough review before being informed, even if the review delayed their learning of the results.

In this industry-sponsored clinical trial, research participants with Huntington disease expressed high levels of satisfaction with personalized, timely, and accurate communication of study results. The participants preferred customized communications rather than announcements aimed at the public or at company investors. Survey respondents also reported a good understanding of the study drug's risks and benefits as well as the next steps for concluding research participation. Whereas previous research has examined communicating research results by mail,^{17 - 18} most of the research participants in this study learned results through a telephone call from the site staff or a conference call with the study's investigators and sponsor; both were well received. The conference call vehicle and format are novel approaches that provide for interactive communication among the study participants, investigators, and sponsor.

A previous study involving breast cancer treatment shared the results with only those individuals who indicated a desire to learn them¹⁷ ; however, we sought to inform all research participants of our trial results. Among those surveyed in our clinical trial, few (<10%) indicated dissatisfaction with the communication of trial results by telephone or by teleconference. We did not disclose any results of genetic testing (eg, CAG repeat length), which was not the primary aim of the trial, to research participants.

The survey results are limited by a modest response rate, the use of a survey that has not been validated, clinical impairments in the study population, and a possible lack of generalizability to other disorders. Only 114 of the total 316 study participants (36.1%) responded to the survey, raising the question of response bias. Although respondents were slightly less affected by Huntington disease than were nonrespondents, the 99 participants who were not surveyed shared similar baseline clinical, demographic, and Huntington disease gene characteristics with the 217 study participants who were surveyed. The generally high levels of satisfaction with and understanding of the results of the study are limited to those who completed this survey. The results may not reflect the perceptions of those who were not surveyed or did not respond, who may have been less pleased about the communication of the trial results and therefore less willing to complete the survey. In addition, no formal test of the understanding of the risks, benefits, and next steps was administered to determine whether the self-perceptions of those who completed the survey were indeed accurate.

Huntington disease affects cognitive function and could have affected the survey responses, but baseline Mini-Mental State Examination scores of study participants were appreciably higher than those that typically indicate competence to consent to treatment.¹⁹ Participants were also permitted to have someone help them complete the survey. To the extent that Huntington disease, a hereditary neurodegenerative disorder, is unique, the results of this study may not be readily generalizable to other medical conditions. However, the results at least may be applicable to other neurodegenerative conditions that affect cognition (eg, Alzheimer disease and Parkinson disease).

Although efforts to protect human research participants have appropriately focused on the recruitment, enrollment, and active follow-up phases of randomized controlled trials, relatively little investigative attention or regulatory concern has been directed to the needs of study participants at the conclusion of trials. Consequently, the obligation to disseminate research findings to research participants often falls through the cracks between institutional review boards⁷ and investigators,^{10 - 11 ,20} who are collectively responsible for protecting and promoting the welfare of participants until a study is concluded. Communicating research results to participants remains an unmet need that would likely benefit from more planning and cooperation, from addressing the issue prospectively in the consent process, and from using a more evidence-based appraisal of clinical research practices. Our study suggests that addressing this need is feasible and highly valued by research participants.

Correspondence:E. Ray Dorsey, MD, MBA, Department of Neurology, University of Rochester Medical Center, 1351 Mt Hope Ave, Ste 223, Rochester, NY 14620 (ray.dorsey@ctcc.rochester.edu).

Accepted for Publication: May 28, 2008.

Author Contributions:Drs Dorsey and Beck had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Dorsey, Beck, Adams, Chadwick, de Blieck, McCallum, Stewart, Shoulson, and Huntington Study Group TREND-HD Investigators. Acquisition of data: McCallum, Deuel, and Huntington Study Group TREND-HD Investigators. Analysis and interpretation of data: Dorsey, Beck, Briner, Deuel, Clarke, Stewart, and Shoulson. Drafting of the manuscript: Dorsey, Beck, McCallum, Clarke, and Shoulson. Critical revision of the manuscript for important intellectual content: Dorsey, Beck, Adams, Chadwick, de Blieck, Briner, Deuel, Clarke, Stewart, Shoulson, and Huntington Study Group TREND-HD Investigators. Statistical analysis: Beck. Obtained funding: Clarke and Shoulson. Administrative, technical, and material support: Adams, Chadwick, de Blieck, McCallum, Briner, Clarke, Stewart, Shoulson, and Shinaman. Study supervision: de Blieck and Shoulson.

Financial Disclosure:None reported.

Funding/Support:The TREND-HD Study and the communications effort were supported by Amarin Neuroscience Ltd.

Role of the Sponsor:The sponsor helped formulate and conduct elements of the communication plan. The sponsor was not involved in the collection, management, or statistical analysis of the data. Dr Clarke and Mr Stewart helped interpret the data and provided critical review of the manuscript for important intellectual content.

Previous Presentations:This study was presented as an abstract at the Huntington Disease Clinical Research Symposium; December 1, 2007; Boston, Massachusetts; and the abstract was published in Neurotherapeutics. 2008;5(2):374-375.

Additional Contributions:The following individuals contributed to the TREND-HD study and this manuscript: W. Adams, A. Agarwal, J. Burkholder, J. N. Caviness, M. Diggin, M. Fite-Weatherford, J. Forsyth, J. Goldstein, C. Gray, P. King, M. L. Klimek, S. Kostyk, P. Lamberton, N. Lucarelli, M. Malikowski, E. Molho, R. Y. Moore, S. Peterson, C. Phillipson, C. Rini, D. Scott, A. Shadduck, A. Sheinberg, R. Vareha, and K. Wernette.