By Janice T. Radak and Lael Stone, M.D.
While controversy continues to simmer over who is accountable for posting results on ClinicalTrials.gov, perhaps the “results debate” masks the more challenging problem: ClinicalTrials.gov is not working for its intended audience.
Recall that ClinicalTrials.gov was conceived as part of the Food and Drug Administration Modernization Act (FDAMA) of 1997 to inform the public about federally funded clinical trials that were recruiting, about to recruit, or had been completed and had posted results of their findings.1
In February 2000, the National Institutes of Health launched ClinicalTrials.gov as a “consumer-friendly” resource designed to benefit patients, their family members, healthcare professionals, researchers, and the public.2 "Through this new database, NIH offers up-to-date information on promising patient-oriented research on hundreds of diseases and conditions,” Acting NIH Director Ruth L. Kirschstein, M.D., is quoted as saying in the launch press release.
There were 4,000 trials registered at the launch. As of Sept. 14, 2018, the website listed 284,374 studies with locations in all 50 states and in 204 countries.1 The number of registered studies has grown consistently since inception, with 2018 averaging 2,316 studies registered per month (though this may fall short of 2017’s monthly average of 2,433). While these numbers serve as evidence of its status as the premier registry in researchers’ minds, its usefulness to its intended audience needs improvement.
Consumers Want Health-Related Information
That consumers want health information is well documented. In 2016, nearly 80 percent of internet users searched for health-related information (up from 62 percent in 2001), and 81 percent of those were willing to click on sponsored healthcare links.3 However, it’s also clear that consumers need information they can actually consume.
A CDC report from 2013 showed that adult literacy skills had changed little since 1992, with only 12 percent of U.S. adults showing high proficiency in literacy tasks, and only 9 percent showing proficiency in numeracy skills.4 As diseases and conditions such as diabetes, Alzheimer’s, and obesity continue to affect larger and larger numbers of citizens with limited health literacy, we see a major disconnect between what ClinicalTrials.gov was intended to be, what it is, and what more it could be.
What Was Intended
We imagine the founders originally envisioned a database of useful information presented in a consistent format written in plain language with a consumer-friendly search function. In other words, a product designed specifically for the end user who is a U.S. taxpayer or consumer.
What It Is
The content execution of any database is driven by the culture of those who input the information. For ClinicalTrials.gov, the content is input by the research teams, who appear to simply copy and paste from the formal protocol. What’s holding back today’s ClinicalTrials.gov, in our opinion, is that the information entered in the database is not written or presented for its true intended audience: the consumer, the critically ill patient (or their caregiver), the future trial participant, or even the community physician who has the power to refer.
As currently presented, the tilt is clearly technical, raising three concerns. The first represents a disconnect between how researchers and patients think. Take the inclusion and exclusion criteria: Researchers categorize patients by certain criteria (e.g., biomarker cut points) in order to answer the research questions they have developed. As such, inclusion and exclusion criteria are entered according to the research team order and level of detail (e.g., blood level counts). But patients don’t think like researchers; they think more broadly, such as by diagnosis, e.g., Sjogren’s syndrome or relapsing/remitting MS. So, when patients search the database, they get every Sjogren’s trial and are forced to review highly detailed criteria that they may or may not understand. This is not patient-centered.
This leads to the second concern: communication between the research world and community medicine. Clinical researchers appear to think that patients know more technical details about their disease than their community physicians may actually share. For the first concern, perhaps there’s a way to standardize these criteria and improve the current search function on ClinicalTrials.gov such that inclusion/exclusion criteria could be searched. For the second, using the health literacy lens would help both patients and community physicians better understand the research.
The third concern is the disconnect between the increasing complexity of protocol design, which is reducing the number of eligible patients, and consumers’ requirements for research participation in general. Several studies have noted that inclusion criteria have increased, thus limiting the number of patients who are eligible in the first place. The July/Aug. 2018 Impact Report from the Tufts Center for the Study of Drug Development5 reports that the number of endpoints rose 86 percent in additional procedures compared to earlier studies. Yet 53 percent of consumers surveyed for a Forte research report6 cited the location, number of study visits, and types of medical procedures required would all be factored into their decision whether to participate. Thus, more complex trial designs may actually be driving future participants away.
What It Could Be
Imagine ClinicalTrials.gov as a patient education tool. Using this lens, several changes present themselves instantly. For example:
Plain Language: Given the low level of health literacy in the general population — the group from which all trial participants will be drawn, language on the site should be geared (as much as possible) to an eighth-grade reading level, and rationale should be added where applicable. For example, in a trial of CoQ10 and D-ribose in patients with diastolic heart failure, the primary endpoint is:
Change in health status of patients with HFpEF [Time Frame: Change from Baseline to Week 12]—Patients’ perceptions of their symptoms will be measured using the Kansas City Cardiomyopathy Questionnaire (KCCQ), a self-administered questionnaire that measures patients' perceptions of five domains of their health status relevant to HFpEF. The KCCQ is a 24 item questionnaire. Scores are transformed to a range of 0-100, in which higher scores reflect better health status.
Why not make it plain language: Patients will be asked a set of questions about their health status and how they are feeling.
Consumer Summary Page: Consider a Consumer Summary as the opening page, with key inclusion criteria written in plain language and a Next Steps section that could include a link to a website with more information (in plain language) and a number to call to speak with someone. Such pages are not unheard of: many grants require a lay summary page that has the researcher translate their work for the lay audience. Adding a number to call to speak to a live person is obvious if you want volunteers.
Links to Definitions: This is research, and some things must remain technical. So, to help patients better understand, add the common question mark in a circle at the end of a line. Today’s internet-savvy consumers know this means click here for more information.
Make Contact Information Real: Make it easy to reach the local sites. Local site contacts were touted as a plus from the beginning. However, publishing that one of the trial locations is University of Alabama at Tuscaloosa is not helpful. More detailed contact data about each site is needed: local website, phone number or email with a name attached. And no, it’s not impossible. ResearchMatch, another NIH/NLM clinical translation project run by contract through Vanderbilt University is already doing it —www.ResearchMatch.org. The website, which appears to match volunteers to research projects, says it has 635 studies for which it is seeking volunteers — both those with a health condition and healthy volunteers. However, you must register and provide your email to browse the studies, meaning use is not anonymous, which many consider an invasion of privacy. Further, the website states that its content “is solely the responsibility of ResearchMatch and Vanderbilt University and does not necessarily represent the official views of the NIH,” and the website is branded at the bottom with “Powered by Vanderbilt,” calling into question who owns the information a user might enter into it.
What More It Could Be: A Recruiting Tool
Imagine ClinicalTrials.gov as the ultimate recruiting tool.
Consider the number of trials that fail due to lack of enrollment: 19 percent, according to the 2016 study of Forte benchmarks data.6 Significantly, in a global study of 12,000 individuals, 57 percent said they are very or somewhat willing to participate in a clinical research study. But there is a disconnect between consumers, the research community, and community physicians. A 2018 Landscape Report from the American Cancer Society Cancer Action Network notes that there is no trial available for 56 percent of cancer patients.7 Of the 44 percent of patients for whom there is a trial, 27 percent will be eligible, but only 8 percent will actually enroll. The reality, according to ACSCAN, is that many of those 27 percent who are eligible are not even asked to join the trial. The report is careful to note these are cancer trials and limited to cancer patients, but the difference between the number of people who are asked and the number who enroll is a serious challenge to the success of any trial. If the information on ClinicalTrials.gov was written with health literacy as a consideration, perhaps more patients could find a trial on their own. If community physicians knew they could reliably send patients to ClinicalTrials.gov for user-friendly information, they may be more willing to make the referral.
Results Matter — Both Positive And Negative
In the 2018 STAT investigation8 into the posting of results on ClinicalTrials.gov, the authors sought to shine a light on the discrepancies between what federal law requires and the reality of what actually appears in the public database. Thanks to STAT’s 2015 report,9 reporting of results has increased from 52 percent to 72 percent, but that means at least one-quarter of trials still are not reporting their results as required.
While many have suggested that more FDA regulation is required, there are already regulations in place that require reporting results. So what is missing is not more regulation, but rather the collective political will to honor the entire research process through to the end.
Everyone agrees that it is the principle investigator’s responsibility to post his/her research results. But PIs do not operate in a vacuum — they report to supervisors, lab managers, tenure committees, hospital medical boards, Institutional Review Boards, to name but a few. Accountability should be part of the institutional decisional milieu — whether it is academic or corporate
And what about a commitment to transparency within the research community? When results are not reported on ClinicalTrials.gov or elsewhere, those not in the in-group have no way of knowing if something worked. Reporting results democratizes the flow of information rather than keeping it in the hands of a limited few. This problem is compounded by journals’ past tendency not to publish negative results. Some journals are trying to address this bias; Neurology, for example, is planning a special supplement specifically to publish negative trial results.
Failure to post results wastes two precious resources: dollars and volunteers. The dollars will continue to be dispersed as the community judges. But the volunteers need a different kind of care and feeding. And this is where ClinicalTrials.gov is perhaps missing the biggest opportunity: a consumer-friendly version could be the ultimate marketing tool for recruitment. As four out of five consumers are already searching online for health information, why not make a premier registry for researchers a premier landing site for consumers activated to improve their health or the health of a loved one?
About The Authors:
Janice T. Radak is the former editor in chief of the journal Geriatrics and current editor of Lower Extremity Review. She has developed patient and professional medical education content for more than 30 years, including time with the American Diabetes Association and the Office of Disease Prevention and Health Promotion (Healthy People), and currently serves on an Advisory Panel for the Patient-Centered Outcomes Research Institute.
Lael Stone, M.D., is a neurologist specializing in multiple sclerosis at the Cleveland Clinic, with a particular interest in the use of MRI outcome measures in clinical trials. She has served various roles in multiple clinical trials, including site Principle Investigator. She is currently on medical leave.