It’s a difficult truth of clinical trials: some participants can be less than honest.
Infractions range from concealing a history of smoking, to lying about following a drug regimen, to sharing study medications with others. Often, investigators are none the wiser — but Clete Kushida, MD, PhD, wants to change that, first by quantifying the problem.
The Stanford professor of psychiatry and behavioral sciences is the senior author of a new study, published by Contemporary Clinical Trials, that catalogs the prevalence and types of deception by participants in clinical trials and suggests ways for investigators to root it out.
“One of the most important things is to realize upfront that deceit in clinical trials is something that is there, that will be there, and to figure out ways to minimize it,” Kushida told me.
The researchers’ first discovery was that data on the topic are scarce, he said, but what they did find “was kind of staggering in terms of the depth and prevalence of these different types of deceit.”
In a literature review of 90 papers, they identified 103 instances of participant dishonesty. In those trials, between 3 to 25 percent of healthy participants were practicing some form of deceit.
In one study of experienced clinical trial participants — those who had taken part in 12 trials a year on average and had received more than $20,000 in their lifetimes as payment — the rate was higher: 75 percent said they had withheld information to avoid exclusion, and significant percentages were reported for other forms of deception.
Dishonesty in clinical trials boiled down to four categories, the study found: concealment, fabrication, drug holidays and collusion.
Concealment, the most common, involves lying about a medical history, co-enrolling in another study, or complying with trial protocol unevenly with improved adherence before a clinic visit, a practice known as “white coat compliance.” Fabrication includes exaggeration of symptoms, falsifying health status and overreporting of adherence. Drug holidays are periods of consecutively missed drug dosages without clinician approval; and collusion, the least common, entails giving study medication to others or sharing privileged recruitment information to help others gain admission to a trial.
All are serious, Kushida said, “because they all can have implications — not only giving false research findings, but also putting people’s lives at risk.”
Case in point: the study mentions an instance in which a trial participant failed to notify investigators of her history of bulimia. Due to that condition, her blood’s potassium levels were too low, and she had an adverse reaction to the study drug lithium, leading to her death.
Though Kushida and his colleagues believe their findings represent “the tip of the iceberg” of deception in clinical trials, they also think investigators can empower themselves to detect or avert such dishonesty before it imperils participants or contaminates the scientific findings.
But it won’t be easy.
Strategies currently available, such as pill counting, serum drug level testing and electronic medication packaging, all have drawbacks. Though many clinical trial participants join in good faith, Kushida recommends that investigators use more than one of these types of techniques to guard the integrity of their studies.
Researchers also could use newer statistical methods to identify people likely to deceive before they enter a trial, and to correct for the probability of deception after the fact.
Kushida and his team are currently working to create code that researchers in academia and industry can use to identify which study participants will likely adhere to treatment based on characteristics, such as demographics, psychosocial factors and medical history.
Kushida explained:
Almost any discipline in medicine that you can think of relies on adherence to medications or treatment regimens in order for patients to get better, from the clinical aspect and clinical trials as well. So, the ability to know which patients might adhere and might not, conceivably, is very important.
Photo by Global Panorama/Okko Pyykkö
