The clinical research community is continuing to take a hard look at why trials of central nervous system treatments have such a high failure rate. The problems are easy to identify, experts say. What is needed are practical approaches to solving them.
The failure rate among CNS trials is second only to oncology. About 85 percent of CNS trials fail, compared to a failure rate of less than 70 percent in areas such as vaccine or ophthalmology trials.
The biggest cause of CNS trial failure is the “placebo response,” says Mark Opler, chief research officer at WCG’s MedAvante-ProPhase. When there’s a strong placebo response—subjects who psychologically believe they are receiving benefits from the investigational drug even though they are unknowingly assigned to the trial’s placebo arm—results can become muddled.
“A number of studies have suggested that placebo response rates are going up in neuroscience in general,” Opler says, adding that the problem seems to be particularly pronounced in trials that look at pain and psychiatric medications.
A second, related problem is an over-reliance on subjective outcomes in lieu of more quantifiable biomarkers. Both of these problems make it less likely that a study will yield reliable, statistically significant results.
“The placebo response is more a problem in clinical [trials] with subjective endpoints rather than objective endpoints,” explains Nathaniel Katz, Adjunct Associate Professor of Anesthesia, Tufts University School of Medicine, and founder of WCG’s Analgesic Solutions.
When patients are asked to report pain on a scale of one to 10, for example, about one-third will not be able to rate their symptoms in a consistent way, Katz explains.
“So if you enroll them in a [trial], they’re going to provide info that’s not useful,” he says. The good news is that there is a solution to this problem. Rather than toss out one-third of potentially eligible patients, you can put patients through a training course that helps them improve their ability to self-report pain. Katz has pioneered such training courses that already have been used to train more than 50,000 patients.
Furthermore, training patients to more accurately and consistently report on their own pain has a doubly beneficial effect, says Katz: “It inoculates them against the placebo response.”
Another problem in CNS studies has to do with adherence to medication. In pain studies, says Katz, adherence rates can be as low as 50 percent. Medication adherence also is problematic in depression and other conditions that affect mood and behaviors.
Within CNS studies, he says, there are numerous examples of “clinical trials that have failed. But if you look at the subgroup of patients who took their drug, the trial was successful.” Once again, there are several techniques that have been demonstrated to bolster adherence, such as medication reminders and “smart” packaging.
Medication adherence problems in CNS studies is a “scandalous cause of failure,” Katz says, “because it’s so fixable.”
CNS studies also may be more likely to fall victim to “trial shopping,” a phenomenon in which patients enroll in multiple CNS studies simultaneously, or the same CNS study twice under two different names, often because of the financial incentives offered for trial participation.
Even as the industry struggles to grapple with the problem of duplicate patients, sites also may be recruiting truthful patients who still don’t fit the study protocol, says Opler.
“In the rush to get warm bodies in clinical trials, we’re putting the wrong people in,” Opler says.
Another major challenge with CNS trials is the operational burden they place on study sites. The technology implemented to overcome problems like placebo effect places a greater burden on providers, meaning few research centers are incentivized to conducted CNS clinical research.
For six years, Sean Walsh, president of independent sites at WCG Clinical, ran an independent research site within a neurology practice. “Over those six years,” Walsh says, “things to do in the studies just became more extensive and difficult.”
For Alzheimer’s disease studies alone, there might be more than a dozen different technology platforms investigators and staff would be asked to use for one study, he says.
“It just kept getting worse and worse,” says Walsh. “The criteria had become so stringent that a lot of the PIs didn’t want to do it anymore because none of the Phase III criteria married up to the standard of care. It’s become so narrow and they want to hit so many secondary endpoints, that it’s become impossible to find those patients.”
This burden, “makes it less likely new investigators will get into these trials,” agrees Opler. Investigator fatigue places a heightened imperative on pharma sponsors to carefully evaluate the utility of tech solutions and the complexity of CNS protocols.
Overall, many of the problems that contribute to CNS trial failure are fixable, experts agree. And considering the dearth of good therapies for CNS diseases, Opler notes that “a failed trial isn’t just a scientific failure, it’s also an ethical failure.”