Progress in the fight against cancer relies on being able to test new drugs and therapies in clinical trials. Yet, nearly 1 in 5 publicly funded cancer clinical trials fail to recruit enough participants to yield reliable results.
“Such trials represent a waste of scarce human and economic resources and contribute little to medical knowledge,” note researchers, who have developed a mathematical tool to predict how difficult it might be to attract trial participants depending on a range of trial-specific factors.
The team, from the University of Washington and the Fred Hutchinson Cancer Research Center, both in Seattle, WA, writes about the new mathematical tool or algorithm – and how they arrived at it – in the Journal of the National Cancer Institute.
To develop the algorithm, the team first trawled through trials published in the last few years and identified several trial-related “risk factors” that are linked to low patient accrual rates.
They found, for instance, that trials requiring patients to give a tissue sample or undergo biopsy to decide if they can enroll tend to find it harder to attract participants than those that do not have such invasive eligibility tests.
Another risk factor is that when patients know they are not going to be – or are unlikely to be – treated with a potentially new drug or therapy, they are less likely to sign up.
First author Dr. Carrie Bennette, an investigator with the Hutchinson Institute for Cancer Outcomes Research (HICOR), says:
“If you have a trial looking at a new investigational drug, it’s much more likely to hit its accrual target.”
She explains this is especially true of phase 2 trials – the ones that test drug safety – which often guarantee participants will receive the new treatment.
However, this is not the case in phase 3 trials, whose purpose is to compare new treatments with the best currently available treatment. For this stage of drug testing, patients typically are randomly assigned to receive either cutting-edge, new therapies – albeit unproven – or approved treatments already in use.
This poses a dilemma; while randomization is essential for a phase 3 trial to yield reliable, robust evidence about how the new drug stacks up against the current therapy, it is this very factor that could make it less able to accrue trial participants. Dr. Bennette explains:
“As soon as you add in randomization, where patients may or may not get [the investigational treatment], it wipes away the higher accrual rates we found among trials studying new treatments.”
To identify the risk factors, Dr. Bennette and colleagues analyzed 787 phase 2 and phase 3 adult clinical cancer trials that launched between 2000-2011. They also interviewed clinical trial experts.
All the trials were sponsored by the National Clinical Trials Network (NCTN) of the National Cancer Institute, which is part of the National Institutes of Health (NIH) in the US.
The team found that 18% of the NCTN trials either closed due to insufficient participants, or they were recruiting participants at less than half of the target rate 3 or more years after they launched.
While the participants in a trial that closes can still get the drug or therapy under investigation, the trial results are not reliable, so the investigators cannot say if the drug works or not, explains Dr. Bennette, who adds:
“What that means is a clinical trial is started, a tremendous amount of resources are invested in designing the trial and finding sites to begin to enroll patients – then that trial doesn’t get used to help advance science or improve clinical practice.”
Another risk factor for predicting low trial accrual is what the researchers call “increased competition for patients from ongoing trials.” The team notes how nationally, only around 3-5% of adult cancer patients enroll in trials.
Altogether the team identified 12 trial-level risk factors that they said showed good agreement between predicted and observed risks of low accrual in a preliminary validation using 46 trials opened between 2012-2013.
The researchers incorporated them into an algorithm to predict how an NCTN trial might succeed in enrolling participants. They note that several of the factors have not been picked up in such an analysis before, and conclude:
“Future work should assess the role of such prediction tools in trial design and prioritization decisions.”
He says we “should strive to improve trial enrollment, giving the associated potential for improved results.” For example, we know trial patients tend to fare better than non-participants, because:
“It is well documented that patients on clinical trials have better outcomes than those who do not participate.”
Dr. Raghavan also notes that for many cancer patients, the barriers to enrollment in trials are not trial-specific but to do with the patient’s situation. These include health insurance constraints, the fact they live far away from the trial center and “various disparities in cancer care.” However, these factors were not covered by this study.
Earlier this year, Medical News Today learned about the first published basket study, a new type of clinical trial design that explores responses to drugs based on the specific mutations present in patients’ tumors rather than where their cancer originated.