In an article in the New England Journal of Medicine on Thursday, former director of the US Centers for Disease Control and Prevention Thomas Frieden calls for greater use of alternative data sources, other than randomized controlled trials (RCTs), for health care decision making.
"For much, and perhaps most, of modern medical practice, RCT-based data are lacking and no RCT is being planned or is likely to be completed to provide evidence for action. This 'dark matter' of clinical medicine leaves practitioners with large information gaps for most conditions and increases reliance on past practices and clinical lore," Frieden writes.
The article comes as the US Food and Drug Administration (FDA) looks to flesh out a framework for the use of so-called real world evidence—such as data from observational studies, routine safety surveillance, patient registries and insurance claims—to evaluate drugs and medical devices.
Last week, FDA announced it will hold a public workshop in collaboration with the Duke-Margolis Center for Health Policy in September focused on the development of such a framework.
While RCTs are considered the "gold standard" for generating evidence to support the approval of new drugs, Frieden argues that in some cases, observational studies may be more appropriate for determining health outcomes.
"Although they can have strong internal validity, RCTs sometimes lack external validity; generalizations of findings outside the study population may be invalid," Frieden writes.
But, Vinay Prasad, a hematologist-oncologist and assistant professor of medicine at the Oregon Health and Sciences University, told Focus that this assertion somewhat misses the mark.
"When are randomized trials desperately needed? To prove that at least in some circumstances a therapy can work … you have to show that something can work under some circumstances to move forward. You cannot assume this. Only an RCT can tell you this unless the effect size is huge," Prasad said. Prasad also said that sponsors can increase the external validity of RCTs by loosening their inclusion criteria to allow for a more varied patient population.
Frieden also argues that RCTs often have insufficient study periods and small sample sizes, which can lead to unanswered questions about treatment effect duration and rare adverse events that are only answered after a product enters the market.
As an example, Frieden points to the case of FluMist, the live attenuated influenza vaccine nasal spray, which was shown to be superior to inactivated influenza vaccine in children in multiple RCTs. Despite the effect shown in those studies, a later observational study found that FluMist was less effective than previous studies showed.
Here, Prasad argues that RCTs are still necessary to hash out the effect of the two treatments. "Most interventions, including the ones discussed by Frieden … have modest effect sizes. For these things you need RCTs to separate true effect from your bias," Prasad said.
Frieden also points to a number of issues that limit the practicality of conducting RCTs, including their high cost and difficulty to plan and implement. As a tradeoff, Frieden says these constraints can lead sponsors to rely on surrogate markers or recruit patients from high-risk patient populations that might not reflect the broader patient population to speed along their studies.
But, Prasad says these issues are not inherent with RCTs, arguing that innovations could be made to bring down their costs and startup times.
Despite these issues, Frieden says that "current evidence-grading systems are biased toward RCTs, which may lead to inadequate consideration of non-RCT data." Though, while advocating for greater use of other data sources, Frieden says FDA's standards for evaluating safety and efficacy should not be lowered.
Rather, "there should be rigorous review of all potentially valid data sources," Frieden writes.