Data Pharmaceuticals Regulation

Why use real-world evidence for clinical trials?

Clinical trials are the hallmark of the drug development process. In most cases, life sciences firms are required to conduct randomized controlled trials (RCTs) to examine their treatment’s safety and efficacy relative to standard of care. While RCTs do a great job at measuring causal inference, there as a number of limitations. First of all, they are expensive and time consuming. We’re not talking pocket change. As reported in a Bipartisan Policy Center’s report:

Clinical trials can take as long as seven years and have cost $1.5 billion of the more than $2 billion spent on drug development

Thus, if real-world evidence could in part replace part of the clinical trial process, it could bring down drug development cost, reduce prices and increase incentives for life science R&D investment. An alternative approach is to use more real-world settings with pragmatic trials.

Often referred to as pragmatic clinical trials or large simple trials, these study designs allow for more diverse patient populations to be enrolled in studies within their own care settings, all while maintaining randomization (through patient, site, or cluster randomization). Early examples include the Salford Lung Study and the ADAPTABLE study from the Patient-Centered Outcomes Research Institute’s PCORnet (National Patient-Centered Clinical Research Network).

So how do we get RWE into more common use without undermining the validity of study designs? This is a difficult question. The Bipartisan Policy Center report recommends

  • Additional funding for RWE. Real-world evidence is great to discuss, but additional funding to increase data collection and access is certainly valuable.
  • Improve regulatory clarity. FDA has previously stated that it is interested in using RWE for regulatory approval, but the details are currently vague. More reliable guidance is needed for life sciences firms to be able to determine of a proposed RWE will stand up to FDA scrutiny.
  • Improve access to data. This includes better interoperability standards to facilitate data sharing, easier and less expensive access for researchers to access data (e.g., Medicare data sources) including mortality data.
  • Use RWE to collect PROs. The patient perspective is vital, but is often expensive to collect. Digital tools such as mobile phones, wearables, biosensors, and other digital health tools can improve convenience and reduce cost
  • Assuring confidentiality/privacy. Use of real-world data is a good thing, but patient confidentiality must also be protected. Clear standards need to be established/followed to insure this goal is met.

Other suggestions were made. Many of the comments are sensible, but it is not clear in which cases RWE will be a viable alternative to the standard RCTs currently required for FDA drug approval. Still, there is much promise–and reasonable concerns–for how exactly RWE will be used going forward.


  1. There is a trade-off here but I don’t think cost is the major one. I suspect the increased difficulty of data collection will make that a wash. The real trade-off is between reliability (how much can we trust the accuracy of the data that was collected—or how much are the results affected by the data that couldn’t be collected) and generalizability (how well will the results apply to populations not studied in the trial— the real world). Randomized clinical trials (RCTs) can do well on reliability (if—and that’s a big if—they’re conducted well) but, as pointed out, poorly on generalizability. But statisticians have been complaining for generations that RCTs could do much better if they loosened up their criteria for including patients. The FDA could help here. Drug companies are so afraid the FDA will reject their trial because it wasn’t “pure” enough, or they worry they won’t show enough benefit because they included too many “poor risk” patients, or that their drug will be pulled for “side effects” due to “complications” from patients with comorbidities. So we get studies of healthy, homogeneous patients that look good, and then we unleash the drugs on an unsuspecting population. However, trying to jump past the RCT is hazardous. Without a rigorous trial design it will be difficult to know the value of a new treatment. This is a true quandary. Progress is expensive. But R&D is not the only reason for the high cost of drugs in this country. Just saying.

Leave a Reply

Your email address will not be published. Required fields are marked *