What if we could better understand the effects of treatments on specific types of individuals, thereby allowing us to personalize drug treatments?
We recently had the opportunity to discuss our efforts toward realizing this goal during a panel session, which was part of a two-day workshop called “Real World Efficacy: Patient Subgroups.”
The workshop was organized and hosted by the RWE Collaborative, of which Holmusk is a member. The RWE Collaborative, organized by the Duke-Margolis Center for Health Policy, is a group of stakeholders that comes together to improve the development and use of real-world evidence (RWE) and inform the FDA of how RWE can be better implemented and leveraged to improve research.
The theme of the workshop aligned quite well with research we have been conducting here at Holmusk using Holmusk’s unique data asset, the NeuroBlu Database. The NeuroBlu Database offers real-world clinical behavioral health data captured from the electronic health records of over 1 million patients across the U.S. Notably, the NeuroBlu Database also contains a more representative population than traditional clinical trials do, making available data from patients with different racial and ethnic backgrounds.
A study we conducted recently underscored the importance of this representation. First, we drew a patient cohort from the NeuroBlu Database that emulated what is typically seen in clinical trials—ethnic minorities, hospitalized patients, and those with failed clinical encounters were unrepresented. In comparing the effectiveness of different antidepressants, we found results that were in line with what has been observed in randomized controlled trials.
But here’s where it gets interesting. When we pulled a cohort from the NeuroBlu Database that contained racial and ethnic minorities—and therefore captured a group of patients typically under-represented in clinical trials—our results looked markedly different.
These results illustrate how findings that appear clear from a randomized controlled trial might not be generalizable to individuals not represented in that trial. In other words, it highlights that when clinical trials only include one type of patients, these results might not improve clinical care delivery for unrepresented patients.
Moving forward, real-world evidence should be used to complement the findings from and help to identify potential limitations of randomized controlled trials. Although our findings from this analysis are preliminary, in the future, we could use this type of evidence for a multitude of use cases, such as identifying which factors may be important to consider when selecting the most effective treatments.
It is encouraging that the FDA has recognized the role that real-world evidence can play in future product approvals. By providing granular data down to the symptom level, real-world data sources like the NeuroBlu Database open the possibility to better understand individual patients’ trajectories and the factors that affect them. Ultimately, we hope this deeper understanding of how subgroups fare in the real world will enable us to improve the care for all groups of patients.
Want to learn more about the NeuroBlu Database? Visit the website.
Image by Gerd Altmann from Pixabay