How to use your real world data (RWD) for clinical studies—Taking a page from the FDA’s playbook II
In a previous insight, we discussed how to evaluate if your real-world evidence (RWE) is good enough. In this one, let’s take a different perspective and consider how we, as digital health innovators in the age of data-driven medicine, can use real-world data (or the data that your solution collects) for clinical study purposes.
What is real-world data (RWD)?
RWD is data concerning patient health status and care delivery, which can come from electronic health records (EHRs), claims, disease registries, patient-reported outcomes, and other sources. For example, a digital solution might gather continuous information through remote monitoring devices. The bottom line: It comes from the real world—the clinical setting—not from a lab or trial.
What can it be used for?
RWD can serve as an integral component in different study designs that validate the effectiveness of a product, treatment, or solution. Whether your goals are to help move research forward with access to more data or to prove the efficacy of secondary use cases, here are key considerations when using real-world data:
Availability: Do you have the user agreement or permission to use data?
Areas: Which treatments, disease types, or patient journeys can you evaluate in real-world settings?
Quality: How high is the quality collected in the clinical setting?
Breadth: How many patients are accessible? Are there diverse populations?
Controls: What variations and missing controls must be accounted for in real-world settings?
Bias: How can you mitigate biases?
Now, let’s look at how RWD plays a role in study designs.
RWD in observational studies
RWD is critical to studying the effects and uses of drugs and treatments (AKA pharmacoepidemiology) in real populations. For this type of study design, RWD may be drawn from many sources (EHRs, registries, etc.) and unlike clinical trials, which isolate the question of interest, observational studies are conducted retrospectively (looking back). This reflective approach, however, leaves the opportunity for relatively inexpensive and manipulative re-analysis of the dataset to support a preferred hypothesis. Thus, measures need to be taken in order to ensure an unbiased result.
RWD as external control in clinical trials
In some clinical trials, it’s not possible to randomize participants (i.e., it’s not feasible and/or ethical to offer treatment to one group and a placebo to the other). Effective clinical study design, however, requires some type of control. In non-randomized trials, RWD and/or historical data (from past clinical trials) can serve as a sufficient external control arm in combination with statistical methods. As is expected with using RWD to generate “good-enough” RWE, there are limitations that must be considered (quality of data, locating comparable populations, etc.).
One last note: Transparency
Using RWD in study designs to create quality real-world evidence presents an incredible opportunity to make clinical studies less expensive and more efficient. However, transparency is a huge challenge that needs to be addressed, and innovators may find themselves straddling the line between patient privacy and the reliability of data for regulatory purposes.
For more on assessing RWD, check out the source for this article, the FDA’s RWE framework, which can serve as a starting point for innovators worldwide.