Dr Schuyler Jones: ADAPTABLE Is a Study Model for the Future

ADAPTABLE was an opportunity to accomplish a large-scale study in a generalizable manner: directly involve patients, partner with them, and cocreate the program, noted Schuyler Jones, MD, associate professor of medicine at Duke University.

Transcript

How did the approach to ADAPTABLE for studying aspirin use in patients with cardiovascular disease come about? As this was a pragmatic trial, what key adjustments did you have to make, and why?

It’s been a question that's been ongoing for probably decades. In fact, I remember Rob Califf [Robert M. Califf, MD, professor of medicine in the Division of Cardiology at Duke, and a former FDA Commissioner] mentioning that he wanted to study this when I was a young physician-clinician at Duke as an intern and resident. So it's definitely something that's been talked about for a long time.

The pragmatic nature and the opportunity presented itself when PCORI [Patient-Centered Outcomes Research Institute] funded PCORnet [The National Patient-Centered Clinical Research Network]. We needed a clinical question that was important to clinicians, important to patients, and then able to be potentially answered with newer methods, like direct-to-participant outreach and an open-label design. So that's how it came about.

The second question, I think, was about did we have to make any changes? Certainly we did make a few changes during the course of this study, but overall we held true to wanting to randomize patients 1:1 to the 2 most common doses of aspirin, 81 and 325 [mg]. We were able to finish the study, complete it in a short period, and provide an answer.

What are the most important takeaways from this study?

The story is really 2-fold. One is about aspirin, and then the other is about the process of the study, how we did it. And so for aspirin, it really is that there aren't differences in the effectiveness or safety of the 2 most common used doses. So we think that in most cases, because most patients right now are taking 81 mg, they shouldn't switch; they should stay on 81 mg.

From a process standpoint of the trial, we think that the way that we did this trial may be a way to do studies like this in the future. So, directly involve patients at every aspect, partner with them, cocreate the program, and stay true to why we're doing studies for our patients, as well as some of the novel methods like direct-to-participant enrollment and recruitment and using some of the data sources, like electronic health record [EHR] data and health insurance claims data.

A New England Journal of Medicine editorial wondered if a pilot study might have revealed the tendency for switching. Did the study group consider this option?

We really didn't. When you go directly to participants, like we did, it's hard enough to talk to patients about research. We really wanted it to be a streamlined, as-straightforward-as-possible clinical study. So if we did a pilot study, that would have required patients to do something different for some period of time, followed by the actual study. So I think that while the intent of the editorial was good, it's hard to think of how that would fit into how ADAPTABLE was run.

To be quite honest, pragmatism doesn’t have to be perfect. We just have to do it better and faster and more focused on the things that are important to people like clinicians and patients.

What lessons about the use of real-world evidence do you think ADAPTABLE gives the research community, especially in cardiology?

Cardiologists are so dedicated to data. And I think one of the criticisms that I always hear is, clinical trials are never done in real-world data. I don't think that's true. Clinical studies sometimes are a little bit more refined than what you would see in a generalizable, kind of everyday practice. But at least for this, it was an opportunity to truly get a large-scale study done in a generalizable manner. I think we showed that, with almost 500,000 patients approached and in the ability to use stuff that's already collected, like EHR data, like insurance claims data. That really, I think, is a model for the future.