By Deborah Borfitz

April 3, 2025 | Pragmatic trials, by name, have been around for well over a decade now and, while well intentioned, they remain poorly understood and currently offer a limited value proposition to study sponsors and contract research organizations (CROs), according to Drew Garty, chief technology officer, clinical data at Veeva Systems and a self-described “optimistic contrarian.” In his view, the ideal starting point is pragmatic innovation where the industry remains focused on simplifying and standardizing what happens in research for the benefit of all stakeholders, especially patients and study sites. 

Although pragmatic trials seek to bridge the gap between clinical research and clinical practice, the chasm persists as a direct reflection of the complexities study sponsors have unwittingly heaped on investigator sites that are trying to simultaneously deliver real-world patient care, Garty says. Technology intended to provide solutions too often has become the central barrier to success, contributing to the perennially high attrition rate among research sites. 

Garty says simplicity is missing in the design of study protocols, in terms of where and how “reasonable and appropriate” data gets collected, as well as the ways in which study sponsors and CROs ensure that information is complete and correct. The industry is quick to talk about “standards” without always fully understanding what the term means, he points out.  

Clinical research might be simplified by leaning on general healthcare standards implemented across large health networks, enabling existing, real-world data to be used in lieu of collecting so much specialized data in a trial setting, Garty continues. In fact, one of the best conceivable ways of conducting a pragmatic trial is to figure out how to collect the necessary scientific data at the time of patient care in an electronic health record (EHR), thereby shortening the journey from EHR to clinical statistical analyses. Attempts have been made for decades with very little success.  

But just because sites are using the same EHR system doesn’t mean everyone is collecting the same data in the same way, says Garty, since the platforms are typically implemented to the specifications of a health network. The FHIR (Fast Healthcare Interoperability Resources) global standard facilitating the exchange of EHR information between different healthcare systems continues to gain traction, but the focus is on standard-of-care data that doesn’t necessarily provide the breadth of safety and efficacy data being sought for trials. 

Most health system data have no regional standards, let alone international ones. “That is the crux of our data problem,” he says.  

If data are being collected in different ways, the question becomes how it can be aligned to have semantic equivalence with real-world patient data, adds Garty, and then be processed so it is usable for analysis. “We don’t today have resources for that.” 

Data management groups weren’t designed to take in any form of data in response to a standardized set of research questions, he says. “Given the real world out there and the wide range of the data, you have to call into question the feasibility of trying to make use of all of that [information].”  

Multiple Challenges

It is likely that fewer than 5% of people working in the industry could articulate anything meaningful about pragmatic trials, and this includes “many of the data management organizations who would naturally carry a lot of that data,” Garty says. The misunderstanding stems from the fact that clinical research is “incredibly complicated” with everyone preoccupied with their own universe of complexities. 

Pragmatic trials have more challenges than enablers, says Garty, in talking about how things came to be so difficult. In the late part of the last century, industry was focused on digitizing information quickly and making it available for analysis. This made the data collection exercise a lot less intuitive to site personnel who first had to learn how to use computers and then contend with an ever-growing volume and variety of technologies—from clinical trial management and eRegulatory management systems to systems for electronic data capture (EDC), clinical outcome assessment, and patient-reported outcomes.  

When teams are implementing a well-defined pragmatic trial, “they are working almost in spite of the traditional trial process,” he continues. “It requires adding complexity yet working very nimbly in usually small teams to work around our normal practice and not by bypassing quality... the reality is that we must engage with sites in very bespoke ways to understand how they work best and then to facilitate data flow that is in cooperation with their goals of patient care.”  

Talk about enabling pragmatic trials is only a statement of intent, says Garty, returning to his “simplify and standardize” mantra. “We have a lot of very smart people in our industry who have solved complex problems, but that doesn’t necessarily make the solution great. A great solution to a complex problem is, again, simplicity, and that word matters” because its absence is the barrier of entry for sites and the barrier of continuity for them from study to study.  

The problems to be solved are about people, which is where process and best practices come in, Garty says. “To simplify... start with the protocol” and not the data or technology solution.” 

The industry overall has a “proliferation of extraneous data” that gets collected in studies out of a desire for more insights with no real checks and balances in place to reduce the data overload, says Garty. The simplicity of the experience for patients, in terms of how many and what kinds of procedures they must endure through the trial, and sites, in terms of how much data they must collect regardless of the method, is what matters—and that requires a laser focus on what is critical for that study. 

Old But Immature Idea

For nearly as long as electronic health records have been around, sponsors and CROs have been trying to use that data for secondary purposes, Garty says. But, as with the similarly ambiguous term “risk-based monitoring,” pragmatic trials generally haven’t changed interactions with sites.   

“There are so many things that we can and should be doing to decrease cost and complexity, but we often don’t do all of them, or any of them,” Garty says. “We just choose one or two.” 

NIH Collaboratory Trials, large-scale pragmatic clinical trials conducted by the National Institutes of Health in partnership with healthcare systems, now number 35. Many organizations are leveraging real-world data and real-world evidence to supplement and, in some cases, replace clinical trial data, he points out, but the practice is still far from ubiquitous. 

Efforts in that direction began 25 years ago with a series of interoperability projects, then centered on DIA Connectathon events held in conjunction with the Drug Information Association annual meeting. These included a collaboration Garty was involved in that successfully put clinical trial questions inside of an EHR. A handful of leading research sites have been lighting the way with pragmatic trials; the problem has been executing those studies on a large scale.  

Perhaps the biggest misconception about pragmatic trials is “the maturity of them,” says Garty. Up until the last couple of years, even his understanding of their scope was limited, he admits.  

“In my mind, it was always about getting the product to market as fast as you can if it is safe, effective, and comparatively better, so patients and doctors would start making use of it, and that is just not the reality,” Garty continues. “We have many layers that stand between pharmaceutical companies and patients who are in need of those therapies, and it takes a lot of effort and a lot of data to actually inform those policy decisions.” 

Data Waste

Sponsors and CROs could accomplish more of their objectives with pragmatic trials by simply collecting less data, he argues. He worked with one sponsor who was “too standardized” regarding what data would be collected study to study, meaning it wasn’t properly tailoring data capture to the requirements of an individual trial. As a result, “less than 40% of the data was actually used for analysis... [and] the value of the data has decreased because you have waste.”  

Massive amounts of waste are happening simply because sponsors do not regularly challenge their internal standards, says Garty. Data managers, on behalf of clinical development, “add things to them but rarely take away because [they] typically don’t know what that data is used for.” With marketing approval potentially on the line, “when in doubt they leave it in, and so the standards grow.”  

The remedy is to “simplify the complex,” he reiterates. “Our solutioning as an industry reminds me of a Rube Goldberg machine [a contraption designed to perform a simple task in an overly complicated way]—we know how to take half a step forward, but we don’t know how to take a full step where the foot behind you comes up off the ground and moves forward again... [and] we end up with this complex hybrid set of processes.” 

His tactic is to challenge the industry’s implementation of technology tools, including artificial intelligence solutions that have created a lot of blind enthusiasm, to be sure they are the most pragmatic choice. It is sometimes necessary or appropriate to “rebuild from the ground up” on a new foundation, says Garty. 

Patient Burden

Consider the patient’s experience, Garty says. Qualified clinical trial candidates invited to participate aren’t going to agree if they view the logistics as impractical, e.g., 12 visits over the next month to a study site 30 miles away during its normal 9-5 weekday hours of operation. But they might, if they are offered an option of using a device at home to wear or complete questionnaires to provide some of the required information, reducing their investment in trial execution to once a week. This could be particularly helpful to people in underserved populations that have a long commute to get to a study site or just don’t have the available time, he adds. 

“You have to understand the patients that you’re serving—are they healthy, are they sick, do they have physical disabilities, are they geriatric or pediatric—what is that burden?” Questions about whether patients can be found and attracted into a trial are being considered today, says Garty, but pragmatic approaches that could help the most have yet to gain a foothold. 

One of the longstanding visions with pragmatic trials is to repurpose general healthcare data to create a synthetic control arm in a clinical study, in lieu of a traditional placebo group, to keep the focus on patient treatment, he notes. “In many cases, we already know the health outcomes of those populaces” and real-world EHR data could therefore be incorporated into conventional trial design.  

Cost of a Data Point

Garty moderated a panel session on pragmatic trials at this year’s Summit for Clinical Ops Executives (SCOPE), where the key takeaways were that “adding complexity to complexity doesn’t typically yield simplicity or scalability” and that clinical trial stakeholders need to be working as a united front to push for pragmatic innovation. He also answered his own challenge with a commitment to positively influence the simplifying of trial design by calculating the cost of a data point over the next year. 

The true cost of asking a patient or collecting a data point about a patient has been one of the elusive numbers that might help drive pragmatism in clinical trials, says Garty. The plan is to give clinical development a way to articulate the real-world implications (e.g., patient time spent at the doctor’s office), and not just the monetary expense, of adding “just one more thing” to a study protocol. Work toward that goal begins at an upcoming Veeva executive summit in New York, where Garty will be addressing about 14 industry executives.  

Master protocols or multipart studies (e.g., phase 1b/2) are ways medical writers can get “more scientific bang for the buck,” he says. These are essentially studies with multiple pathways that can replace various separate studies by bringing them together into one and managing them “seamlessly and elegantly.” 

Veeva is not the only vocal champion of pragmatic trial design, he says. So is Gracy Crane, international regulatory policy lead for real-world data at Roche, another speaker at SCOPE, who has been helping to drive better industry understanding of the regulations supporting pragmatism in trials. 

Many solution providers are also tackling EHR-to-EDC integration, he adds, while companies like OM1 are taking on the business problem from a data perspective. “In conversations leading up to the conference and after, we talked a lot about the overlaps and how we might enable each other’s initiatives.”