Object moved to here.

Why Should the FDA Focus on Pragmatic Clinical Research? | Research, Methods, Statistics | JAMA | 西瓜视频


[Skip to Navigation]
Sign In
Integrating Clinical Trials and Practice
闯耻苍别听3, 2024

Why Should the FDA Focus on Pragmatic Clinical Research?

Author Affiliations
  • 1US Food and Drug Administration, White Oak Campus, Silver Spring, Maryland
JAMA. Published online June 3, 2024. doi:10.1001/jama.2024.6227

Traditional randomized clinical trials (RCTs) have long been a key tool underpinning drug and device development. The use of individual participant randomization and active or placebo controls in RCTs, combined with comprehensive collection of highly structured data, supports assay sensitivity. At the same time, focused enrollment criteria and careful attention to the collection of adverse events for specified follow-up periods promote detection of toxicities and risks. These trials support a system, regulated by the US Food and Drug Administration (FDA) and other global regulators, that allows the majority of candidate therapies whose risks outweigh benefits for intended use to be screened out while enabling safe and effective medical products to advance to market. However, the next stage鈥攁fter product development and marketing authorization are completed and a therapy is integrated into clinical practice鈥攏eeds serious attention.

The US system for evaluating risks and benefits of medical products before initial marketing approval often does not provide all the information needed to inform decisions concerning diagnostic and therapeutic strategies in routine health care practice. This includes questions regarding comparative effectiveness of different products or optimal sequences of treatment. An evidence generation system integrated into routine health care practice after initial market approval holds promise for answering these questions. This is true in many therapeutic areas but especially so for common chronic diseases (eg, cardiovascular disease, pulmonary disease, kidney disease, type 2 diabetes, and Alzheimer disease), for which large clinical trials are required because treatment effects are often modest and risks and benefits are heterogeneously distributed across affected populations.

One approach for generating evidence after initial approval is real-world evidence, which the FDA defines as 鈥渃linical evidence about the usage and potential benefits or risks of a medical product derived from analyses of real-world data (RWD).鈥 Spurred by technological advances, sources of RWD have grown substantially. Electronic health records are now ubiquitous, and the quality of data collected is constantly improving as shared standards and terminology are developed and adopted. The uptake of smartphones and wearable technologies with sophisticated sensors enables data to be collected directly from patients where they live. Additionally, large datasets now offer insights into social and environmental determinants of health.

In concert with technological advances, methods for analyzing and applying lessons from observational data have progressed. The use of RWD to better understand the quality of health care delivery and the risks and toxicities of interventions is well established, and methods to account for confounding have improved, which has enabled attempts at causal inference.1 Real-world evidence derived from observational study designs has informed clinical care and in some cases contributed to FDA regulatory decisions on the effectiveness of medical products.2 However, real-world evidence derived from observational studies is limited by inherent biases that cannot be entirely overcome with improved analytical methods. Data quality remains variable, and many end points are not available in current RWD sources. As a result, many RCTs used in drug approvals cannot be replicated using observational designs and RWD that do not include study-specific data collection.3

Pragmatic Clinical Research as a Middle Ground

We at the FDA believe that pragmatic clinical research represents a middle ground between traditional RCTs and observational studies and that this area of research deserves much more attention and emphasis. Pragmatic clinical research uses RWD, often combining it with data collected specifically for research. It features streamlined, prospective trial designs fit for answering a specific, focused question and often applies randomization, although other designs may be used. Such research holds special promise when embedded in health care systems, where it can help investigators avoid some of the most critical drawbacks of 鈥渢raditional鈥 clinical trials, such as substantial, sometimes prohibitive costs, as well as the loss of diverse representation that can arise from conducting trials in environments that are unrepresentative of all clinical practice settings and patient populations.4 However, evidence from pragmatic clinical research can present organizational, methodological, and ethical challenges that require experience and focused expertise to navigate.

The FDA has an interest in pragmatic clinical research after initial market approval in several important areas. There is a need for robust data to inform clinical practice, particularly confirmatory postmarket studies. These trials may more closely resemble the real-world settings of clinical practice than they do traditional clinical trials and may leverage more pragmatic designs to focus on key data elements needed to confirm benefits and risks for the intended populations. Ideally, these trials would be integrated into routine clinical care and conducted in diverse practice settings to ensure that the populations enrolled reflect the disease burden.

Pragmatic evidence is also needed to evaluate new technologies such as gene editing to treat rare diseases. Because the long-term consequences of such interventions (including off-target effects of genetic changes) are unknown, the FDA recognizes the need for cost-effective methods for assessing patients鈥 progress long after the immediate intervention is completed. Furthermore, these medical products raise unique issues for clinical trials due to the small target populations, with some treatments that address not a disease but a specific pathogenic variant. For these types of therapies, combination approaches are needed that fine-tune the use of pragmatic methods to answer specific questions. For example, this could entail an RCT at the time of approval that includes consent and procedures for long-term follow-up using RWD, and that is designed to be used as an external control group for future pragmatic studies.5

Common chronic diseases present different issues. Approvals of drugs targeting amyloid plaque in Alzheimer disease and glucagon-like peptide 1 pathways in the treatment of obesity and diabetes have raised questions about optimal treatment strategies in real-world settings. After drugs are approved according to the FDA鈥檚 rigorous standards for safety and efficacy, questions remain that are better suited for trials in real-world clinical practice, such as the optimal order of therapies or the populations most likely to benefit. These questions could be answered with streamlined, patient-centered trials. Such trials may especially benefit sites in underresourced settings that might otherwise be unable to sustain the effort of participating in traditional clinical trials. This is particularly important in chronic diseases such as obesity and diabetes that are highly prevalent in racial and ethnic minority communities and in patients with lower socioeconomic status. To enroll a large, diverse sample of patients at a low cost per participant, these trials could take advantage of streamlined designs and use data sources such as electronic health record and claims data, wearable sensors, and virtual methods embedded into clinical practice.6


Pragmatic clinical research offers a middle path between traditional RCTs and observational evidence to inform decisions about the use of medical products after initial market approval. Pragmatic clinical research should be leveraged to answer questions about medical products in the context of real-world clinical practice with appropriate transparency and ethical safeguards. Technological and methodological advances in RWD and pragmatic trial design make it possible to assess risks and benefits of medical products across heterogeneous populations at much lower cost than would be typical for traditional RCTs, while also yielding more valid findings than observational studies. Pragmatic clinical research can inform regulatory decisions while also supporting patients, clinicians, health systems, and payers in assessing the comparative effectiveness of treatments and making determinations about the relative value of different interventions.

Back to top
Article Information

Corresponding Author: Robert M. Califf, MD, US Food and Drug Administration, White Oak Campus, 10903 New Hampshire Ave, Silver Spring, MD 20993 (commissioner@fda.hhs.gov).

Published Online: June 3, 2024. doi:10.1001/jama.2024.6227

Conflict of Interest Disclosures: Dr Abbasi reported receiving consulting fees from MIS Technologies LLC and Le Maitre and receiving a National Institutes of Health T32 training grant (5T32CA251070-04) on assignment at the US Food and Drug Administration (FDA). Dr Curtis reported receiving past consulting income from Regeneron Pharmaceuticals, NFL Players Association, and Boehringer Ingelheim and receiving institutional grants from GlaxoSmithKline and Novartis. Dr Califf reported that prior to his appointment to the FDA as Commissioner for Food and Drugs, he was an employee of, and held equity in, Verily Life Sciences and Google Health (Alphabet) and served on boards of directors for Cytokinetics, Centessa, Clinetic, Keystone Symposia, the Critical Path Institute (CPATH), the Clinical Research Forum, and One Fifteen.

Hern谩n 聽MA锘, Wang 聽W锘, Leaf 聽DE锘. 聽Target trial emulation: a framework for causal inference from observational data.聽锘 听闯础惭础. 2022;328(24):2446-2447. doi:
Concato 聽J锘, Corrigan-Curay 聽J锘. 聽Real-world evidence鈥攚here are we now?聽锘 聽N Engl J Med. 2022;386(18):1680-1682. doi:
Wang 聽SV锘, Schneeweiss 聽S锘, Franklin 聽JM锘, 聽et al; RCT-DUPLICATE Initiative. 聽Emulation of randomized clinical trials with nonrandomized database analyses: results of 32 clinical trials.聽锘 听闯础惭础. 2023;329(16):1376-1385. doi:
Phillips 聽CM锘, Parmar 聽A锘, Guo 聽H锘, 聽et al. 聽Assessing the efficacy-effectiveness gap for cancer therapies: a comparison of overall survival and toxicity between clinical trial and population-based, real-world data for contemporary parenteral cancer therapeutics.聽锘 听颁补苍肠别谤. 2020;126(8):1717-1726. doi:
Eichler 聽HG锘, Pignatti 聽F锘, Schwarzer-Daum 聽B锘, 聽et al. 聽Randomized controlled trials versus real world evidence: neither magic nor myth.聽锘 聽Clin Pharmacol Ther. 2021;109(5):1212-1218. doi:
Ishani 聽A锘, Leatherman 聽SM锘, Woods 聽P锘, 聽et al. 聽Design of a pragmatic clinical trial embedded in the electronic health record: the VA鈥檚 Diuretic Comparison Project.聽锘 聽Contemp Clin Trials. 2022;116:106754. doi: