July 16, 2024 | A recent analysis of interventional clinical trials by the Tufts Center for the Study of Drug Development (CSDD) has found that use of digital endpoints has been growing rapidly since 2015 with drug developers now initiating an average of about 35 such studies each year. News of the uptick isn’t entirely surprising, given mounting evidence that the approach “can improve operational performance and ease patient burdens in clinical trials and in clinical practice, while [potentially] yielding financial net benefits to developers,” according to Joe DiMasi, Ph.D., research associate professor at Tufts CSDD.
Willingness to experiment with digital endpoints is partly a reflection of the growth and use of wearables, DiMasi says. The high adoption rate of internet-connected sensors (55%) and continuous glucose monitors (CGMs, 24.7%), the technologies involved in the creation of almost 80% of digital endpoints in clinical trials, is likewise a mirror of their widespread use in clinical settings and increasing prevalence in the general population.
Data used in the Tufts CSDD Impact report included 164 clinical trials and 393 digital endpoints taken from the Digital Medicine Society’s (DiMe) Library of Digital Endpoints. The library has grown over tenfold in four years, reports Partnerships Lead Sarah Valentine.
The impetus behind the investigation was to better understand the economic value of digital endpoints in a clinical research setting, the expectation being that this will further spur their adoption, she says. Evidence has long existed that connected sensors have a lot of scientific value, “but up until this study we didn’t have evidence of the impact digital endpoints have on study size and duration.”
Scientific value has been gleaned from continuous sensor-based measures such as wearables and CGMs that “provide a unique snapshot of patients’ lived experiences inside and outside the clinic walls,” says Valentine. With the advent of CGMs, “we don’t need to go finger prick to finger prick anymore. We have visibility into patients’ glucose level over time, which enables better glycemic control and measurement of time in range.”
Central nervous system (CNS, 35.4%), diabetes (20.7%), and cardiovascular (14%) clinical trials are the therapeutic areas that most often employ digital endpoints, since “technologies [e.g., CGM devices and actigraphy/wearables] exist that can continuously measure markers of benefit in these areas,” says DiMasi. These are also the disease categories seen most frequently in the DiMe’s Library of Digital Endpoints.
“Diabetes and cardiovascular are relatively straightforward [since] the sensor-based products that exist within these spaces are relatively mature and easily understood by the field,” says Valentine.
With CNS, it is less a matter of asking “why digital” than “why not,” she adds. “The ‘gold standard’ scales that exist for so many CNS indications are frankly insufficient at capturing patients’ lived experiences,” increasing the scientific value of the digital alternatives for historically underserved populations such as individuals with Parkinson’s disease, major depressive disorder, and rare developmental conditions.
The Tufts CSDD analysis was restricted to interventional studies where industry was either the lead sponsor or a collaborator, since “they may be run differently or for different purposes than non-industry studies,” DiMasi says. “In addition, the DiMe data are almost exclusively for industry-involved studies.”
Numerous observational, behavioral, and other types of studies that populate ClinicalTrials.gov—the primary data source—were excluded from the study to keep the focus on the development of drugs, biologics, and devices, he adds.
Data related to the cost of creating, implementing, and validating a digital endpoint measure came from online surveys completed by 36 endpoint developers and 35 sponsor companies. Developer costs to create a digital endpoint measure was pegged at a median of $3.8 million, significantly more than the follow-on investment to put it to use in studies.
Sponsor costs to implement and validate a digital endpoint varied widely, from less than $500,000 to more than $15 million (mean of $3.4 million). Here, says Valentine, it is important to consider that when sponsors implement a digital endpoint at scale, they incur costs beyond the digital measurement product itself—e.g., the logistics of getting these devices to and from sites, aggregating and statistically evaluating the data, and executing study startup and support processes.
The report was part of a larger effort by the Tufts CSDD to quantify the net financial impact of deploying digital endpoints in later-stage phase 2 and 3 clinical trials that has been submitted for publication.
This is believed to be the first study to comprehensively quantify benefits and determine financial value of digital endpoints, DiMasi says. “There was an earlier study [Digital Biomarkers, DOI: 10.1159/000525255] that posed some hypotheticals and did simulations that suggested that enrollment sizes could be lower with the use of digital health technologies, but the cases were hypothetical and there was no financial component to them.”
As for case examples where the DiMe has seen similar results, Valentine points to acceptance by the U.S. Food and Drug Administration (FDA) of a reduction from 300 to 140 subjects in a phase 3 fibrotic interstitial lung disease study based on the effect size observed in phase 2 (as measured by actigraphy). The Lilly-Verily PRESENCE study of Lewy Body dementia also showed detection of drug effect with a smaller sample size with the use of wrist-worn sensor-based measurements (Parkinsonism & Related Disorders, DOI: 10.1016/j.parkreldis.2023.105355).
Roche and digital endpoint developer Sysnav additionally co-authored a 2022 paper that discusses the potential for stride velocity 95th centile (SV95C) to reduce sample size in Duchenne muscular dystrophy studies from 105 to 14 per treatment arm (Journal of Neuromuscular Diseases, DOI: 10.3233/JND-210743).
The Impact report indicated more than half of digital endpoints are being used to support secondary clinical trial outcomes. This reflects “an element of caution” on the part of pharmaceutical companies, says Valentine, and rightfully so given the importance they place on evidence and rigor, and “ensuring that measures are verified, analytically validated, clinically validated, and usable within a given patient population.”
Moving forward, Valentine says she’s confident more studies using primary digital endpoints will emerge. The Digital Medicine Society, for its part, has been establishing field-wide best practices to support these processes. Regulatory bodies, notably the FDA and European Medicines Agency, have also continued to introduce policies guiding fit-for-purpose use of digital endpoints.
Progress is being seen by regulatory agencies in other parts of the world, including the Asia-Pacific region and Latin America, in clarifying their perspective and supporting innovators who are adopting these new technologies in their trials, says Valentine. Although the report shows a greater use of digital endpoints in trials conducted in North America, “there’s likely some bias in this initial study.”
International studies are typically only reported to ClinicalTrials.gov if they involve an FDA-regulated project or funding by the National Institutes of Health, she explains. European trials are mandated to report to EudraCT.
On a practical level, clinical trial operators tend to rely on “tried and true” primary endpoints for clinical trials, says DiMe Program Director Victoria Bangieva. “Digital endpoints have not been around as long, and the experience and trust aren’t fully there yet.”
For a new, related project, the Digital Medicine Society recently announced it is developing a framework and resources to help stakeholders build a compelling case for continued investment in digital strategies that align with business goals and industry standards. Its Digital Health Measurement Collaborative Community (DATAcc) with the FDA’s Center for Devices and Radiological Health has already brought together over 30 partners from industry, government, and the startup community to advance the use of digital endpoints in clinical research.
The FDA has issued draft guidance on the use of digital health technologies for remote data acquisition in clinical trials. The agency also recently qualified the first digital endpoint for clinical trials—specifically, the atrial fibrillation history feature of the Apple Watch for use as a secondary endpoint in trials for cardiac ablation devices.
The EMA made the first move in 2019 by qualifying SV95C as a secondary endpoint (Digital Biomarkers, DOI: 10.1159/000517411) and last year as a primary endpoint. The FDA still has SV95C under review.