Digital Endpoints: Trends and Opportunities in Drug Development and Beyond

By Paul Nicolaus

January 17, 2024 | When the Digital Medicine Society (DiMe) launched its Library of Digital Endpoints, a listing of industry-sponsored studies of new medical products or applications that use sensor-derived endpoints, there were a total of 35. Four years later, that same library now includes over 400 unique digital endpoints collected by over 60 sponsors to answer questions about the efficacy of new medical products.  

“And it’s important that they are actually answering questions,” Jen Goldsack, DiMe CEO and founder, told Clinical Research News. “These are not pilot studies. These are not validation studies. These are not exploratory studies.” The library’s extensive growth reveals greater trust in these measures from industry and a recognition of the value and ROI, whether that’s reduced sample sizes, reduced timelines, or more generalizable findings. “Otherwise, we wouldn’t see such enormous adoption.”  

Regulations have advanced during this time, too. The FDA has issued draft guidance on using digital health technologies for remote data acquisition in clinical trials, for example, and in Europe we have seen the qualification of the first digital endpoint. 

According to Goldsack, several exciting innovations are at the heart of the digitization of clinical research. “The first is the ability for data not to live in a file cabinet in a manila folder with only the PI having the key,” she said. Instead, that data ought to be far more liquid in the digital era so that—with appropriate permissions—it can be used and reused. This alone should speed up the clinical research process and reduce the burden on participant costs.  

Another exciting part of this digitization is the ability to use new technologies, like sensors, that can be seamlessly included in people’s daily lives. In addition, algorithmic technologies and new analytics capabilities can be used to capture and analyze high-resolution data. Doing so can help arrive at “clinically interpretable information… about each individual’s lived experience of their disease or condition with or without an investigational new product.” 

Traditionally, there has been a reliance on snapshots of a person’s life. A 12-visit protocol, for instance, provides 12 individual snapshots of someone’s life, and perhaps some self-reported information between those visits. “Now we get this really high-resolution data and information and much deeper holistic understanding that’s much more generalizable to how new products and therapies are actually affecting people’s lives,” she said.  

These technologies offer better information, and the higher resolution means improved statistical power. Assuming the right measures are being selected, there is more relevant information about how a therapy affects a person. Goldsack also highlighted the upside of measuring things we haven’t been able to measure in the past. “And we can reduce participation burden because we can measure these things outside of the clinic,” she added. 

Evolving Use of the Digital Endpoints Library 

“The original purpose of the library was to be able to shine a light on areas of unmet need and potential redundancy in developing these measures,” Goldsack said. Current data shows that of those 400 plus endpoints, more than 80 are sleep measurements, and over 90 are focused on physical activities. “That’s 92 unique measures of physical activity, and frankly, there aren’t 92 interesting things to measure about physical activity.” The numbers reveal that a lot of redundant activity is taking place. 

One way that industry has used the library to make advances is to consider where activity is abundant without consensus about what the most valuable measures are. “Where we’re seeing this proliferation of competing measures, we are doing pre-competitive work to identify the most valuable measures and reduce the risk and increase the ROI for those industry players,” she said. “We are also doing similar work to identify high-value digital measures that could be used as endpoints where we are seeing a paucity of digital measures being used.”   

Consider, for instance, the unmet need for more sensitive higher resolution measures in an array of central nervous system diseases where it is known that early detection makes a big difference regarding outcomes. “We only need to think about the new monoclonal antibodies and how they only work for people with early-stage conditions,” Goldsack said. The library reveals very few endpoints devoted to Alzheimer’s and related dementias, and there is also “a huge gap and a huge lag in the use of digital measures in oncology studies.” 

We’re already seeing broad adoption of these measures as efficacy endpoints, or indicators of whether a therapy has the anticipated treatment benefit—and doing so in ways that can gather more meaningful data that can help reduce sample sizes. “What I think is interesting as we look ahead is that these digital measures that we are developing for use as endpoints are also being used in other ways to increase ROI,” she pointed out. 

Because these tools are high-resolution, they can be used to make better go/no-go decisions. In other words, they can help figure out more quickly and with a greater sense of certainty which new products are working for purposes of fast-tracking them. Similarly, they can help figure out what may not be working as intended earlier on, she added, “so we fail fast and don’t spend time and dollars developing a molecule that actually isn’t effective.”  

With sensor-generated measures, she said it is possible to imagine arriving at a “more sensitive read on different subpopulations within a broad diagnostic group” that could be especially responsive to a new drug. “So we can screen them using these tools, we can then enrich the trials, and we’re much more likely to be able to get effective therapies for subpopulations to market much more quickly.”  

Another body of work centered on ROI will be launched in early 2024 in response to requests from industry. There are questions at large life sciences organizations about the value of these measures and similar questions are being posed by investors in companies that develop digital products, often start-ups.  

There is a lingering need to “more clearly articulate and quantify the ROI that cascades from using these digital endpoints,” Goldsack continued. DiMe has been working on a study in collaboration with Tufts that does just that, and it is anticipated that the findings will be published in early 2024. “We are going to be advancing pre-competitive industry work to not only look at the hard, direct cost savings from using these endpoints but also to quantify and create business and strategy modules to support an entire digital strategy around these tools.”   

So when will we see the first medical product that is either approved based on a digital endpoint or gets a label based on a digital endpoint? Goldsack said this is a question she is frequently asked. The library has only been around for about four years, and the greater than tenfold increase in the number of digital endpoints being used by industry “speaks to enormously rapid innovation, adoption, trust, and evidence generation in this new category of products,” she said.  

“Remember, it takes 10 years to develop a new drug, and the technical success rate is 5%,” she continued. So if we consider the notion that not only do therapeutics need to go through an entire drug development lifecycle but also have technical success for an endpoint to be held up as an example of the first, Goldsack believes that “we are very much ahead of the curve.” 

If we look at the pace with which industry has adopted digital endpoints, along with the emergence of FDA guidance and a qualified endpoint by the European Medicines Agency (EMA), “this is rapid adoption,” she continued. “This is a bellwether of the future. I think it also shows that these measures can solve enormous unmet need around the meaningfulness and the relevance of endpoints to the patients that our industry is here to serve and the ROI that they offer.” 

Industry Report Sums Up Progress Made, Suggests Path Forward 

While DiMe’s Library indicates that digital endpoints are being used to develop product types that extend into categories like behavioral, biologic, and genetic, the most extensive focus areas have been devices (24%) and especially the development of drugs (68%).  

In November, Boston-based clinical research and care delivery company Biofourmis and Dublin, Ireland-based market intelligence platform and community HealthXL announced the release of an industry report that draws insights from trends in digital endpoints and biomarkers related to pharmaceutical R&D in particular. 

Entitled “Unlocking the Value of Digital Measures in Drug Development,” the report’s analysis of DiMe’s database of digitally collected endpoints used in industry-sponsored studies found that smartphone apps, activity monitors, and continuous glucose monitors have been the most frequently used digital health technologies for endpoints collection. Among industry sponsors with registered digital endpoints, Novo Nordisk, Pfizer, and Novartis have the most. 

Experts surveyed in the Biofourmis/HealthXL report predicted that oncology would see the most significant uptake of digital endpoints, followed by cardiovascular diseases. Some also see opportunity in the rare diseases space, considering the ability of digital biomarkers to pinpoint meaningful changes within relatively small cohorts.  

Survey participants indicated that the most significant benefits of leveraging digital endpoints in clinical trials are trial efficiencies, improved patient experience, and continuous disease monitoring. Additional benefits include enhanced disease insights and reduced cost of drug development.  

According to Jaydev Thakkar, COO at Biofourmis, digital measurements are becoming a boon to the industry because they can cut down on sample size and, in turn, the time it takes to carry out a clinical trial. “The time it takes for a drug to come to market is hugely dependent on the sample size because that’s what drives the amount of years it takes to recruit patients and observe,” he told Clinical Research News.  

However, the survey results also pointed to an array of barriers currently preventing more widespread adoption of digital endpoints in research and development, especially the lack of awareness among patients and healthcare providers. Other notable issues include a lack of technical expertise, regulatory uncertainty, and steep development costs.  

One reason the report remains optimistic despite these barriers is that regulatory agencies have taken an active role in coming together with industry and beginning to standardize and define what they expect the rigor to be when validating these measurements, said Thakkar. With that sort of regulatory participation, this is no longer just an innovation idea but something becoming quite real.  

“There’s certainly a lot more work remaining,” he added. But seeing standardization and rigorous validation happening and regulatory agencies accepting digitally connected measurement as the primary endpoint is a “solid sign of what’s to come.”  

As more of this work occurs across larger therapeutic areas and more disease conditions, he predicts that digital endpoints will see more widespread adoption over the next several years. However, this would rely upon greater levels of industry-wide pre-competitive collaboration. 

** 

Paul Nicolaus is a freelance writer specializing in science, nature, and health. Learn more at www.nicolauswriting.com

Load more comments
comment-avatar