Some of the evidence for digital therapeutic suffers from selection bias and relatively short studies, panelists at the Pharmacy Benefit Management Institute said today. They also discussed making the flood of data from the digital therapeutics accessible and useful to payers and clinicians.
The promise of digital therapeutics to transform healthcare, particularly when it comes to the management of chronic disease, is there, agreed the participants in a lively panel discussion on digital therapeutics today at the 2022 Annual National Conference of the Pharmacy Benefit Management Institute (PBMI) in Orlando, Florida.
“We have always understood that we can’t tell people what to do,” said Jeffrey Dunn, Pharm.D., MBA, chief clinical officer, of the Cooperative Benefits Group. “They have to buy into the plan, they have to be engaged or their behavior is not going to change. So that is where these things (digital therapeutics) potentially can play a huge role.”
But Dunn and the other panelists explored the welter of issues and obstacles that stand in the way of great acceptance of digital therapeutics by the payers who would be shouldering their cost and physicians, nurses and other clinicians who would be in the position to recommend or prescribe them to patients.
The other members of the mid-afternoon panel were Patty Taddei-Allen, Pharm.D., MBA, BCACP, BCGP, vice president, clinical programs and services, at WellDyne and Bill Rush, senior director of value and access for digital healthcare at Sanofi. Timothy S. Regan, RPh, CPh, a vice president at AmerisourceBergen/Xcenda, moderated the discussion.
Taddei-Allen talked about some of the weaknesses of the evidence that the digital therapeutics companies offer up as proof that their products are effective. She mentioned selection bias — study participants who are not representative of a population — and study periods that are relatively short. Few studies are designed to compare a strategy that uses a digital therapeutic with the standard of care, she noted.
But Taddei-Allen mentioned some issues inherent to digital therapeutics that make studying them difficult. Because of changes in operating systems and other software, digital therapeutics may need to be altered, making them a moving target for study. “It is not like a medicine — this is the molecule, this is what it is,” she said.
Taddei-Alen also noted that some digital therapeutics are designed to prevent an event or conditions. Evidence that an intervention kept something from happening is difficult to collective relative to, say, one designed to lower a particular biomarker.
Diabetes was Taddei-Allen’s first choice for a disease for which digital therapeutics hold the most near-term promise, with inflammatory conditions, such as rheumatoid arthritis, a close second. For Dunn, mental health was the first choice, and he noted that there are many prescription digital therapeutics on the market now that are based on cognitive behavioral therapy.
Rush said full-fledged, long-term randomized clinical trials for digital therapeutics would be enormously expensive. He advocated for an evidence base that would include real-world evidence research and retrospective cohorts.
Rush also said that successful pilot projects for digital therapeutics and their evaluation need to be careful about selecting a patient population and what problems (end points) they are trying to solve.
Access to the data generated by digital therapeutics and rendering it in a way that it is useful to payers and clinicians came up several times. Dunn addressed the issue, in various ways, several times.
“Can it (the data) be put somehow an through an API (application programming interface), or something into the systems that we already have because it is not feasible to expect our nurses or our pharmacists who are doing care management to have to remember or use 15 or 20 different systems,” said Dunn.
“It still comes down to…getting the data and getting it in a way that we can actually use it—and that has been a challenge so far,” Dunn said later in the discussion.