What’s this Button for? Digital Health is Being Held to a Higher Standard
Digital health (DH) continues its relentless march into the hands of patients, and into the heads of clinicians, purchasers, and payers. Despite great promise, DH also presents many practical stumbling blocks and open questions that trouble post-regulatory decision-makers. This may delay adoption of these technologies, frustrating consumers, and may sink companies entirely, frustrating financial backers. What’s going on?
Similar to ISPOR EU, Digital health (DH) was well-accounted for at this year’s international meeting. The title of Session 207 (“The Market Access Conundrum in Digital Health: Balancing Rapid, Robust Validation with Integrated Evidence Plans”)1 succinctly summarises one of the issues facing DH: how can one adopt DH when supporting evidence is either absent or dubious at best? Whilst many DH technologies have sailed through FDA approval, companies appear to be snagging on post-regulatory evidentiary requirements, quickly finding that the bar is higher than expected.
Whizzy startups that burn through money with questionable business plans is nothing new. However, for DH, having clinical value is their raison d’etre, making the lack of foresight on evidence requirements puzzling. That is, until you consider that a lot of these players are tech companies, not manufacturers, thus lacking the knowledge and patience in getting products to the market. It is possible that many outfits may be spending more time and effort trying to cut corners (rather than just getting on with data collection) and aiming for regulatory approval only (effectively ignoring the requirements of those that actually hand over money). These companies may also be focusing on today’s perceived minimum evidence standard, rather than the direction of travel.
Another important issue was raised: what is a “meaningful” change in a DH intervention that necessitates new evidence and re-evaluation? For pharma and other medical technologies, this is an easy question to answer, but is fraught for DH. This featured a (mercifully) brief hypothetical involving artificial intelligence (AI): if an approved DH intervention then integrates AI into its offering, is it still the same intervention? The perception of a material change is a lot more nuanced in DH, and small changes could easily lead to unintended, even harmful consequences. Maybe it’s not a bad thing that payers are reluctant to fund a lot of them?
On payers, what about the route to market and reimbursement? Costello Medical’s research explored value frameworks for DH and painted a fragmented picture across jurisdictions, with different frameworks assessing a menagerie of different characteristics.2 This nonetheless supports the view that manufacturers should be aiming high from the outset, rather than simply at the step immediately in front of them. Session 148 also explored this area, touching upon the US landscape, where in lieu of a national reimbursement pathway there is a fragmented and siloed patchwork of health portfolios.3 The endorsement of the Institute for Clinical and Economic Review (ICER) is widely seen as important in the US, recently collaborating with the Peterson Health Technology Institute (PHTI) to develop a joint assessment framework for DH. However, ICER has no mandate. So criteria on decisions surrounding procurement and funding among payers remain relatively opaque, and theirs alone. Though, if session 238 was anything to go by, or indeed anyone who has had to deal with a procurement process: being cheap helps.4
Overall, though, there appears to be a trajectory in the right direction, and payers are right to be cautious. Balance is key. There is a middle ground between a large RCT and a 10-patient case series. There are likely innovations in trial design; is the current model, skewed heavily towards pharma, necessary or even appropriate? Tackling these questions will be key in reducing outlay, complexity and shortening the path to market. Otherwise, companies might have to make the unpalatable choice between going broke conducting clinical studies versus going broke pushing an app with marketing authorisation but no clinical value. Tech companies, as a general rule, like to “move fast and break things”. They are quickly discovering this won’t wash for medical technologies. Is that really such a bad thing?
