History and literature are replete with cautionary tales on the dangers of pretending to be someone — or something — you aren’t. Many digital health companies failed to heed the lesson, and the results have been predictable.
Technology has upended virtually every known industry and customer experience, health care included. Sadly, health care is at or near the bottom when it comes to providing users with good experiences.
Most people use their phones and devices day and night, and in every setting. Digital experiences influence social lives, purchasing habits, and nearly everything else. Given that up to 50% of what makes people healthy is directly attributable to their behaviors, technology should be able to help people develop and sustain habits that influence their health.
That thesis is obvious. Yet many digital health companies ignored it and instead pretended to be medicines. Of course, anyone’s grandmother could tell you that apps are not pills. You don’t swallow them, they generally don’t have side effects, they don’t interact with each other, and they are far safer than pills.
The allure of being treated like medicines, being prescribed and delivered like medicines, and, most importantly, being reimbursed like medicines was too seductive for many digital health companies to ignore. They insisted they were something they aren’t and demanded to be treated as such.
While there are a few exceptions, the results have been devastating for several of the companies that took this route: Pear and Better Therapeutics are recent examples of companies that suffered the consequences of pretending to be medicines. Akili shifted its model and has so far survived. The digital-health-as-a-medicine business strategy contained serious flaws. The formularies of pharmacy benefit managers were not structured to treat digital health products like medicines, payers did not have processes to reimburse them like medicines, and requiring prescriptions for these products meant that too many people had to be convinced of efficacy before a product could be monetized; payers, PBMs, prescribers and patients all needed to act before a product could be monetized.
Since many players in the digital health industry insisted on being treated as medicines, it stands to reason that they would also be evaluated as medicines. This has been the case for several years, with a number of meta-analyses and evaluations of digital health that used the narrow endpoint definitions traditionally used to evaluate medications. And digital health experienced this recently as well, with one popular report garnering attention for claiming that the therapies don’t work while essentially repeating the mistakes of prior misguided analyses.
A key tenet of health care is that design should be informed by an intervention’s intended use. Most digital therapeutic solutions are not intended to be replacements for medications, and are not intended to influence a single outcome, whether it is systolic blood pressure in hypertension, HbA1c in diabetes, or pain in musculoskeletal conditions. These digital solutions are certainly therapeutic (hence the term software as a medical device), but they aren’t medicines. Instead, they do what medications never could: influence the behaviors and habits of those who use them.
Good or better health is generated between doctors’ visits rather than at them. It’s built in the home, at the gym, in the grocery store, and at work. It’s the decision to go for a walk rather than to let Netflix autoplay the next episode of a TV series. It’s generated at the dinner table and in all the moments that lead up to what’s on that table. None of these are influenced by medications; all of them can be influenced by digitally supported interventions. Astute observers of digital therapeutics and health economics outcomes research experts have understood this for some time, with research indicating that evaluations of digital health need to differ from standard pharmaceutical outcomes measurement, in a concept termed fit-for-purpose evaluation.
The International Society for Pharmacoeconomics and Outcomes Research is a global nonprofit dedicated to advancing the understanding of health economic outcomes. Its recent newsletter noted that the strategies typically used for evidence generation for new medical products are not at all well suited for studying digital products.
Put simply, products that influence a range of behaviors, conditions, and outcomes should not be narrowly evaluated on point outcomes in a single condition. Instead, they should be evaluated on the clinical and economic value of improving those behaviors, conditions, and outcomes. Many behaviors, such as eating better, exercising more, and sleeping better, produce improvements across a range of conditions. To capture the value of these interventions requires measuring the impact of these solutions on whole health outcomes and the total cost of care, not just the impact on a single measure like blood pressure. Further, economic predictions loosely derived from estimates of the financial impact of influencing a single endpoint won’t capture the broad benefits of digital therapeutics, such as the ability to capture behavioral and health data; the ability to use those data to influence habits and health behaviors; to hyper-personalize interventions; and to make access to care easier, more equitable, and more affordable.
All interventions that could enter the market should be rigorously and independently evaluated for their clinical and economic benefits. Criticism is essential to innovation and progress. If health care is going to solve its fundamentally broken user experience problem and ever-growing inefficiency, digital health is essential. But to get there, solutions need to be measured on what they actually do, not on what some of them pretended to do.
Omar Manejwala, M.D., is the chief medical officer of DarioHealth, a digital health company that helps users manage chronic conditions.