Digital therapeutics after Pear: separating the clinical evidence from the payment model

In September 2017, the US Food and Drug Administration authorised reSET, a smartphone programme from Pear Therapeutics for people with substance use disorders, to be used alongside standard outpatient care. According to the company, it was the first time the agency had cleared prescription software with clinical data showing improved outcomes in a disease. The programme ran for twelve weeks and delivered cognitive behavioural therapy in interactive modules, with a dashboard the clinician could consult.

Cognitive behavioural therapy, or CBT, is a structured psychological treatment in which patients learn to identify automatic thoughts and behaviours and then change them, session by session. It is a first-line treatment for insomnia, addiction and several anxiety disorders. Its main limitation is access: it requires trained therapists, and there are not enough of them. The founding idea of digital therapeutics was that a protocol this structured could be coded, tested in clinical trials like a medicine, and delivered at scale.

Five and a half years later, in April 2023, Pear filed for Chapter 11 bankruptcy protection. The story of how both events happened to the same company is the subject of a French-language episode of Tech Anatomy's ERROR 404 series.

What a digital therapeutic is, and is not

The term covers a narrower category than "health apps". Meditation, step-counting and sleep-diary apps can be useful, but they are sold as wellness products and nobody had to prove they treat a disease. A digital therapeutic, or DTx, is defined by three features: its effect has been measured in clinical trials, a regulator has authorised it for a specific indication, and it is often prescribed within a care pathway. On a phone screen the two look alike. The difference lies in the evidence and the authorisation behind them.

Most products in the category digitise CBT, which explains why they cluster in sleep, addiction, attention and mental health, the areas where CBT is already standard care and where therapists are scarcest.

Three authorisations in three years

Pear followed reSET with reSET-O for opioid use disorder in 2018, and Somryst for chronic insomnia in adults in 2020. Somryst delivers CBT for insomnia over six to nine weeks, including sleep restriction adjusted by an algorithm, and was one of the first products reviewed through an FDA pilot programme designed for software.

Pear was not alone. In June 2020, Akili Interactive obtained FDA authorisation for EndeavorRx, a video game prescribed to children aged 8 to 12 with attention deficit hyperactivity disorder. The child steers a vehicle through obstacles, and the attentional training is the treatment. The authorisation rested on five clinical studies involving more than 600 children, including a randomised controlled trial.

In December 2021, during the surge of investment in remote care that followed the Covid-19 pandemic, Pear went public by merging with a SPAC, a listed shell company created to take a private firm public faster than a conventional listing. The transaction valued Pear at around $1.6 billion, according to the deal documents.

Where the model broke

The FDA had authorised the products. The difficulty was payment. In the United States, regulatory authorisation and insurance coverage are separate decisions taken by separate actors with separate criteria. A prescription digital therapeutic was neither a drug, handled through pharmacy benefits, nor a conventional medical procedure, and insurers had no established category for it. At the end of 2021, Pear reported agreements with around thirty organisations covering more than 26 million lives, according to the company. A coverage agreement, however, is not the same as claims being paid at scale.

In April 2023, about sixteen months after its listing, Pear filed for Chapter 11 and announced it would sell its assets, which went for a few million dollars. The filing itself gave no cause. Corey McCann, Pear's chief executive, who stepped down at the time of the filing, pointed to payers in a LinkedIn post quoted by Behavioral Health Business: "[Payers] have the ability to deny payment for therapies that are clinically necessary, effective, and cost-saving." Akili, which had also gone public through a SPAC, was sold in 2024 to Virtual Therapeutics for $34 million, well below its peak valuation.

Pear's bet was reasonable when it was made. FDA authorisation is a demanding standard, and a company that cleared it had grounds to expect that payers would follow. What the episode describes is a system in which that expectation did not hold for a new product category.

Competing explanations

The first reading, the one the episode develops, separates science from economics. The clinical case was made: the products were authorised on trial data, and the bankruptcies did not withdraw that evidence. What failed was the payment infrastructure, which did not exist yet. The technology was ahead of the financial plumbing.

A second reading gives more weight to adoption itself. Physicians had to start prescribing a type of product they had not been trained on, and patients had to download, open and complete a programme over several weeks. In this reading, payment was a necessary condition, but even full coverage would not have guaranteed use.

Both explanations are compatible with the record. What distinguishes them is what they predict for markets where payment is solved.

What the comparison with Europe can and cannot show

Europe offers something close to that test. In 2020, Germany created the DiGA pathway (Digitale Gesundheitsanwendungen, digital health applications), under which a physician can prescribe an authorised app and statutory health insurance pays for it. A fast-track route allows provisional reimbursement while evidence is still being gathered. Dozens of products have been listed, most of them in mental health, and prescriptions have grown year on year, according to an analysis published in npj Digital Medicine in 2024. France followed with PECAN, an early-coverage scheme introduced in 2023 that funds a therapeutic digital medical device presumed innovative for one year while the data for standard reimbursement are collected. In the United Kingdom, Big Health's Sleepio, a CBT programme for insomnia, received guidance from NICE in 2022 covering both clinical effect and cost-effectiveness.

These schemes show that payment can be designed in advance. They do not, on their own, show that digital therapeutics are widely used or that they improve outcomes at population level. Prescription volumes measure uptake, not completion or benefit, and the schemes are recent.

What to watch

The useful indicators are the ones that separate the two explanations above: how many DiGA and PECAN products move from provisional to permanent coverage once their evidence is reviewed, how many prescribed programmes are actually completed, and whether a US payment category for prescription digital therapeutics emerges. If products with secure reimbursement still struggle to be used, the adoption reading gains ground. If they succeed where payment exists, the Pear story will read mainly as a problem of timing and infrastructure.

Listen to the episode

ERROR 404 - La folle histoire des DTx

Listen · 16 min