May 11, 2023

How FDA's pandemic flexibilities boosted mental health app development

AdobeStock_479650030-768x432

Adobe

Last year, I wrote about how a pandemic-era Food and Drug Administration enforcement policy had afforded mental health app developers the ability to experiment with their treatments in the real world without seeking marketing authorization. With the end of the official public health emergency today, these companies will need to either submit for FDA clearance by early November or pull them from the market. In a new story, I checked in with two companies, Big Health and Limbix, about their plans.

Big Health's two products, Daylight, which treats anxiety, and Sleepio, which treats insomnia, were actually available before the enforcement policy, but the company took advantage of the flexibility to more aggressively position the products as treatments. CEO Arun Gupta said that it has treated more that 150,000 patients since the beginning of the pandemic, which he attributes in part to the company's ability to make stronger claims about the products. Now Big Health must decide whether to keep the treatment language and seek clearance or scale it back.
Meanwhile Limbix, which is developing an app that treats adolescents with depression, took advantage of the flexibility to release SparkRx despite only preliminary efficacy data. Chief science officer Jessica Lake told me SparkRx has served patients in 39 states, with half of its fills coming from primary care providers and the highest fill rates from rural and community health centers, "highlighting existing access issues and the ability for digital therapeutics to fill critical gaps in care."
Lake also told me Limbix recently completed data collection for its pivotal trial and plans to submit SparkRx for regulatory clearance before the November deadline. Lake gave me an early look at the data, which is statistically significant.

Artificial Intelligence

Making insurance claims less of a nightmare

Rajeev Ronanki, formerly a senior vice president in charge of many digital initiatives at Elevance Health, has taken over as CEO at Lyric, where he hopes to help transform the confusing world of insurance claims into something more user-friendly with the help of hyped technologies like AI.

"We've been talking about transparency for consumers and all that for nearly a decade now, but without much progress," he told me.

Until recently, Lyric was known as ClaimsXten, a claims editing business which was formerly a subsidiary of Change Healthcare. ClaimsXten was sold to private equity firm TPG Capital for $2.2 billion to appease regulators scrutinizing UnitedHealth Group's acquisition of Change.

At the time of the sale, ClaimsXten was working with 9 of the top 10 insurers — missing only UnitedHealthcare — to help them pay providers accurately and efficiently, giving the company a great deal of visibility into millions of health care transactions. Ronanki thinks this provides an opportunity to improve the patient experience of claims, including that confusing piece of mail that says "this is not a bill."

"We have the rules and all of the knowledge of how transactions ought to be settled," he told me. "So we can use that knowledge, combine it with newer machine learning and AI technologies to really make life easier and simpler for consumers."

Ronanki imagines within a year or so the company might launch a large language model that's able to answer questions about a bill or potential care costs. The service would be made available to the members of health plan customers, but a version might be made more widely available. While Lyric will fundamentally remain a business-to-business service, he said, "the key beneficiary of seamlessness and eliminating friction will be consumers."

artificial intelligence

Google and IBM ramp up generative AI plans

Google is training its generative AI system, Med-Palm 2, on imaging data such as X-rays and mammograms to help it communicate with doctors about data routinely used in patient care. The move highlights the broader push to get these models accustomed to working with the many kinds of data — from genetic sequences to billing codes — clinicians must use in real-life care. Greg Corrado, a senior research scientist at Google, spoke with Casey about the opportunities arising from the rapid advances in AI — and the risks of moving too fast. Read the full story.

Speaking of moving fast, IBM, which flopped badly in the AI race that surrounded deep learning, unveiled a new platform called WatsonX to help companies in multiple industries incorporate generative AI into their businesses. The company's CEO, Arvind Krishna, said the lower costs of implementing large language models will quickly attract a wide variety of new users. Whether that will be for better or worse is anyone's guess.

Evidence

Data on digital CBT and diabetes

This week we saw new data from two different collaborations between digital health startups and more traditional health care players.

Magellan Health and NeuroFlow released data about their year-old Digital Emotional Wellbeing suite, which combines Magellan's digital CBT modules with NeuroFlow's software infrastructure that enables diagnosis, triaging, and patient access. Users who completed 75% or more of the FearFighter module saw an average reduction of 41% in anxiety assessment scores. The data is drawn from a sample of hundreds of users. Those are promising results, but the sample is quite selective, so we'll be watching for additional details when they are published later.
Next, Sanofi and DarioHealth released data showing users of Dario's Diabetes solution had a 9% reduction in all-cause health care resource utilization and a 24% reduction in hospitalizations compared to non-users. The retrospective cohort study compared 2,445 Dario users to 7,334 non-users matched using statistical methods. This is the first of several studies being conducted as part of Sanofi and Dario's strategic partnership.

medical devices

Lawmakers call out CMS on breakthrough devices

At a hearing yesterday, lawmakers called for a simpler reimbursement pathway for breakthrough-designated devices, echoing a familiar argument from medical device makers: that when companies struggle to obtain insurance coverage, it discourages innovation and limits patient options.

Republicans called out CMS for taking two years to release a replacement rule for a Trump-era policy rolled back in 2021. That rule would have guaranteed FDA-designated breakthrough devices four years of Medicare coverage. But CMS has a different, and often higher, data standard from the FDA on devices. STAT has found that the FDA's breakthrough devices program, meant to benefit patients, has delivered the biggest gains for device companies.

Just one voice called for caution in automatically granting breakthrough devices coverage: Rep. Lloyd Doggett (D-Texas), who stressed safety and efficacy concerns and raised questions about the quality of data coming out of breakthrough device studies. Read more here.

More around STAT