You’re reading the web edition of STAT Health Tech, our guide to how tech is transforming the life sciences. Sign up to get this newsletter delivered in your inbox every Tuesday and Thursday.
Omada Health announced it will launch a new service to help employers respond to the soaring demand for expensive GLP-1 weight loss treatments like Wegovy. When employees of participating companies seek to have their treatments covered by health plans, they’ll be asked to join Omada’s program that encourages lifestyle changes that might help people keep weight off.
Omada, which offers a host of virtual chronic disease management programs that focus on behavior change, won’t actually write or fill prescriptions for the drugs, a service that CEO Sean Duffy told STAT has been commoditized to the point that people don’t need more options.
Where Duffy thinks Omada can help is improving longer-term outcomes for people who choose to take GLP-1s. Though it’s unclear what long-term coverage for novel weight loss drugs might look like, Duffy said the company’s clients are leaning towards covering the drugs for a year or two. Many people who use Wegovy regain weight once they come off the medicine.
“If you’re not taking advantage of the positive inertia that a medicine like this can build with someone by inspiring a different way of thinking about habits, maybe a different relationship with food, you’re less likely to get sustainability,” he said.
Duffy acknowledged that while it’s plausible that behavior change programs like Omada’s coupled with drugs could yield longer-lasting benefits — or at least “stem the slope of the regain” — it’s a hypothesis without evidence at the moment. He said the company is committed to producing evidence that supports the claims in the long run.
Ultimately, Duffy said the company’s approach makes sense because it can encourage the change that’s written right on product label: “Wegovy should be used with a reduced calorie meal plan and increased physical activity.”
In related news: Consumer weight loss company Noom this week launched Noom Med, which offers GLP-1s to “medically qualified individuals.”
Microsoft health chief talks Nuance adoption, preaches gospel of low-risk AI
During his keynote at The New Wave of AI in Healthcare conference in New York this week, Microsoft’s global chief medical officer & vp of healthcare David Rhew laid out his vision for the opportunities for AI to improve care, which perhaps unsurprisingly included a good chunk of marketing for ambient clinical intelligence enabled by Nuance, which Microsoft acquired last year for $16 billion.
Pressed on the company’s progress on adoption, he said that the biggest driver has been hospitals trying to address clinician burnout. With growing excitement around AI, Rhew suggested that potential customers will be more open to piloting Nuance to see how it might work with their systems. In March, the company announced DAX Express, a version of its dictation tech that uses OpenAI’s GPT-4 to enable access to clinical notes faster after exams. Since then, he said, “demand has been through the roof.”
Earlier in his presentation, Rhew urged the leaders in the room to approach use cases for emerging technologies enabled by cutting edge technologies like GPT-4 with caution.
“Rather than start with the hardest one, let’s start with the ones that have a big impact but have the lowest risk,” he said. “… It’s almost like if we just got a pair of skis and we see that, wow, we can go down these hills, let’s not go on the double black diamond. Let’s try to master the bunny hill first.”
Overheard: Data challenges of AI
Following Rhew’s keynote at the conference, which was hosted by Mount Sinai’s Icahn School of Medicine (ISMMS) and The New York Academy of Sciences, experts wrestled with some of the biggest challenges of developing and deploying AI. A few choice bits I overheard:
- “If you think that you’re operationalizing race to get at some type of social or socio-structural something, then I would say put the socio-structural something in your model. Don’t think that you’re going to capture that. ” That’s what Emma Benn, director of the center for scientific diversity at ISMMS, told an audience member who asked whether researchers should include or exclude race and ethnicity as variables in AI models if they seem to have predictive value. Some experts suggest including that data in models can perpetuate disparities. Benn said there remains debate about how best to handle the issue.“When one assumes that, well, it’s racist that race is in the algorithm, so I’m going to remove race from the algorithm. I would say that it’s not that simplistic and the algorithm could still be racist,” she said.
- “I think sometimes it’s difficult to conceptualize pathology data for people that have not dealt with pathology data,” ISMMS computational scientist Gabriele Campanella told an audience member who asked about the scalability of designing AI models based on pathology slides that average 1.7 gigabytes and hundreds of thousands of pixels each. Mount Sinai has its own dedicated high performance computing cluster crunching one of the largest repositories of slides in the world — resources available to only a handful of institutions. But Campanella said one day it may be able to release foundation models that others can fine-tune for their own purposes.
FDA prepares rules for drug-related apps
The Food and Drug Administration is expected to release new guidance on apps designed to work with prescription drugs this year.
Drug companies already have products that are intended for promotional purposes, to support adherence, help patients manage their conditions, or provide dosing guidance. Martin Culjat, SVP and global head of digital medicine & regulatory innovation at consultancy Eversana told STAT that the forthcoming guidance, which could come as soon as this summer, will provide clarity to companies developing apps that have an “additive clinical effect.” That’s where the rules “get a little bit more fuzzy.”
Eversana is helping risk-averse drug companies with bigger ideas navigate the murky waters.
“The intention is often to have some kind of impact on patient outcomes,” said Culjat. For companies who want to do more with apps, “there’s a lot of uncertainty. Can you reference the app on the drug labeling? What kind of clinical trials do you have to run? Do you have to submit something as part of a new drug application?”
Right now, it appears FDA anticipates most drug-related apps won’t need regulatory approval.
In those cases, “only the output of these apps will need to be reviewed as promotional drug labeling,” Culjat wrote in an article on the topic. The new framework, however, “does not alter and is independent from the regulatory framework for medical device software.”
FDA started thinking about this long before the current explosion of interest in digital health: In 2018, it posted some of its thinking on the topic, and received comments from representatives from companies like Omada, and drugmakers Novartis, Otsuka, Takeda, and Teva.
Boston Scientific acquisition nixed
Boston Scientific is scrapping its $230 million plan to buy a majority stake in M.I. Tech, a Korean company that builds surgical tools. Instead, the company will buy around 10%.
“Our agreement to purchase the majority stake of M.I.Tech Co., Ltd, from Synergy Innovation Co., Ltd, required global regulatory approvals that we were not able to obtain in some countries,” Boston Scientific spokesperson Kate Haranis told STAT’s Lizzy Lawrence.
The Federal Trade Commission said the deal fell apart apart in response to “investigations by FTC staff and our overseas enforcement partners.”The medical device giant had agreed to purchase around 64% of the company last June. Boston Scientific focuses on minimally-invasive medical devices, and M.I. Tech would have contributed to its portfolio of non-vascular stents, devices that clear blockages in the body.