Issue 113

Welcome back to E&O Fridays, a paying subscribers-only weekly newsletter focused on the world of digital pharma products and FDA-regulated digital health.

 E&O Fridays.

Here’s what’s happening this week in the world of pharma digital products and FDA-regulated digital health:

• Physical therapists came out in force to comment on the 2022 Physician Fee Schedule and ask CMS to let PTs bill for services using the forthcoming remote therapeutic monitoring codes. Almost one-third of the more than 2,000 comments on the fee schedule so far are form letters from PTs organized by groups like the American Physical Therapy Association. Each letter closes: “I respectfully disagree with CMS’ determination that physical therapists are not permitted to bill remote therapeutic monitoring codes. Although the services may be “incident to” when billed by physicians, they would not represent “incident to” services when billed by therapists. I encourage CMS to reevaluate their interpretation of these codes to permit physical therapists to bill and be paid for these services.”
• FDA cleared Abbott’s FreeStyle Libre 2 iOS app, which connects to the company’s continuous glucose monitoring wearable.
• The FDA also granted Breakthrough Designation status to Altoida’s SaMD: “The non-invasive software device which measures and monitors neurocognitive function is backed by two decades of global research, and is now being evaluated to predict conversion from Mild Cognitive Impairment to Alzheimer’s disease with artificial intelligence.”
• The UK’s Closed Loop Medicine said that its latest clinical trial met its primary and secondary endpoints. That tees it up to submit its drug+ digital combination product to UK and EU regulators. The combination product is codenamed CLM-IN01 for insomnia. CLM developed it in partnership with Taiwan-based Orient Pharma.
• Pear is hiring a number of sales reps to cover statewide territories in places like Texas, Alabama and North Carolina for its three PDTs. The company also launched a coupon program for people paying out-of-pocket for Somryst, its insomnia PDT: “Pay as little as $100, with the Somryst Savings Card, through an entirely digital prescription process that fits right in with an on-demand world.”
• One more thing… Interesting marketing move by Akili: The company bought a sponsored review on the popular parenting website Fatherly for its newly commercialized EndeavorRx PDT for ADHD.

Did a friend forward this to you? You, too, can win friends and influence people via newsletter forwards by clicking right here. Become the newsletter-forwarder we both know you were meant to be. Act fast! This offer won’t last forever (I am not immortal).

Baby step toward more billing codes for AI?

One more note on the agenda for the September meeting of the American Medical Association’s CPT Editorial Panel Meeting. One of the sections in the proposed agenda suggests — not a new code — but a discussion about how to define and categorize artificial intelligence in medicine:

• “Taxonomy of Artificial Intelligence (AI) for Medical Services and Procedures: Establish an Appendix that will improve the terminology and understanding of AI as it relates to the CPT code set”

This isn’t particularly important on its own but it got me thinking about the recently proposed 2022 Physician Fee Schedule, which had a lengthy section about AI that was filled with questions boggling the minds over at CMS — here’s a snippet: “We are also more broadly soliciting public comment to help us better understand the resource costs for services involving the use of innovative technologies, including but not limited to software algorithms and AI. Specifically, we are requesting commenters consider the following questions:

• To what extent are services involving innovative technologies such as software algorithms and/or AI substitutes and/or supplements for physician work?
• To what extent do these services involving innovative technology inform, augment, or replace physician work? For example, CPT code 92229 is a PE-only code in which the software algorithm may be substituting for some work of an ophthalmologist to diagnose/detect diabetic retinopathy. CPT code 77X01 is a service in which the trabecular bone score software may be supplementing physician work to predict and detect fracture risk. CPT code 0503T may be both substituting for, and supplementing physician work to detect coronary artery disease.
• How has innovative technology such as software algorithms and/or AI affected physician work time and intensity of furnishing services involving the use of such technology to Medicare beneficiaries? For example, if a new software algorithm or AI technology for a diagnostic test results in a reduction in the amount of time that a practitioner spends reviewing and interpreting the results of a diagnostic test that previously did not involve such software algorithm or AI technology, and if the software algorithm or AI could be considered in part a substitute for at least some physician work, it may follow that the intensity of the service decreases. It is also possible that a software algorithm for a diagnostic test that is supplementing other tests to establish a diagnosis or treatment pathway for a particular condition could result in an increase in the amount of time that a practitioner spends explaining the test to a patient and then reviewing the results.
• How is innovative technology such as software algorithms and/or AI changing cost structures in the physician office setting?”

Part of what trips CMS up in this area is that it has historically categorized software expenses as “indirect” rather than “direct” expenses for physicians. Not to get into the weeds too much, but when it comes to the AI and software algorithms CMS discusses above, it’s likely that the software itself is doing some of the direct work for practices. It’s notable that both the AMA’s CPT Editorial group and CMS are having these discussions. If the AMA group ends up establishing a taxonomy, it will likely lead to more AI-focused billing code proposals in the future. And if you’re not already aware of this (been a while since I linked to it): Check out this great database of FDA-authorized AI-based algorithms that The Medical Futurist put together.

Is Natural Cycles’ latest FDA clearance a first for consumer electronics as SaMD data inputs?

I mentioned Natural Cycles’ recent FDA clearance in Issues 108 and 109, but I still had some questions about it, so I managed to get a few answers this week from the company’s co-founder and CEO Raoul Scherwitzl.

What the clearance was for: In case you missed it, Natural Cycles offers an FDA-cleared app for contraception. The company’s initial De Novo clearance obviously broke new ground, but its recent 510(k) did too. Natural Cycles users have to use a basal body thermometer to take their temperature in the morning and manually input that data into the app. (These thermometers are FDA-regulated medical devices.) The new clearance added the Oura ring as a validated, automated data input for temperature. Thanks to this clearance, Natural Cycle users might be able to rely on their Oura ring passively sensing their temperature and sending that data to their Natural Cycles app.

When will the Oura-powered Natural Cycles app launch? Maybe never. It’s currently just a private beta. Oura offers an API and Natural Cycles made use of that for this clearance. It had no agreements in place with Oura as of the time of its new FDA clearance. The clearance marked the beginning of a process that Natural Cycles hopes to build on. The company is currently conducting validation studies of other wearable devices that have temperature-sensing features. It expects a validation study is all it needs to expand its latest FDA clearance to include more wearables. It doesn’t believe it will need to file a 510(k) for each new off-the-shelf data input option it validates.

Quality agreements will likely be necessary. Scherwitzl told me that Natural Cycles will probably need to put together agreements with the manufacturers of the wearables it validates as new data inputs. NC will need a heads-up about upcoming iterations of the devices that might change how NC functions.

Is this a first — having FDA unregulated wearables feeding data into SaMD? I’d argue it is. Here are three buckets that I see emerging in the category of FDA-regulated software meets unregulated hardware:

1. Same company makes unregulated hardware platform and FDA-regulated software that runs on it: Apple’s two De Novo clearances are the prime example of this category. Apple’s hardware — the Apple Watch — is not regulated as a medical device. The unregulated sensors on the Watch send data to Apple’s regulated software, the ECG App. Fitbit and Samsung have similar clearances for the ECG apps on their respective wearables. The software (ECG app) is running on the device or platform (smartwatch) that includes the unregulated hardware too.
2. One company makes an unregulated hardware platform and another makes FDA-regulated software that runs on it: This scenario is similar to the above but the regulated software is not made by the platform maker. NightWare might be a good example. The heart rate data and movement data from the Apple Watch feeds into NightWare’s FDA-regulated software, which takes action based on that data. Like the first example, however, NightWare’s software runs on the platform that contains the unregulated hardware, the Apple Watch.
3. One company makes unregulated hardware (not a platform) and another makes FDA-regulated software (running on some other device): This is the Oura-Natural Cycles category. I’m not sure I’ve seen this before. Natural Cycle’s software runs on a smartphone — not on Oura’s ring. Oura’s device is not FDA-cleared, but it is now validated and can send data into FDA-cleared software — replacing an FDA-regulated thermometer.

What do you think? Is that a first? Do you know of other unregulated sensor-equipped wearables validated as data inputs for SaMD running on another device? In any case, we are going to see a lot more of this. Read more about the Natural Cycles’ 510(k) and its validation study of the Oura ring in the FDA documents here.

Clinical trial updates: Verily-Parkinson’s, Woebot for GenZ, ResApp COVID Dx, Pear Somryst

Verily and UCB Pharma share details on Parkinson’s study

Remember back in May when the FDA rejected Verily’s bid to use its Study Watch for a Virtual Motor Exam for Parkinson’s disease in a clinical trial in the Netherlands? This week Verily and its partner UCB Pharma posted details of a two-year study in the Netherlands that will make use of the Study Watch with recently diagnosed PD patients. The Watch, however, will passively monitor 140 participants to help Verily develop

More on the study:

Woebot parallel assignment study pits its GenZ DTx up against PDFs for mild to moderate depression in teens

Woebot’s latest study will compare:

The control group is further explained:

Here’s the digital therapeutic description:

ResApp’s new COVID cough evolution study

This next study from ResApp might include as many as 1,500 participants over 25 days:

Pear’s DREAM study set to finish up sooner Small update: Pear Therapeutics changed its anticipated completion date for its chronic insomnia (Somryst) study from January 31, 2025, to January 31, 2023. That’s unusual — typically those changes are to push the date back.

Links to E&O’s reports, databases, newsletters

The Exits & Outcomes site is designed to make it easy to find long-form research reports, databases, and past newsletter editions. Click below for dedicated pages for each of those categories:

• Read through the long-form E&O research reports here.
• Search and sort the E&O databases here.
• Skim more than 150 past issues of E&O newsletters here.