Fortnightly Healthtech Update #30

New to me Vocalis Health uses machine learning to analyze breathing and vocal patterns for signs of disease.

Also respiratory-related, CurieAI continuously monitors breathing patterns for signs of deterioration. Good for conditions like COPD, the company is now seeing early success with COVID-19.

I touched on Binah.ai back in June. Forbes has a longer piece on the company’s vision and business model.

Not sure how I feel about this application of leading edge tech, but it was bound to happen: Sift Healthcare raised $3m to apply AI to healthcare billing (the revenue cycle). If it’s a way to help providers cut their processing costs and speed the flow of cash from payers that’s one thing. But then there are marketing messages like “The 10% of patients least likely to pay their bills account for 50%+ of outstanding patient payment bills, and pay less than 5% of their bills”. Don’t know about you, but that doesn’t make me feel all warm and fuzzy inside.

Delivery of prescription meds is potentially caught up in the USPS political wrangle. Uber partly steps into the void, partnering with NimbleRX to provide delivery. As noted, this solution helps to fill a COVID-19 shaped hole where virtual visits have become norm. That might persist, post-pandemic. Either way, if you’re in rural American without Uber coverage, I guess you’re out of luck.

Does not compute: Large employers expect to pay 5.3% more per employee for health insurance next year. OK, so telehealth usage is up. But….shouldn’t telehealth help to increase the productivity of clinicians, so dropping costs overall? We need to use virtual visits to re-invent how clinicians work with their patients to grow productivity. How can we use telehealth to enable a doc to consult with 5 patients in the time they would previously have seen 4? Or 10, in the time they would have seen 1…? That’s the promise and opportunity of telehealth. If all we’re going to do with telehealth is layer it over the existing models of medical practice, we’re just adding another layer of cost. Bloomberg has more on that topic, with management consultant Oliver Wyman suggesting the cost to deliver a telehealth visit is about half of an in-person visit.

Just to pull on that thread a little more, that would also be my concern with solutions like Vocalis Health, CurieAI, and Binah.ai at the top of this note. Med device innovation is great if we can get better outcomes. But, if we’re still working within a fee-for-service model, we’re adding on another layer of cost each time. In a value-based care framework, we should be able to get better outcomes and lower costs overall. My concern is that we’re just not setup to do that. Because it really seems to have stalled: There’s been no real change in the amount of revenue at risk in value-based care from 2018 to 2019. Almost half of execs taking part in the survey say 10% or less of their revenue is linked to value-based contracts. The risk of financial loss is what’s holding them back apparently. And there was me thinking that was the whole point.

Still plenty of business activity around telehealth, Amwell takes a strategic $100m from Google and files for IPO. With almost 3m telehealth visits in the first half of the year, can’t argue that it’s a good time for investors to cash in. Telehealth bubble, anyone…?

The clinical data on the early detection of sepsis using automated patient monitoring is mixed. The key quote for me is this one: “…as with manual screening tools, a patient monitoring system (PMS) will only be effective if the system has a high level of sensitivity and specificity, to engender clinician trust and reduce false-positive alerts. However, the nonspecific nature of sepsis makes achieving a highly predictive system difficult, whether on paper or in an automated PMS.” Sounds like a problem for lots of data and machine learning to me.

More Google, with Verily moving into the employer self-insured market. Many people worry about big tech’s use of data and possible invasion of privacy. Honestly, I’m not sure if those are genuine concerns, or just a smokescreen put up by those with an interest in maintaining the status quo. Personally, if Amazon or Google can drop my healthcare costs by oh, 10 or 20% a year, by making better use of my data I’m fine with that. It’s way past time my healthcare data was used to bring value to me, not just to create a margin for someone else.

And what about all those elective surgeries that have been postponed this year? Well, payers are making too much money now, so they have to give some back. Meanwhile, the Commonwealth Fund warns of an impending insurance crisis resulting from mass unemployment among other factors. Then, in 2021/22, I imagine we’ll see the backlog of elective surgeries landing, so premiums will go up again…

This is really the way ACOs were meant to work, but I can see this being a problem for people going forward. Cleveland Clinic and Aetna setup an ACO and suggest employers could save 10% over conventional plans. ACOs by definition offer a narrower network of providers. People with long memories will recall that was one of the concerns with managed care organizations from the past. Honestly, I think we’re at a point where there will start to be a much greater divergence in what insurance covers and the premium charged. We already have that in the different plans offered by employers, but I think it will become more pronounced.

A timely reminder that the practice of medicine is not the same as the business of healthcare: Hospitals still suing patients in coronavirus hotspots

Will consumer wearables ever become medical devices? Following on from the Apple Watch, and the Samsung Galaxy Watch 3 both getting FDA cleared ECG, is Fitbit – and perhaps the Amazon Halo. Fitbit’s latest has ECG, but is awaiting FDA clearance. The Amazon Halo meanwhile doesn’t seem to have any path towards FDA clearance yet. So, what’s the difference between a consumer wearable and a medical device? A smarter man than me explained it like this: It’s pretty easy to get pretty good data from most people, most of the time. But, it’s really difficult to get really good data from anybody, all of the time. That’s the difference between a consumer wearable and a medical device. Really good data, in all circumstances, whatever the body shape, age, gender, or race.

A longer read: McKinsey assesses the path forward for ACOs.