By MATTHEW HOLT
The big news in the comeback of digital health is that Hinge Health filed its S1 and is looking to go public soon. I suspect that they’d have preferred to get the IPO done late last year when the AI bubble was expanding rather than deflating, but timing the market is tough! Nonetheless hinge is almost profitable and at over $350m in revenue at a growth clip of some 75% last year, in terms of a show pony to trot out, it’s about as good as the digital health field has got. The problem is that the last round in 2021 was at a $6bn+ ZIRP era valuation with Tiger & Coatue paying the idiot price because Teladoc was trading at $15bn market cap then (albeit down from $30bn a year before that!). That is, err, no longer the case. There’s a bunch of weirdness in the IPO structure to pay those guys back, but the main point is that the likely valuation will be in the $1.5-2.5bn range.
But there’s another problem. And it’s one I have some personal experience with. I must stress that my experience is not with Hinge.
As it happens I did a video interview at Hinge’s booth at HLTH in 2022 when my back collapsed, and I got to try out their Enso device (it helped a bit but not much after the first few minutes using it). I discussed the process with PT Lori Walter and got a quick interview with President Jim Pursely (an old Livongo hand BTW).
But this past summer I used the services of their main competitor, Sword Health. As far as I can tell the two companies are very similar in their process and services, both with self-service exercises delivered via the smartphone and both moving from remote care from therapists to AI therapists. But I could be wrong. So for this article I am extrapolating from one company to the other to look at the field of MSK digital services overall.
In total, I thought the Sword experience was good as a standalone program. But the problem was that it was standalone.
My problem was with my left knee. I had a lot of knee surgery in 2002-4 as the result of snowboarding into a tree (Hint. If you snowboard, try to make sure you and the board go the same side of the tree). More than 20 years later in 2024 I managed somehow to induce terrible pain in the knee running for a ferry in January, a train in May and an airport shuttle in June. (It seems that travel and my knee disagree). This didn’t stop me strapping up, taking drugs and snowboarding in the 2024 season but it certainly slowed me down a whole lot. Around this time there were many reports of people much younger than me getting their knees replaced.
So I thought I should do something about it. My Blue Shield of California plan offers Solera which is an agglomeration marketplace of digital health apps and services. Sword Health is their PT app, so I selected it, enrolled and off I went.
Note that there was zero integration with my PCP, any orthopedic surgeon, any clinical person at the health plan or basically anyone. This was purely patient-driven and managed.
With Sword I had a 15 min intro call on June 6 – then was sent a box containing a generic tablet and six sensors which fit into straps that you attach to your lower and upper legs and arms.
There was a conversation in the app with a PT and then it spat out a selection of exercises for me. The example below is my second exercise session. If you want to check out more, I have put more of the exercise and the chat with the PT here.
Sword suggested, instead of regular 45-60 minute physical PT sessions, that I did four 15 minutes sessions a week. Essentially one every other day.
The end result was that I did eight sessions between June 12 & June 30.
I then lost motivation and took 10 days off, but then got back into it and did another nine sessions between July 10 and July 29 (just over three a week). Most of the time the system worked fine. For some reason in their deal with Blue Shield was sending out the sensors plus a dedicated tablet device, rather than letting me use my own phone for the video-based tracking that is now standard for Sword, Hinge and their competitors. There were times the sensors had problems being recognized by the device and one or two sessions I cut short as the data I was sending back was wrong. But in general the technology worked well and I suspect the video-based monitoring is even better.
All the while there was some discussion with my (non AI) therapist– actually two as the first went on vacation. This was pretty limited, but the therapist asked me some questions, I was able to send some photos of my knee to identify where the pain was, and she made some adjustments to the exercises per my requests and reporting back of pain levels. Obviously there was no physical manipulation (a big part of the PT I have had before) and no electro stimulation. Possibly Hinge with its Enso device is doing some of that now.
Then I stopped the program as I went on a vacation in Europe–where I spent way too much time walking around museums and big cities. I actually felt a lot better after all that walking. When I came back I had vague intentions of starting the PT again, but I never did. My therapist sent me one message in the app trying to get me to start again, but after that I never heard from Sword. They didn’t even ask for my equipment back.
So what was the result? I had 17 sessions over a 7 week period.
Clinically my range of motion increased, my ability to do a little weight bearing improved and by the last session I was doing 20 minutes of exercises. As I mentioned, having had a fair amount of knee surgery I have done plenty of PT. My conclusion was that 1) this gave me a program I more or less stick to for some time, which was good, and 2) it tracked range of motion and improvement. In other words it was a level above me following along to an exercise on Youtube.
But after I came back from vacation, I fell off the program. Then my knee slowly got worse. This didn’t stop me going back to snowboarding on it. It works but it hurts.
In the fall I met my friendly neighborhood UCSF knee surgeon at a party. He played with my knee and said, come see me and get it replaced. Meanwhile the greatest downhill skier of all time Lindsey Vonn got her much surgically altered knee (partially) replaced with an implant and is back on the World Cup circuit at the age of 40.
Of course I am not Lindsey Vonn and there is also the apocryphal tale of Eric Topol’s knee replacement which went horribly wrong and ruined his life. And I know what knee surgery (and the recovery) is like. So I am still pretty unsure as to what to do.
So what did this all cost?
Solera (which I assume takes a cut and passes a percentage onto Sword) billed Blue Shield of CA four times. Once was a set up cost of $497 billed the day before my first intro call. Then there were 3 subsequent bills, on 6/16 $300, 7/11 $310, 7/13 $300. There was no patient copay of any kind.
Blue Shield paid all of these bills with no discount–clearly a pre-agreed rate. I was trying to figure out what the trigger for the billing was. It looks like the first $497 is for me signing up, and the next bills for a specific amount of sessions.
But the total bill was $1,407. For this Sword provided one introductory call, many messages with my PTs (but those probably took a total of 30-45 minutes, less if they have pre-written or AI-delivered answers) and 17 actual sessions.
That’s $82 a session (assuming that the cost of the equipment is built into the session as it is at regular PT clinics). My rough asking around is that the cash price of a regular in-person PT session is between $75 and $125. So it doesn’t seem that my self service session is much, if any, cheaper than if I had traveled to a specific in person PT provider. Bear in mind that these sessions ranged from 12-20 minutes and I had 2-4 a week. Usually even post-operative PT sessions last closer to an hour (with one PT running 2-3 patients per our) and get phased out relatively quickly–you may get 2-3 a week for 4 weeks, then one a week then two a month. Meanwhile for a 15 minute session of “Therapeutic Exercise” (CPT code 97110) Medicare pays $28.82!
To me that means that Blue Shield spent roughly as much as if I had a short course of in person physical therapy. (Which I probably wouldn’t have gone to!).
What I do not know is what would have happened if I had kept going after I came back from my vacation. Would Solera & Sword have been able to keep billing Blue Shield? Or is there a capped amount per patient. No way to tell unless either side is prepared to show what their contract says.
The other problem with my care
What I did not like, and regard as a big problem with Sword, Hinge and the rest, is that there was zero integration with the rest of my care. My PCP was unaware that I was doing this. I have no current orthopedic surgeon (other than my UCSF party-going friend). And Sword made no effort to connect me with anyone else on their team other than the PT, who gave up on me pretty quickly after I gave up.
They may have reported this back to Blue Shield as a win. As a patient I did the course of treatment, ostensibly improved and did not result in a claim for surgery in the next 6-9 months. But they have no idea whether I was seriously considering surgery, or whether my improvement lasted.
And will I get my knee replaced? My UCSF friend says, get it done and get back to your favorite activities. I am sort of doing them, but in pain. But I spoke to a European orthopedic surgeon who said, our indication is the replace the knee if the pain is so bad the patient cant sleep overnight. I’m sleeping fine!
The IPO and what’s next
What does this all mean for the Hinge IPO? Assuming that their deals with plans and employers are similar to Sword’s I see two big challenges.
The first is integration with the rest of the health system. You’d think that a service like Hinge or Sword would work best in some kind of integrated medical group that had some kind of capitated payment, and was able to sub out expensive human PT with cheap virtual PT and see a reduction in expensive surgery as a result. Right now, at least in my n of 1, there’s no way of telling what happened next to the patient post virtual PT program. Did I have surgery? Did I see a surgeon. What happened? No one knows. Presumably if I am an end-payer I would want to know, at least eventually.
The second is on price. Again this is n of 1, but it seems to me that $1,400 for 2 months of virtual PT is a hell of a lot. Assuming Hinge is roughly price competitive with Sword, how long before this gets knocked-off at a much lower cost–especially if you can replace the virtual therapists with AI. We’ve seen a similar cram down in cost in other areas of health AI, notably in ambient scribing. It’s hard for me to imagine that Hinge and Sword can keep this price level. And if they can’t, one can imagine that Wall Street might get grumpy.
Matthew Holt is the publisher of THCB and his knee still hurts!
