Abbreviated show notes below. View all links and the full notes in our blog post about this episode here.
ACA Preventive Care mandate challenged
SVB / Banking / Private Investment in Healthcare
John
ATA2023 quick recap
Elena
Fatma
TEFCA and QHIN update
Zus Health Funding
Jody
Meta Pixel vs HIPAA
AI / ChatGPT
Abbreviated show notes below. View all links and the full notes in our blog post about this episode here.
ACA Preventive Care mandate challenged
SVB / Banking / Private Investment in Healthcare
John
ATA2023 quick recap
Elena
Fatma
TEFCA and QHIN update
Zus Health Funding
Jody
Meta Pixel vs HIPAA
AI / ChatGPT
John: [00:00:14] Welcome back to the Chilmark podcast. Today we are introducing a new format for podcasts, which will be a monthly review of any of the significant news and events that have happened over the previous four weeks since the last podcast. Today we have on the call Jody Ranck, Fatma Niang and Elena Iakovleva, who will be discussing their recent research and the news that has fallen into their various categories of expertise and what they look at following market. On today's podcast, we have Jody Ranck discussing the role that the megapixel lawsuits are playing around driving a new privacy discussion around patient rights to their data and how it is used and how it is disseminated to other organizations, especially because the way that this tool was implemented exposed a lot of AI in a context that was not actually under a business associate agreement like it's supposed to be due to HIPAA. He will also be discussing new developments in the digital therapeutics space because of some recent developments that have shown that this is a fairly weak business model and it's just not been great going for the companies that are trying to develop solutions in that sector. Elena will be discussing telehealth in 2022 versus 2023 and kind of how the dialog has shifted.
John: [00:01:33] She will also be discussing how more retail players are entering the space like Best Buy and Whole Foods. And she will also be providing some insights into her upcoming Hospital at Home Report. Then we have Fatma who will be discussing updates to TEFCA as well as the recent funding round for Zus Health. Before we jump into that though, as a team, we'll just discuss some of the general topics that have affected the industry over the last week or two. We will start with a quick discussion about a recent lawsuit that was filed and decided in Texas around whether or not the preventive care mandates in the ACA are actually legal and constitutional. So to start the day before we recorded this, a Texas judge decided that private insurers can use religious exemptions to no longer have to cover preventive care that was covered or initially discussed in the ACA and mandated by the Preventive Care Task Force. This includes cancer screenings, prep for HIV prevention, as well as other basic preventive care. So does anybody want to share some thoughts around what impact this might have on accessibility of care as well as care outcomes going forward?
Jody: [00:02:49] Okay, I'll weigh in. I think cutting access to preventive care. I'm a public health guy, so we're all about prevention. And, you know, this is quite, quite troubling that, you know, the use of religion to deny access to care screening and diagnostics and prevention. That's that's medical care. So I have a long background in health and human rights. And it almost goes back to being a human rights issue in preventing access to medical care and using religion to do that. And I'm not and I'm also not sure what Jesus would think about that if you're using religion. So those are my thoughts on on it. It's a very, very troubling, problematic development that, you know, definitely has a lot of other things going on around it and playing politics with people's access to health care, pure and simple.
Elena: [00:03:45] And I will probably insert my $0.05 as well. Um, I just see it as a global trend. I don't know why exactly, but preventive care in general getting slimmer and slimmer. And it just my heart breaks to see that happening. And not that long ago, probably 8 or 7 years ago, they made this major cut on ob gyn annual screenings. So instead of annual screenings, they allowed to have just 1 in 2 years. And it did affect health of multiple people. Just who I am closely talking to, generally speaking, the whole basis and foundation for preventive care is just to cut the costs of the treatment. So it's much cheaper to actually pay for the screening, right? So preventative manipulations rather than to cover the entire cost of treating a specific condition that was developed due to lack of preventive screening. So I'm not sure how they look at the whole model itself, but it just highly concerning. It's either we're interrupting something just from the wrong side or those policy changes just directly happening in some someone else's interests.
Jody: [00:05:04] But can we say that it's a general trend to cut preventive care when a big part of value based care is to prevent readmissions and hospitalizations? I mean, doesn't it totally contradict a broader trend? I mean, you know, going back to the probably late 80s, early 90s, there were some economic health econ studies showing that prevention actually didn't save money. And then there's a big debate over how the studies were done and so forth. And now, you know, value based care, it's all about identifying high risk folks and intervening with preventative measures to keep them out of the the hospital. Right. So it seems to me it's a bit of a you know, there are a lot of contradictory things going on when it comes to prevention.
Elena: [00:05:56] 100%, Jody. So that's pretty much what's the main focus is on that Logic is suffering. So it just doesn't fit the general model of how we think of health care in the United States. And that's concerning.
John: [00:06:10] Yeah, it's a bit ironic considering that there has been a notable uptick since the pandemic around the shift to value based care, because, you know, one of the only areas that we're seeing profitability around care services is in the Medicare Advantage market. We also saw during the lockdowns that organizations that had a decent amount of their services prepaid through these contracts, they did a lot better than their peers who were entirely dependent on fee for service when all the hospitals kind of got locked down and prevented elective surgeries and elective procedures all completely got shuttered. So it defies logic that we would be moving backwards in this way, as the entire industry is trying to shift more towards value based care, because preventive care is a core foundational piece of that, like Jody was just saying. And it's also just really odd that an employer would want to gut preventive care measures for their employees considering that that's just going to raise their premiums over time. It's going to raise their their costs because people are going to be going in later after they've already come down with something instead of having that screening and those those kind of pre-diagnosis tools that are currently covered by preventive care, it just I don't see any rationale for an employer to want to do this. It's totally ludicrous to me. The cost of treating HIV, the cost of treating cancer, you know, if it's caught later in the process, it's so much higher than if you're just doing constant preventive measures and implementing best practices around keeping people healthy.
Jody: [00:07:41] And then there's the issue of going beyond cost to compassion and empathy for our fellow humans, which is why a lot of people go into health care.
John: [00:07:52] Yeah, but when we're talking about employers, they don't give a shit. At least some of them. Sorry. Maybe we should cut that one.
Jody: [00:08:01] Other than when they get the bills and people aren't at work and taking care of children and things like that. So yeah, us is full of lots of very expensive. Non-logical contradictions. I think that's what we can say about this one.
Elena: [00:08:19] Actually, I just had an idea. So maybe all that actually cleans up the field for the retail, health and digital health separate just because when you're told like, okay, dude, you're not going to get any any pre-screenings you're not going to get anything, you're not going to get any tests, then you're more in charge and you'll be looking for other options where you can get tested, when you can get all those preventive care options. And there is like a realm of companies stepping into this market, like from labs to screening tests, like let's get checked, Quest Diagnostics, developing the whole line of products to patients directly surpassing providers. So maybe it's just their golden hour.
John: [00:09:07] Yeah, it's definitely an opportunity for the direct primary care and retail side of things to gain some market share. The problem is, is that those services, generally speaking, are catering to a better off population, or at least a population with more means than the people that really need the free preventive services.
Jody: [00:09:27] Exact shifting euphemism for consumer focused health care, cost shifting to those You know, the whole notion of the patient is now the payer. And at what threshold? You know, you know, the usual discussions about medical bankruptcies and so forth come forward at some point when you start factoring in inflation and if you have a recession on top of that and then, you know, the last 10 or 15 years is shifting medical costs to people, at what point do you break their backs?
John: [00:10:04] Exactly. I'm just hoping that the feds appeal this because, I mean, it's a huge blow to the moon, to the Cancer Moonshot effort, because, I mean, one of the biggest things about reducing cancer risk and reducing cancer deaths is screening for cancer.
Jody: [00:10:21] And then at one point you mentioned Medicare Advantage and to Elaina's point about cutting preventive care and so forth. We had news in the last week of the use of algorithms to cut people's Medicare Advantage benefits and so forth, and a single doctor denying an enormous amount of claims, I forget the exact numbers, but, you know, is virtually impossible for a human being to deny that that number of claims and adequately review them. But using algorithms to automate denials, increase denials and so forth. So I know we're going to get to this later in our discussions about Chatgpt and so forth. But this whole algorithmic society as it impacts health care is becoming a really big problematic area of it doesn't get enough attention.
John: [00:11:16] All right. Next up, we've got hims in a few weeks, so we want to make sure that we prepare our listeners for what to expect from the event. I know personally, after going to and following the news, I'm very much curious to see how much of the conversation is dominated by the themes that we've been writing about recently. And then there's also this general shift, like I was saying, more conversations. So talking to people on site and seeing exactly what is driving that beyond the COVID pandemic fallout, because I feel like there's something else going on that's finally getting a lot of these provider organizations to take BBC more seriously. And maybe that's the shift of retail providers getting in. Maybe it's more of the burden of the census falling to patients, but I'm definitely curious to see how that plays out.
Elena: [00:12:01] Yeah, so I am going to hymns this year with my own agenda as I did last year. And precisely I'll be looking for those provider payer platforms that are emerging on the market. And I do have high hopes that I'm going to see some solid progress in these regards because I am a huge advocate for patient rights when it comes to medical billing, and I am expecting this new technology to have a massive effect on how affordable health care can be for patients in the United States. That's probably the major one, and the secondary is going to be remote patient care and how it's evolving, new initiatives, new products on the market, new technologies. And probably what's more important is how those providers are making partnerships and acquisitions recently so that we can understand their growth strategy and how they see the future market that's going to be shaping health care in the next five years.
Jody: [00:13:06] And I'm going to be taking a look at any of the announcements I see about new AI models, things coming up for FDA approvals in the AI space. And I'm also constantly following the these sort of health care platform players from Mayo Clinic platform and beyond. And then the role of AI in all of these platform business models, because we've seen how that's played out in other sectors. But I think it might have a slightly different it has more complexity in health care. And then the second area is going to be robotic process automation, RPA, which I'm currently working on a report for. So meeting as many of those vendors and kind of mapping out the whole range of offerings and, you know, the general market and market space understanding that better and diving deeper into it. So, Fatma, what are you looking at?
Fatma: [00:14:05] Thanks, Jody. So my primary interest is kind of feeling the temperature towards attitudes regarding the information blocking rule last year towards the end of 2022, when I went to conferences, there was a lot of outrage and skepticism about how it was going to play out. So I want to kind of assess the field and see if that attitude is still consistent across the board. And secondly, I just want to talk to all the guys with the APIs and learn more about what their APIs do in general.
Jody: [00:14:34] The interesting to take a look at APIs in some of these platform plays as well because their central component, yeah, I'm.
John: [00:14:42] Still waiting for there to be a Zapier for health care that actually aggregates all these different APIs and lets you set up your own triggers and actions. Okay. So one of the big things that happened this year or this past month, obviously, was the collapse of Silicon Valley Bank, which is the primary bank for the innovation economy. Obviously, as the investigation has proceeded, we have learned that their banking practices were. Fairly questionable. They were fast and quick with making new accounts, and there were a bunch of other things that they had done that are now coming to light that are a little bit sketchy. That said, they were one of the biggest banks for startups and for the VC community. And so we have seen a massive pullback on pretty much everybody that's in the investment community really moving forward with investments while they wait for the fallout and wait for the dust to settle from this and make sure that they're actually secure in their finances. The government did say that they're going to back SVB and all the depositors. So there's not really much risk right now. But the tantrum that the investment community threw around the collapse of this bank when they were the ones that precipitated the run on the bank has been absolutely astounding to watch.
John: [00:15:59] There was a great article in Slate that Jody had shared with me that just kind of lays into the absolutely childish behavior of a lot of this community, which was pretty unfortunate to see play out live. It was just not a great look on the investor community. And I think that as we look at the slowdown in overall investments in the health tech space, we really need to think about whether or not VC and PE money were the primary objective is financial return, and value extraction is really what we need in the health care sector. Is this really just contributing to higher costs and other kind of economic problems that we look at in health care, or is there a true need for private money to help support innovation in this space? If you look at the investments, the only real kind of broad category of investors that have remained active over the last two years is the corporate VCs. So VCs that are actually aligned with corporate interests that are already out there and were the sole goal is not to just to have a ten X return on the initial investment. So hopefully we can see a bit more of that thesis and thinking shift over into the VC side of things and the private equity side.
John: [00:17:16] Granted, there are plenty of venture capitalists out there that have the best interests of the public in mind, and there are plenty of funds that I have nothing but respect for. But as a whole, we've seen a lot of people enter the space that are not healthcare natives that are not privy to what this industry needs and how health care actually has evolved over the decades. To understand where the pain points are that truly need addressing. And these are the actors that we need to be careful for. I mean, it's things like the cerebral fallout and some of the other cases like that that really give this whole industry a bad name and deteriorate the trust that providers and health care systems have in new innovations. And so as a whole community, we really need to take a look at ourselves and identify how we can filter out the bad players that are really just in this solely for finance reasons and not actually here to help improve the quality of life, improve health care in this country while making a dollar on the side or, you know, making money off of that. Do you guys have any thoughts on that? Anything to add around the collapse?
Jody: [00:18:17] I think well before SVB collapsed. The economist Mariana Mazzucato has been writing about this, this whole, you know, basically the whole libertarian side of the VC world. You know, that's who we saw screaming bloody murder and then get bailed out in a way. But long before that, Mariana Mazzucato wrote about how if you take the iPhone, for example, 85% of the technology in an iPhone was originally funded through government grants, whether it's NSF, DARPA and so forth. So what I think the broader trend is we have the the foundational R&D slash innovations are funded by the taxpayer and some government bets that did very well in creating these innovations. And then the profits are privatized by Apple and on down the line. And VCs who aren't often the earliest investors, they invest kind of midstream and or further downstream. So the whole the mythology around Silicon Valley and so forth needs to be taken apart because the idea that who are the risk takers and who are the profit takers, it's not that fair to the US taxpayer in many ways. So that, you know, there's a bigger problem kind of further upstream in the way markets work, the mythologies around justifying who are the winners and losers and all of this, and often it's the taxpayer and and then we can connect that. I mean, I'll talk about this later, but, you know, you have a market of health data and and data brokers now that are collecting all the data that's created through apps and so forth. And what we're getting is surveillance, capitalism and concentration of capital and all of that. So I think it's a it's a little window that got a lot of attention, but it's a little window into a much bigger problem in my book.
John: [00:20:33] Exactly. That's kind of I guess what I was trying to get at is that this is just one small slice. It's exposing a lot of the bad activity that's been going on within that private investor community.
Jody: [00:20:44] And then the private equity impact on rural Americans, access to health care, closing all these hospitals and your average rural American lives further and further away from a hospital. Is that right? Should that be regulated? Can that be stopped? And private equity isn't serving the common good there. It's serving private investors over the common good and the health needs of rural communities and beyond.
John: [00:21:13] There's a great podcast on Fierce Healthcare's website where they were discussing exactly that after a new report came out showing the full extent of how negative of an impact has had on the rural health care market, as well as the nursing home market back during the pandemic. And the very obvious difference in outcomes between private equity owned nursing homes and care facilities versus the other ones, you know, privately owned or nonprofits.
Jody: [00:21:39] I am just surprised that we haven't had hearings about this in in Congress and so forth, because it's becoming a serious public health issue.
John: [00:21:49] At the beginning of the month, I was lucky enough to attend ATA, which happened to be in my backyard. I recently moved down to Texas and it was held in San Antonio this year, which was a beautiful city that had not yet visited, had a lot of fun. So one of the big takeaways from ATA for me personally that I was not particularly aware of, I had seen news about it, but I wasn't really fully aware of the full extent. Was the new DEA proposed changes to tele prescribing and the requirements for basically any scheduled drugs to You have to go in person to get your refills and to get your prescriptions filled, which is a huge barrier for a lot of patients to actually make sure that they are staying consistent with their medications. As somebody with ADHD, I can absolutely attest to that. I strongly benefit from being able to just call up my doc and let them know that I need my refill sent into the pharmacy. And it was actually pretty sad to hear some of the cases that doctors were getting up and discussing and talking to in front of the audience around exactly how their patients were freaking out and panicking about the impact of this new legislation that the DEA is trying to enforce because they know that it is going to substantially limit their access to their medication. It started out with a focus on trying to reduce opioid overdose and diversion, and it extended into the stimulant category.
John: [00:23:09] And it's just unfortunate because as they quoted during the session, less than 1% of actual prescriptions are diverted to illegal uses. So all that this legislation is doing is punishing the 99% plus of the rest of the population that actually uses their medications as they are indicated. So it was nice that they actually had a DEA regulatory person on site that's helping with this rule to hear these case studies. But we have yet to see how it may affect and impact the final rule that they are developing around how we can continue to have access to medication, enable digital and tele prescribing, especially as Jody just pointed out, with the rural facilities where I mean, a lot of people in this country do not have convenient access to care. So making them go in for an appointment is a substantial lift beyond just if they're in pain, it's hard to move or if you have ADHD, you forget. I mean, it goes beyond just personal considerations. This is a very practical problem that needs to be fixed, and it's all tied to the kind of the shutdown of the public health emergency, because that's what these provisions were initially implemented for, was the public health emergency when people couldn't go to the office. But it's been a huge boon to patients everywhere. So we'll see where that goes.
Elena: [00:24:27] Okay. So John, thank you so much for pronouncing my last name so accurately in the beginning of this podcast. I'll say it one more time. My name is Elena Yakovleva and I'm a research analyst with Chilmark, so here are my $0.15 for today's podcast. I would like to speak a little bit about the previous ETA, which was happening in my backyard in Boston. It was 2022 and it was pretty remarkable in a bad way just because we all were observing this rapid fall of telehealth market. So all everybody were just walking around saying like, finally we see the bubble bursting and everybody were just so down and it was pretty empty and people were not as active as they were expected to be. During ETA, the exhibition hall was like 10% full. I would say a lot of vendors, they were not, you know, closing as many deals as they were hoped for. Nevertheless, it was quite remarkable to see the progress in the past, what, ten months the last ETA was in May. So what are we seeing now? Telehealth is back to where it was practically. And I've been talking to multiple vendors through the the lens of my hospital at home report that is spoiler coming out soon. So and all those folks, what are they telling me? They're like, oh my God, we have so much work. Like, we're not planning on doing any more functionality, we're not planning to switching markets.
Elena: [00:26:10] We're just trying to keep on growing and just to meet the demand for telehealth visits. They're just rocket high and most likely in the few like two, three years from now. I don't see any reasons for telehealth visits just to go down. So a lot of people competing in this area, but still demand is higher than the market can offer. So that's a pretty remarkable change in the last ten, ten months. And talking about telehealth, there is something I came across last week that I think is pretty cool. So John McKay, the founder of Whole Foods, recently entered Telehealth. Yay, yay, yay. So McKay had already raised $31 million for new business venture described as an evidence based lifestyle company. So what is the deal with that? They acquired the plant based telehealth. So that's just the name. It's not actually plant based, but the name was plant based and they rebranded it to Love Life Telehealth. I don't see any significant traction from that yet. I just think it's wonderful. It's just something that caught my eye. Also, I'm adding it to the folder of folks with experience in retail stepping more and more into health care. And here I would like to jump to another biggest news of March, which was Atrium Health and Best Buy health partner to end Hayes Hospital at Home Experience.
Elena: [00:27:53] Yay. So what is pretty much about Atrium health is massive. I mean, you are thinking about atrium health as hospital at home. Think twice. They are much bigger than that. So when you look pretty much at functionality of atrium health, what immediately caught my eye is provider directories. They are adding to that big time and this is probably the main entrance into health care. So there are a few thoughts what Best Buy is doing. We all were trying to wrap our heads around big companies that are pretty much coming from their retail with extreme, amazing experience in a consumerization strategies and how to treat your client the best, right? Like bringing all those higher standards into health care that we were praying for the last 150 years in the United States. So what are they going to be doing? And I think it's important just to highlight what Deborah Decenzo said, who is the president of Best Buy Health. She said, we are never going to provide health care, Right? So before we were thinking that maybe they're going to be actually taking an active lead in primary care as primary care. But it seems like Best Buy through multiple acquisitions, just like current health in 2021, that Best Buy acquired. And we also see current health just growing, growing like a mushroom after the rain and now atrium.
Elena: [00:29:26] So this Frankenstein seem to be powering health care. Right. It's not a direct medical care that they are offering. They are trying to do the best they know, which is the supply chain and meeting the demand of service that's mostly needed on the market. So those are my $0.02 on that subject. And again, I just finished my Hospital at Home Report and Grace Winton was kind enough to invite me to participate in her podcast, hit Like a Girl, where I was sharing my thoughts on hospital at home and the future of it, and how difficult for some providers can be to enter this space and also how it can benefit the community and people that we live with or our friends that we've never seen for many years that still need medical help. Also, I think Hit Like a Girl is something that we really need to pay attention from from time to time just because they are inviting amazing speakers and they are talking about really hot topics on top of it. It was a special moment for me to be able to share some leadership thoughts and tips for women just in the industry, how to grow, how to feel more confident, and how to know that there is always a space for you if you look hard enough for it. And I'll pass it to Fatma.
Fatma: [00:30:51] I love that segment, Elena. I feel like that's really important, especially for women in the industry. So for me, this is my first podcast, which I'm so excited to record. So initially when I started with Chilmark, I was researching TEFCA heavily and it kind of came full circle because on February 13th, the agency announced that six organizations were approved as QHINs. And becoming a QHIN is a really rigorous and elaborate process, and these six players made the cut. We have Commonwealth Health Alliance, eHealth Exchange, Epic Health, Gorilla Health, Konza and Kno2. So by signing this common agreement, the prospective twins agree to fulfill the legal, technical as well as security criteria and have also proven their ability to the OMC to do so. And they agree to enable data exchange amongst one another. So we're witnessing this concept that the health IT space has been on for years just come to life and the future of interoperability lies within these six QHINs. So the aim of TEFCA for those that aren't familiar, is to provide a standardized, seamless as well as secure approach to health information exchange while promoting health exchange between different health care stakeholders, which ultimately can help improve the patient care, experience and health outcomes while reducing the health care costs.
Fatma: [00:32:13] And the stakes are really high for these QHINS as they have a lot to prove. And next, I'm going to talk about Zus health funding. So Zus Health is a health data startup that was founded by Jonathan Bush that aims to link health information so providers can access a comprehensive view of their patient's history. They recently acquired 40 million in financing, and we actually had a conversation with Jonathan last May, which will be linked in the show notes for those who are interested. But back to the topic at hand, the startup intends to utilize the latest funding to add new data sources, introduce integration pathways and build workflow and referral tools. So in addition to the funding, Zus has partnered with Elation Health, which is a primary care tech company, and it expects that Elation will integrate the Zus aggregated profile into its systems so clinicians can easily access patient records from any location with ease. So it's safe to say, I'm excited to see how this plays out and I'll pass it on to Jody.
Jody: [00:33:15] Thanks, Fatma. So I've been going down the rabbit hole of privacy and health data and secondary uses of data for is largely, largely provoked by the megapixel and health care systems. Controversy arose last year when the markup that looked at the Newsweek's top 100 hospitals in the country and found that 33 of them were using Meta's Pixel as a button when patients wanted to schedule appointments. And what happened was about 26 million patient appointments. You know, the data around patient appointments was fed into Facebook or Meta's data repositories or databases. And this provoked a quite a few lawsuits over, you know, the obvious privacy violations. But the interesting thing is the lawsuits are largely based on the Video Privacy Act, not HIPAA itself, which raises a lot of issues about, you know, HIPAA that was implemented in 1996 when we had a totally different technology landscape and technology business models where we now have these mega platform business models hoovering up as much data as possible. Many of them are coming into health care. We're talking about retail coming into health care and this broader ecosystem of companies coming into health care, Many of them, the standard business model is largely surveillance capitalism. It's collecting as much data on people as possible to personalize things and ads and so forth to people. And then we intersect with the health care system, which has very intimate private data. And in fact, in the last several months, Betterhelp got a $7.8 million fine from the FTC for their data or data privacy or data sharing practices.
Jody: [00:35:27] GoodRx. Got a $1.5 million fine. And then meanwhile, while this is going on, we have these data brokers that are making massive fortunes and it's largely a cartel of a small number of companies and putting more and more data behind paywalls, everything from Elsevier to, you know, a lot of these journals that all of the scientific research is published in data being put behind paywalls. But then in health care, we have NIH where if you accept an NIH grant, your data is supposed to be made open and available. How much of it actually is does anyone get held accountable for this? How does how is this structuring markets for the market for data and then increasing risks of privacy violations? Because we have privacy regulations that are based on a certain American history of privacy that well exceeds 1996, but now with a totally different technology infrastructure, it's pretty it's pretty worrisome when you look at what we have. And so I've been digging into this and a great deal. I've written a blog post about it and we'll do more on it because I think it's really important. And there are some interesting debates coming up in legal circles around privacy rights. And regulations are based too much on data, and we need to use a human rights discourse that protects humans rather than just data.
Jody: [00:37:05] Because when you look at, for example, some of the femtech apps where you could go out and for $160, buy all the data for a week's use of users have that used period tracking apps and find out who went to Planned Parenthood in the wake of the rolling back of Roe v Wade. This is really, really problematic, obviously, and I don't think it gets definitely gets attention, but I don't think it gets nearly enough attention. And we need to look at more at how it could be framed in a way to really protect people and and their data and so that people really and also people understand what they're signing off on when they enter a doctor's office and consent to have their data shared to data brokers and probably don't know it. And those forms are designed intentionally so that they don't read them and don't know what's going on. That's the system that's in place. And so where do patients and citizens. How do we benefit from this whole oligopoly of data brokers and this massive exchange of data? And will these lawsuits around the meta pixel thing make any difference in that? I don't think they will. And I think, you know, it looks like some of the companies will get off with this for, you know, you know, who's ultimately accountable for these issues and and who's going to stand up for the citizen and Congress around this will be an interesting issue to watch going forward.
Jody: [00:38:47] The other issue I'm looking at found rather interesting was when Pear Therapeutics, one of the most important players in the digital therapeutics space, announced they were looking for a merger and acquisition opportunities or sale of their IP and so forth. And it really highlights what a number of folks have been saying all along, that the the business model for for digital therapeutics was never really going to work. And it's kind of interesting when you look at what digital therapeutics are doing. They're trying to take digital health tools, many of them in the behavioral health space and treat them like a drug at the time when pharmaceutical companies are trying to go beyond the pill and have a business model that's less like a drug and more like digital health. Digital therapeutics tried to be more like a drug requiring a prescription going through doctors. And then the sort of barriers to entry into the market required the FDA, where the FDA is behind the times and regulating algorithms and to some degree, broader digital health tools. You know, many you have when you have the gatekeeper as the physician, then how many physicians actually know about these tools and are actually going to prescribe them how well they have to integrate them with the EHR and other tools they're using and the EHR.
Jody: [00:40:20] So it's kind of set up for failure the way it's currently, the way they're bringing these tools to market. So in many ways it's no surprise that pear is in the situation that it's in, and I think they need to really think, rethink the business model and how you introduce these tools, because I don't see this current. Framework going to offer many roads to success. I think they sort of backed themselves into a corner on that front. And then they also will have data privacy issues. And trust is going to be a big issue, especially when you're dealing with a large percentage of them are dealing with opioid addiction, substance abuse and so forth. You're getting very intimate data about people. And across the board we've seen pretty lax cybersecurity with a lot of digital health tools over the years. And now this. Also this week, the FDA announced that they're not going to approve tools that don't reach a sufficiently high enough cybersecurity standards. So that's a that's a good thing that happened this week. So that's those are the areas that I've been looking at. Other than beyond the Chatgpt issue, I don't know if the John we want to talk about that as a team because I know we all have thoughts about that. And given our focus on AI and trust, it's a timely issue for us.
John: [00:41:44] Yeah, I'd be happy to discuss that as a group. I want to lead off by just bringing to light the fact that yesterday or between yesterday and today and this is March 31st that we're recording this on Italy sent openai a cease and desist essentially to stop tracking their citizens and to stop operating in their borders. So we're starting to see some governments actually take action. And, you know, there was this absurd letter that a bunch of tech influencers wrote earlier in the week petitioning for AI developers to hold off on developing anything more powerful than GPT four. And I get where they're coming from. I am totally on board with their concerns around deploying things that are not ready for mass consumption, but the expectation that companies and private interests are going to stop developing these tools is absolutely ludicrous. I just maybe it was meant to catalyze more conversation around the topic, but I think that the expectation that will actually have an impact on continued development is mind numbingly ludicrous.
Jody: [00:42:48] And in some of the critics within the AI machine learning world have pointed out that the the letter itself buys into a lot of the hype and is framed in a way that buys into that. And also the organization that sponsored the creation of the letter and publication of it is one of Elon. It's in Elon Musk's orbit. So what's the what's in it for Elon Musk publishing this letter other than to be back in the center of attention around AI? And meanwhile, you know, there are critics of his autonomous vehicle ambitions as well.
John: [00:43:30] There's this there's been this trend of this or there's this like cohort of billionaire elites that have been trying to sound the alarm about AI. Sam Bankman-fried was one of them. Elon is another one. So I think Elon is legitimately concerned about some of these general AI and generative AI tools. But to your point, I think that his conflict of interest and his desire to always be in the limelight is a huge factor to consider when you're actually reading the letter and determining whether or not his input matters.
Jody: [00:44:02] Well, I'm glad you mentioned that in Sam Bankman-fried, because the a lot of the folks that signed it are part of the long termist, whole long termism, philanthropy, which has come under fire, you know, well before Sam Bankman-fried went down, but increasingly so. So I think that's where a lot of the criticism of the the letter comes from, where it's coming from and what's the other motive behind it. But I think, you know, going going beyond that letter, you know, to what we think about how these things need to be regulated, one thing people should know is we set up a LinkedIn group on generative AI and ethical frameworks in health care. So if you want to join a longer term conversation where, you know, look us up on on LinkedIn or look on my profile, I can get an invitation to folks that are interested in engaging in debates on how we're going to create stronger regulatory, ethical, political frameworks for generative AI. But there last year I wrote this report on AI and trust in health care and looking at a lot of the frameworks that have been developed for responsible AI. And if you look at just at the very basic, these are like 30,000 foot guidelines that have been out there in the field that go some way in regulating AI, but never as far as we would like.
Jody: [00:45:34] But around transparency and explainability and data governance. If you look at that and these like GPT four, we don't know all the data that this was trained on. So how do we know what biases are in that data? And there will be biases in the data. All data have a have a bias. How what's the logic of the model? We don't know and they're not releasing it. So this I mean, this makes it of limited utility in health care without substantial human involvement to sort of correct anything that because it can reproduce disinformation and scale it up rather readily as well as biases. And so I think to me, what I'd be I'm more interested in is what are the open source models that are much, much smaller, not nearly as expensive to develop, but focused on medical data that's been governed and de-identified and so forth in the in the right way to create health care specific tools that could be useful.
John: [00:46:42] So, Jody, to that point, I know that you're talking about equity a second ago. I don't want to lose the thread right now because I'm thinking about it, but there was an article that was just published in Nature about how fairness and AI research is being held back by lack of diversity. And to your point around making sure that these tools that we're developing are actually sourcing data from, you know, representative samples and collecting data in an equitable manner, we don't have any clarity to that. There's no transparency around how these models are really being trained, especially to ensure in health care we need to ensure that we're accounting for diversity and accounting for potential biases that are in the data sets that the I wouldn't know about that we as humans need to filter out ahead of time before actually utilizing them in these training sets.
Jody: [00:47:30] They mean that you have things beyond health care. The algorithmic Justice League timnit Gebru, whose group focusing on a lot of the very problematic issues in in AI and racial bias, gender bias, on and on and on. But we really need something very health care specific with people very familiar with health care, health data and, you know, HIPAA and beyond and hopefully do better than HIPAA. As we develop these tools and, you know, increasingly, one of the things I want to write write about going forward is we talk about social determinants of health. There's this last week there was an article that came out on the commercial determinants of health, largely around food and other businesses that outside of health care, the formal health care system that have an impact on people's ability to, you know, be healthy. But I think probably one of the most important things going forward we need to think about are the algorithmic determinants of health and sort of algorithmic slash technological determinants of health. I mean, at Vive I saw they were talking about equity and what I saw saw what they're doing on that. You know, it's an important thing to do. I've nothing against it, but just a very, very strong focus on the algorithms out there.
Jody: [00:49:04] I remember back in you may have seen in July 2021, the New England Journal of Medicine. There was an article by a whole list of authors. First one is, as I recall, if you want to look it up, they went through just the number of clinical decision support tools that have racial bias of one form or another in from algorithms used that were created by the American Heart Association to the American College of Thoracic Surgeons. You know, just in January, we finally the whole glomerular function for kidney disease. Finally, after years of people noting that the 1999 study that treated race as a as a biological category, where it should have been treated as a social construct that was used to justify race adjusting, glomerular function, which led to blacks being lower on the kidney transplant list and getting medications later, so forth. So these are the issues that, you know, have been out there before the chat, GPT, threes and fours came along. These are only going to grow and they can grow exponentially. And so we really I'm very deeply concerned about, you know, we talk about bias all the time, but what are the institutions that we're building to actually make people accountable and companies and the government, everybody accountable?
John: [00:50:37] Yeah. You actually had a really great podcast about this last year where you interviewed Dr. Tanya martin, Mercato about specifically these hidden biases and how it's so hard to extract some of that out because even if you remove race as a parameter, it can still get at race through other means, like your location of residency, zip codes, things like that that are proxies for race.
Jody: [00:50:59] And and, you know, the work is Ziad Obermeyer at Berkeley. When he found the the algorithm or the model that Optum was using that created neglect of African American patients for, I think it was around heart disease, disease management or something. But the broader point he saw is that, you know, these things often are not they're not in what Tanya mercado martinez is very good at articulating. Was this a lot of the racial bias? It's not intentional by the data scientists, but you see these these examples where like with with the Optum algorithm, they were framing the problem to be solved as controlling costs rather than, you know, I guess, access to care. And they use the proxy data they used was insurance coverage where African Americans have lower levels of insurance insurance coverage. So that got integrated into the model and then, you know, created a biased model which ended up referring people less to the care they should have gotten. And so there's just a lot of this out there that I'm just deeply worried if we just jump into these large language models or generative AI and we don't know anything about the data and it's just like massive amounts of data with potentially massive amounts of bias and other problems, What are we going to create? It may not be good. And platform business models coming into health care where that's a very you know, AI is a very big component of it of the business model. And with these outside players coming into health care, with these practices of using these things and sort of the data surveillance business model as part of the platform business model, we've got a lot to worry about.
John: [00:52:57] We absolutely do. I think that's one of the big things that you've been pushing on since you started covering for us is the need for these solutions to be transparent and for them to be auditable. And I think that a lot of people in the health care community that are data scientists and are paying attention to the AI space, they are all in agreement that that is an absolute must have if you are going to be working in the health care space, because you need to be able to make sure and dig into the data to to be able to test it with your populations, with synthetic data, whatever it is. But you need to be able to test and identify where potential confounding factors are coming in and we're biased might be getting introduced in the model so you can go in and fix it if you actually want to keep using that tool. All right. Well, thanks for that very detailed breakdown of some of these considerations, Jody. And what's going on in the space is everybody else is trying to cover this. I think that was a really good breakdown of what the key considerations are for our sector. Well, thank you for joining us today. I hope that you found this podcast informative and if you have any questions or would like to follow up, please feel free to reach out to us. We will include any of the links and references cited during this podcast in our show notes and that will be up on our blog in a few days. So thank you for subscribing and listening in.