Can AI Save Healthcare? HLTH Conference Explores the Possibilities
A Physician’s Take on HLTH 2023
October has been and still is the “busy” conference season. Some might consider this the second year of a return to the conference activity post-pandemic but even last year things were still a little muted and not quite there.
This year conferences have come roaring back and HLTH was no exception with a big uptick and attendance and plenty to look forward to. The conference appears to have settled on Las Vegas which no matter if you love or hate the place is a good choice for a conference for its ease of getting to and from by air and is well equipped for volume…mostly.
This conference is definitely flashy and glitzy with lots to entertain as well as stimulate discussions. I confess to being torn on the issue – I recognize the value of creating a dynamic and high-energy environment and have certainly benefited from some unique experiences but this all comes at a cost. Given the dire statistics and data on the state of the US healthcare system and its many failings to deliver good, let alone great care to all comers mentally I struggle with this. That said creating these reexperiences creates opportunities for mixing and matching people and ideas in unique and wonderfully innovative ways. In some respects, this approach answers one of my initial pushbacks on the conference which felt to me to be getting too big. I felt very much like Sheldon from the “Big Bang Theory” and his selection of his spot on the couch. It suited him because it was just right and that’s probably similar for many. But economics and demand push for an increase in size and attendance so bigger while not a certainty is often the goal and direction for conference organizers.
That’s a hard problem to solve and not one that everyone agrees on. Bigger can mean better with more opportunities to interact and engage with others and the potential for chance meetings. But as the numbers increase so does the challenge of finding your place in the plethora of opportunities. With all the people there I know there were old friends I’d like to have seen and certainly others I would have enjoyed getting to know but finding them was not always easy.
But the ingredients are there and contribute to creating those sparks. HLTH has made smart choices from the start – bringing a decent range of food to the floor to keep people there. Same principle with sessions and they do a bang-up job on the sound to keep the background noise both while listening as well as when not listening to a minimum. So no emptying of the halls as people head out for food and educational sessions. The floor was always busy even in the furthest reaches thanks to the placement of the various stages in distant locations around the extensive floor layout.
It is a mix of entertainment and business designed to bring people together and continued to do a good job as an astute friend noted to me when I asked what he was doing in our random meeting as I roamed the halls, he said:
“I’m scavenging the Halls for connections.”
I shared this with a number of others I met as I walked the halls and got universal nods of agreement.
I know I missed plenty but came away with some helpful insights, some new connections that I look forward to continued discussions, and a renewed sense of optimism while many believe we are in “The Year of AI” (as stated by Aaron Mauck from the Advisory Board in his presentation – available here) not in my opinion. As my brother pointed out to me some years back the technology is not new (accepting there are new developments within the concepts of it that are very significant). But Thanks to the emergence of ChatGPT and some incredible advances in one of the domains of the technology the general awareness of the technology and its rapidly advancing capabilities is bringing significant positive perception and interest to its application in medicine. There are some pretty clear use cases and examples of successfully building solutions to reduce the administrative burden and overhead associated with medicine.
From Therapy Dogs to AI
Wearing my clinician hat I came away with a few insights
- Healthcare remains a massive industry. There is no escaping the engine this is in our economy from jobs and general productivity and that means continued focus even in the light of some significant pressures and the massive complexity of the interconnected systems, processes, regulations, people, and legislation.
- Despite the complexity companies are still jumping in with both feet and General Catalyst’s announcement that they were going to buy a health system. Plenty of armchair commentators across a range of perspectives including myself. Certainly interesting to watch but doing that won’t be quite as easy as some might think. And the competing interests in that ownership model overlaid on what is already a broken incentive system that is draining the life force of the clinical staff on the front lines already is questionable imho
No matter the headwinds keep expecting non-healthcare companies to jump in insistent that they can solve this/the problem. - The jobs (in healthcare) are a changing. It’s not just in healthcare and none of this is new but the wide array of adjunct technologies and capabilities is bringing impact on the way we deliver healthcare. You must adapt to survive – not replacement but adjunct assistive in nature although some things will be replaced by technology. All this comes with the continued decline in resources and the aging of many of those still there. 60% of cardiologists (and many other specialties) are over 55. Given the extended training requirements backfilling those voids is going to be challenging
- We lost about 200,000 nurses during the pandemic – there is no switch to turn on and replace quickly and with all the widespread coverage attracting new entrants remains challenging. Keeping those who have already invested huge amounts to ascend the ladder is also a challenge.
- Doing more with less will be the mantra for the next few years at least and maybe longer. See above on jobs changing. More efficiency cannot come at the expense of compassion, caring, and most importantly the personal resilience of the people actually delivering the care. 70% of junior doctors’ time is spent on administrative tasks that have to change quickly so they can return to the patient and focus on caring about people, not systems and processes.
- Expect a continued nuclear arms race between the opposing sides of healthcare systems. Those who deliver and bill for care and those who pay for it. The insurance system is fundamentally broken, the cost shift has devasted many and the incentives in those cases are focused on maximizing profitability. They are using AI intelligence to create new and unique ways to optimize their financials. On the other side of that equation are the clinicians dealing with the patient in front of them but forced to jump through increasing administrative demands to deliver the care who are also turning to AI to automate the administrative elements and fulfill the requirements
- Employers are rising up and there are increasing options and focus on helping them make better choices and provide better more economical care. For a long time they assumed that their agents and parties involved in delivering were always acting on their behalf, but as I learned later than I should in life, look to see how is paying the bill of your advisor before you trust the guidance
- Drugs, drugs, drugs. Biggest expanding costs with many areas that will on their own break the bank. As one commentator noted – for a small company employing a hemophiliac could mean they go out of business just based on the associated medical costs. Let that sink in for a second. Put aside that we have been successfully treating this disease for many, many years. I should know as I trained at one of the early centers of expertise. But its not just that condition but many more and even if the employer is being responsible and compassionate the impact of that even unintentionally could contribute to decisions on employment. I heard nothing at the conference but can’t shake the notion that employers should be out of the business of healthcare. I have said this repeatedly and while some agree I do hear others push back and say it is a way to differentiate on benefits. But the harm and national cost of this is hurting our economy, our business, and above all our patients. But there seems to be no appetite to take this on.
- Consumerism was alive and well and expanding – sooo much dissatisfaction in the healthcare experience is evident so plenty of cracks to open up to find point solutions but as we have seen with Portalitis there are only so much of individual solutions and options that the consumer can tolerate or is willing to accept let alone pay for. This will shake out but there are certainly interesting examples and opportunities.
- The emergence of large data and a range of possible markers of disease buried in a complex array of interconnected data points creates an opportunity for new and interesting testing. But much like my pushback on the notion of Whole Body MRI being advocated by “influencers” the equation and decisions on this come with significant risk-reward assessments. The Galleri Cancer Screening test was on “offer” but that option was only available if you represent a system that is interested in buying a big order of them. Let be clear in the fractured population already bleeding out from healthcare costs – another test priced at $949 has an uphill struggle for the majority imo. For the worried well it might be the hot ticket. But I have yet to look at the studies to see where this falls in value
Willow was Such a Good Girl
My annual recurring highlight was the Therapy Dog booth where you can spend time with a number of dogs. I know not everyone is lucky enough to have or sometimes even like a pet. My own father was terrified of dogs especially for much of his life so I understand – but as I managed to demonstrate to him, a well-trained, calm dog can overcome even a PTSD World War 2 linked phobia as it did in his case (n=1). YMMV
