Welcome to Thoughts on the Market.
I'm Terrence Flynn, Morgan Stanley's U.S. biopharma analyst.
And I'm Erin Wright, U.S. healthcare services analyst.
Thanks for joining us.
We're actually in the midst of the second day of Morgan Stanley's annual global healthcare conference, where we hosted over 400 companies.
And there are a number of important themes that we discussed, including healthcare policy and capital allocation.
Now today on the show, we're going to discuss one of these themes healthcare spending, which is one of the most pressing challenges facing the US economy today.
It's Tuesday, September 9th at 8 a.m. in New York.
Imagine getting a bill for a routine doctor's visit and seeing a number that makes you do a double take.
Maybe it's $300 for a quick checkup or thousands of dollars for a simple procedure.
For many Americans, those moments of sticker shock aren't rare, they're the reality.
Now, with healthcare costs in the US higher than many other peer countries on a percentage of GDP basis, it's no wonder that everyone, not just investors, is asking not just why is this happening, but how can we fix it.
And that's why we're talking about AI today.
Could it be the breakthrough needed to help rein in those costs and reshape how care is delivered?
Now I'm going to go over to you, Aaron.
Why is U.S. healthcare spending growing so rapidly compared to peer countries?
Clearly the aging population in the US and rising chronic disease burden here are clearly driving up demand for healthcare, seeing escalating demand across the senior population, for instance.
It's coinciding with greater utilization of more sophisticated therapeutics and services.
Overall, it's straining the health care system.
We are seeing burnout and labor constraints at hospitals and broader health systems overall.
Net-net.
The US spent 18 of GDP on health care in 2023 and that's compared to only 11 for peer countries.
And it's projected to reach 25% to 30% of GDP by 2050.
So the costs are clearly escalating here.
Thanks, Erin.
That's a great way to frame the problem.
Now, as we think about AI, where does that come in to help potentially bend the cost curve?
We think AI can drive meaningful efficiencies across healthcare delivery, with estimated savings of about 300 to 900 billion by 2050.
So the focus areas include here staffing, supply chain scheduling, adherence.
These are where AI tools can really address some of these inefficiencies in care and ultimately drive health outcomes.
There are implementation costs and risks for hospitals, but we do think the savings here can be substantial.
Great.
Well, let's unpack that a little bit more now.
So if you think about the biggest cost buckets in hospitals, where can AI help out?
The biggest cost bucket for a hospital today clearly is labor.
It represents about half of spend for a hospital.
AI can optimize staffing, reduce burnout with a new scribe and some of these scribe technologies that are out there and more efficient healthcare record keeping.
I mean, this can really help to drive meaningful cost savings.
Just to add another discouraging data point for you.
There's estimated to be a shortage of about 100000 critical health care workers in 2028.
So AI can help to address that.
AI tools can be used across administrative functions as well.
That accounts for about 15% to 20% of spend for a hospital.
So we see substantial savings as well across drugs supplies, lab testing, where AI can reduce waste and improve adherence overall.
Great.
Maybe we'll pivot over to the managed care and value-based care side now.
How is AI being used in these verticals, Aaron?
For a healthcare insurer, and they're facing many challenges right now as well.
AI can help personalize care plans and they can support better predictive analytics and ultimately help to optimize utilization trends.
And it can also help to facilitate value-based care arrangements, which can ultimately drive better health outcomes and bend the cost curve.
And ultimately, that's the key theme that we're trying to focus on here.
So I'll turn it over to you, Terrence, now.
While hospitals and payers could see notable benefits from AI, the biopharma side of the equation is just as critical here, especially when it comes to long-term cost containment.
You've been closely tracking how AI is transforming drug development.
What exactly are you seeing?
Yeah, a number of key constituents are leaning in here on AI in a number of different ways.
I'd say the most meaningful way that could help bend the cost curve is on R&D productivity.
As many people probably know, it can take a very long time for a drug to reach the market, anywhere from 8 to 10 years.
And if AI can be used to improve that cycle time or boost the probability of success, the probability of a drug reaching the market, that could have a meaningful benefit on costs.
And so we think AI has the potential to increase drug approvals by 10% to 40%.
And if that happens, you can ultimately drive cost savings of anywhere from 100 billion to 600 billion by 2050.
Yeah, that sounds meaningful.
How do you think additional drug approvals lead to meaningful cost savings in the health care system?
Look, I mean high level medicines at their best cure disease or prevent people from being admitted to a hospital or seeking care at a doctor's office.
Equally important, medicines can get people out of the hospital quicker and back to contributing or participating in society.
And there's data out there in the literature showing that new drugs can reduce hospital stays by anywhere from 11 to 16.
And so if you think about keeping people out of hospitals or physician offices or reducing hospital stays, that really can result in meaningful savings.
And that would be the result of more or better drugs reaching the market over the next decades.
And how is the FDA now supporting or even helping to endorse AI-driven drug development?
If companies are applying for more drug approvals here as a result of AI discovery capabilities, without modernization the FDA could actually become the bottleneck and limit the number of drugs approved each year.
And so in June, the agency rolled out an AI tool called ELSA that's looking to improve the drug review timelines.
Now, ELSA has the potential to accelerate these timelines for new therapies.
It can take anywhere from six to 10 months for the FDA to actually approve a drug.
And so these AI tools could potentially help decrease those timelines.
And are you actually seeing some of these biopharma companies actually investing in AI talent?
Yes, definitely.
I mean, AI related job postings in our sector have doubled since 2021.
Companies are increasingly hiring across the board for a number of different parts of their workflow, including discovery, which we just talked about, but also clinical trials marketing, regulatory a whole host of different job descriptions.
So, whether it's optimizing hospital operations or accelerating drug discovery, AI is emerging as a powerful lever here to bend the healthcare cost curve.
Exactly.
The challenge is adoption, but the potential is transformative.
Erin, thanks so much for taking the time to talk with us.
Great speaking with you, Terrence.
And thanks everyone for listening.
If you enjoy Thoughts on the Market, please leave us a review wherever you listen and share the podcast with a friend or colleague today.
The preceding content is informational only and based on information available when created.
It is not an offer or solicitation, nor is it tax or legal advice.
It does not consider your financial circumstances and objectives and may not be suitable for you.
