AI in Healthcare Providers and Services - People are Engaging w/ the Substack!
The following is from Arnav Murudkar, a sophomore in Kelley. Arnav and his fellow KC members, Joel Zechariah and Arjun Iyer, were kind enough to allow us to post their industry report (linked below)
From Arnav: I explored the AI in HC topic briefly for my Knall-Cohen (KC) report. We covered HC Providers & Services - I've attached the pdf (at the bottom) if you want to check it out. In the interest of page count and the shorter 1–2-year outlook requested for the assignment, we sparingly mentioned AI in the report given it will likely take several years for this technology to lead to material impacts on the sector.
That being said, for my sub-sector — Managed Care (Health Insurers) — the primary use case for AI I encountered is better processing and managing of customer data leading to cost efficiencies. Insurance stocks trade largely on their Medical Loss Ratios (MLRs), which represent how much of the company's revenues (insurance premiums) are spent covering medical costs; effectively, this can be thought of as a margin measure. UNH fell 8% Tuesday due to a higher-than-expected MLR.
If these companies can find some efficiency through AI to better sift through data — which, in my view, is possible — think about how many different "service providers" provide care to someone with serious conditions.
Start with your primary care provider → go to specialist(s) and potentially get scans, tests, etc. → get inpatient (hospital) care → recovery care/physical therapy.
An insurer, especially one offering a high-quality plan with broad coverage will intake patient data from all of these service providers for each patient. This creates ever-scaling inefficiencies with data processing/analysis from multiple sources, particularly considering the amount of legacy software in the industry. I see the potential for AI to unlock solutions to these inefficiencies.
Some questions that I contemplated along the way…
Can AI be used to standardize data from different providers?
Can AI provide faster, more informed care decisions in patient-present scenarios?
Can AI improve efficiencies to compare a patient's current conditions with patients with comparable conditions in the past to determine the best/most cost-efficient care?
Even a sustained decrease in MLRs by 10bps would be significant for these companies. I like to think of MLRs as the opposite of R* for the Fed. The Fed controls short-term rates and does not (directly) control R*. Insurers have less control over their short-term MLRs, which are affected by year-old negotiations with providers and can fluctuate with Medicare/Medicaid enrollment impacts, usage trends, etc.
However, insurers can to a larger extent control their long-term MLRs by negotiating better future contracts, better cost management (through AI?), geographic coverage, Medicare/Medicaid vs. private insurance exposure, etc.
I hope this is somewhat relevant and gives some insight into AI usage in a sliver of the HC industry.
From Nikhil & Xander: It’s pretty awesome for us to see people engaging with the blog like this. Thanks to everyone who’s been reading our stuff, always open to any feedback you have.
From Arnav, Arjun, and Joel
