🏥 There is no labor shortage
 |
|||||||||
|
|||||||||
|
|
|||||||||
Hospitalogists, Me, watching the US Open carnage today: https://tenor.com/view/cheer-intense-yelling-cheering-gif-4968107 Earlier this year I had the distinct pleasure of attending ShiftMed’s Transformation Summit in Tampa, at Raymond James stadium. It was a great time with some amazing discussion around workforce transformation. I kicked the event off on stage talking pie-in-the-sky AI macro, then sat back and let a parade of people who actually run staffing offices, float pools, and innovation shops tell me how the sausage gets made. I came in to evangelize about robot overlords and personal health records and left Raymond James stadium thinking about operational challenges in labor supply & demand and flexible staffing models. Here were my main takeaways including consensus views, interesting claims, and other factoids from the event. I’d recommend putting it up for consideration next year for your organization given the depth of experience in the room (and the fact that I was there, of course). PS - shout out to my guy Jacob Laufer who is an awesome dude and literally booked Raymond James stadium because he’s a massive Bucs fan. Next Hospitalogy event in 2027 will be in the biggest suite at DKR in Austin. 
The infamous pirate ship. Is it Oscar's? IYKYK Was this email forwarded to you? Sponsored by R1 Mid-cycle revenue operations have long been a highly manual and fragmented dysfunction in healthcare finance. Clinical documentation moves from coders to CDI teams to auditors — each reviewing the same chart at a different stage, creating a pileup of duplicate effort, handoff delays, and administrative drag that most health systems have quietly accepted as table stakes. AI is starting to change the norm. R1 breaks down how intelligent automation is beginning to connect coding, CDI, and auditing workflows, reducing repetitive manual review and freeing staff to focus on exceptions, quality, and decisions that actually require human judgment. Worth the read. And if you want to go deeper, their new report on the revenue cycle workforce in the age of AI is available to download. BLAKE'S BREAKDOWN The Labor Shortage That Isn't (and 5 Other Things I Took Away From a Room Full of Workforce Wizards)TL;DRConsensus themes from the room:
Some of the more provocative takes from the event:
We're in peak hospital mode. 2026 is the calm before the 2027 stormAfter a high utilization environment and a rebound post-pandemic, 2027 is now looming for hospitals and health systems. Over the last decade, the government payer mix has crept from ~40% to ~45%, and operating margins have slid from ~3.5% to ~2.5% in lockstep. Over 42% of hospitals are losing money. We are over-bedded: inpatient admissions per capita keep drifting down. And the demographic freight train hasn't even peaked — 10,000 boomers turn 65 every day, Medicare service lines run at roughly a -13% margin, and systems are staring down something like a full point of margin erosion over the next five years just from mix shift they can't control. Layer OBBB on top — call it ~$1T out of the system, plus the uninsured-care tab climbing toward $85B — and you get a sector bracing for impact in '27. Uninsured rates rise, Medicaid work requirements bite, ACA subsidies lapse, automatic re-enrollment goes away. The efficiency and margin gains everyone's chasing through AI and workforce redesign are real, but in my experience as a former valuation guy, margin accretion in healthcare is almost always a short-term loan from the government. CMS has clawed back reimbursement in MA, in imaging, in home health and hospice via PDGM, in VBC through rebasement and moving benchmarks. DRGs get reweighted to bake in your "efficiency" as the new baseline. So when a vendor tells you AI is going to permanently expand your margin, my honest (but unfortunate) read is that some value accrues to systems, but more of it gets recaptured by payers, by government, and by the tech and vendor ecosystem selling you the tools. Take advantage of the windfall when you have it, but don’t expect it to last. The labor shortage that isn't
There is no nursing shortage. There's a coordination problem and a packaging problem dressed up as one. The supply is there, and increasing. We've gone from roughly 1 in 5 to nearly 1 in 4 working Americans employed in a health system or hospital. Hospital unemployment sits near 1.6%. Many states are projected to have an RN surplus. What's actually broken is the fit. Searches for fractional and part-time work run ~77% of clinician demand; full-time postings run ~80% of supply. We are advertising the exact opposite of what the workforce is looking for — and if you understand anything about how LLM and search results surface jobs, those full-time reqs aren't even showing up in the feeds where clinicians are actually looking. Nurses over 55 have dropped from ~43% of the workforce to ~16% in five years — we're hemorrhaging our most seasoned clinicians and turning them into patients at the same time. Behind them sits Gen Z, soon to be 60%+ of the workforce, with completely different expectations. Stop asking professionals to conform to the institution and start redesigning the institution around the professional. The ones who don't will keep paying premium-pay and travel rates to plug holes that fractional and float-pool design would close for ~6% less. And the size of that prize is not theoretical. One system banked ~$130M in labor savings across 1M+ hours by leaning into flexible models. The operational version of this — going from "22 spreadsheets and a two-hour phone tree" to filling a shift in 58 seconds — isn't an AI moonshot. It's plumbing. It's the least glamorous, highest-ROI work in the building. 
What the workforce cares about MORE than moneyWhen you survey nurses on what they actually want, pay lands around #4 or #5. What sits at #1 is control over their own schedule. We keep throwing money at a problem that isn't fundamentally about money. The most resonant story of the day — and I'll spare the speaker the spotlight — was a float nurse who picked up a specific shift, only to have someone literally take an eraser to the paper schedule and reassign her to the worst unit in the building. She never picked up a voluntary shift again. Clinicians won't commit if they don't trust that what they committed to will still be true tomorrow. The bot wars and the self-driving car problemOn the clinical and administrative side, we're watching an arms race I've started calling the AI bot wars. Providers got to the technology first — it's far easier to bolt a scribe onto your CDI function and document more completely (providers say "more accurate acuity"; payers say "upcoding") than it is for a payer to respond. You can see it in HCA's net revenue per adjusted admission compounding at ~4.2%. Payers are now firing back with billions in AI investment, more denials, and research showing the same diagnosis pays wildly differently across hospitals. My fear isn't that one side wins. It's that we automate and entrench the worst parts of the system — prior auth becoming bot-versus-bot, with the human friction now running at machine speed. Which brings me to the frame I'm most attached to: healthcare AI has a self-driving car problem. A Waymo gets stuck at an intersection and it's national news; the ~40,000 annual road deaths — a 737 falling out of the sky every day — are background noise. We judge the machine against perfection and the status quo against nothing. Healthcare does the exact same thing. Nearly every clinical AI study measures the model against an idealized ground truth, not against relative human performance, which is the actual bar. Over-reliance is a real failure mode and I'm not waving it away. But until we start asking "better than the tired resident at 3am?" instead of "perfect?", we'll keep slow-walking tools that are already safer than what they replace — and the convergence point, where it's simply cheaper and lower-risk to deploy AI than not, is coming whether the liability and reimbursement questions are settled or not. The cop-out heard 'round the summitI asked a version of this question to a lot of leaders and never got a straight answer: if you could automate your entire revenue cycle function end-to-end and fire 2,000 people tomorrow, would you? What does that mean politically for your organization? The responses all rhymed — "we have plenty of roles, we'll reskill them." And look, I understand why a CEO says that on a stage. But it's a cop-out, and the fact that nobody has a real answer is itself the story. The honest version is that the role of the clinician is going to change in ways the comp model isn't ready for — panels expanding from 3,500 toward 20,000 patients, the physician becoming an orchestrator of agentic work rather than the worker, the hospital subsidy-employment model wobbling the moment referral patterns slip out of the walled garden. I also posed this take to the room - critical thinking is becoming a premium good. I use AI to write and outline and learn, and I can feel the temptation to outsource the thinking itself. The sharpest take in the building came from someone half-joking that they want the pen-and-paper veteran running the AI tool — not the person who only ever knew the tool. AI outputs are sterile and standardized by design; a narrow band of competent, average answers. Great for reducing unwarranted variation. Quietly terrible for the eureka moments and the clinicians who are supposed to catch the model when it's confidently wrong. Consumerism is already walking through the front door plus other predictionsThis is a regularly recurring talking point of mine. Roughly 30% of people are using AI for health questions; OpenAI alone is fielding ~40M health queries a day. That's an existential rerouting of the front door. The patient without a PCP — which is a growing segment — isn't calling your access line. They're asking a chatbot, and increasingly that chatbot will be wired to a payer's network or a Zocdoc-style action layer that steers them somewhere that isn't you. Stack the rest on top: primary care stratifying into a two-tier system (cash-pay longevity clinics for the affluent, APP-heavy MA models for everyone else, and a genuinely interesting AI-enabled opening on the Medicaid/rural side). GLP-1s torching once-lucrative elective lines — more drugs, fewer bariatric surgeries, and that's before the next wave of AI-accelerated drug discovery potentially compresses development timelines from a decade to a few years. The personal health record, healthcare's white whale, getting another flood of venture and frontier-lab money. None of these is fully here. All of them rewire demand. Systems that aren't building their own consumer-facing front door are ceding it. Sponsored by Innovaccer Context: The Healthcare Autonomy Conference kicks off June 23–24. This fully virtual event features live Agentic AI Workshops where real operational challenges get worked through in real time, unscripted, across 5 tracks: • AI-Enabled Physician Groups If you’re a CMO, CFO, CIO, ACO exec, or payor leader navigating the shift from AI that assists to AI that executes, this is the event you’ve been waiting for. What still can't be automatedThe best innovation story of the day wasn't about AI at all. It was a government agency — about as status-quo as it gets — solving a transportation problem not by building anything, but by renting Uber's logistics and integrating it. You don't have to invent everything in-house. you have to be willing to bring the right outsiders to the table and align everyone around a shared win. The same instinct shows up in the workforce world, where the systems making real progress are stitching scheduling, credentialing, and even transportation into one ecosystem rather than buying 12 point solutions and praying they can work together...and that you can manage those contracts. The biggest through-line that every convo circled back to is that the more we advance technologically, the more the human part of this business becomes the differentiator. Empathy, trust, schedule control, the bedside time we keep stealing from managers and handing to spreadsheets. The organizations that win the next few years won't be the ones with the most AI. They'll be the ones flexible enough to absorb it while keeping the humans — clinicians and patients alike — believing the loop works. That's the whole thing, really. We spent a day talking about robots, routing algorithms, and air taxis. But healthcare's hardest problems are still coordination, trust, and packaging — and those have always been human jobs. Hospitalogy Top Reads & Resources
*This post is brought to you by one of our amazing ad partners. MISCELLANEOUS MADDENINGS Right now the U.S. Open has kicked off and it is absolutely fascinating to watch the best golfers in the world take on a grueling challenge in brutal conditions - hitting balls into knee-high rough, balls not holding greens, massive wind gusts…and yet there’s a handful of them who shoot under par rounds from that distance. I can’t even begin to fathom how good of golf a round like that requires. But I’ve always wondered…just how BAD I would play in those conditions, and what my score would be. Call me a masochist but I kind of sounds fun. Have a great weekend, Hospitalogy fam. Thanks for the read! Let me know what you thought by replying back to this email. — Blake |
|||||||||
|
No comments