Washington Is About to Grade Your Technology on Labor

So here is the number that belongs on the first slide of every medtech strategy review this fall, and it did not come from a medtech company.

Over roughly two decades, labor productivity in US clinical care organizations fell about 1%. Across the rest of the American service economy it rose more than 55%. In the same stretch, healthcare organizations went to spending north of $150 billion a year on information technology.

Fifty-six points of divergence. Bought and paid for.

Everybody reads that as an IT failure story. I read it as a purchasing test, and I think it is about to become the purchasing test, because a productivity gap that large stops being an operating problem once it becomes an election issue.

Here is where I think this goes. I want the conclusions in front of you before the evidence, because the useful part of this issue is the doing.

One. The payment floor under hospital-based procedures is already moving toward the office and the ambulatory surgery center. That part is finished rulemaking, not forecast.

Two. Washington is shifting from a coverage frame to an affordability frame, and the two ask completely different questions about your device.

Three. The affordability question is whether your technology lowers the cost of producing one successful episode of care or adds another billable input to it. A $1 million technology that releases $3 million of labor, complications and capacity becomes politically protected. A $1 million technology that delivers a 4% clinical improvement and raises cost per episode becomes politically exposed. A $1 million technology where nobody can demonstrate either one is the most exposed thing in the building.

Four. The scarcest input in American healthcare is a trained clinician, so the strongest answer to that question is measured in labor hours returned per thousand patients.

Five. 2027 and 2028 produce more healthcare legislation than 2026 did, because divided government pushes both parties off system redesign and onto discrete deals against cost pools they already agree they dislike.

Six. Premium technology gets redefined, from the best clinical result to the best clinical result per unit of scarce labor. The commercial organization barbells with it.

At LiquidSMARTS℠ we have taught Clinical, Operational and Financial value for a decade. COF has held up because those are the three arguments a hospital actually runs on. What changed this year is that the field started answering a fourth question before any of us wrote it down, and the fourth question is the one Washington is walking toward.

Now the six things underneath that, in the order they will hit your number.

---

The Mechanic

1. The purchasing question already changed. The policy is the lagging indicator

Sit in a capital committee this quarter and count how the conversation actually runs.

A surgical robotics rep presents a 14% reduction in a complication rate. Good data, well controlled, peer reviewed. The CFO listens, nods, and asks a question that has nothing to do with the paper: how many OR turnovers does this add or remove, and does it let us run the same room with the team we can actually staff on a Thursday.

The rep does not have that number. Nobody built it for them. The deal goes to committee review and dies quietly in the second budget cycle.

Now run the same meeting with an imaging AI vendor who walks in and says their triage tool returns 40 minutes per radiologist per shift, that the department is running eleven radiologists against a fourteen-body plan, and that 40 minutes across eleven people is roughly a full additional reading day every week without a recruiter, a signing bonus or an agency rate. The clinical claim in that pitch is smaller than the robotics claim. The deal closes faster.

So the question that decides the deal has already moved from what the technology does to a patient toward what it does to a schedule. The buyer changed the test. Most sales organizations are still answering the old one.

Chapter and verse.

The Conference Board published the ratio in June. More than 700,000 job openings in healthcare every month, against 306,000 unemployed workers available to fill them. Their own recommendation list ends where you would expect it to end, with deploying technology in ways that expand effective capacity.

That ratio is the whole story, and the reason it matters commercially is that hiring has already been tried at scale. Over the past two decades US clinical care organizations grew their workforce 2% to 3% a year and added more than five million clinical roles. Patient wait times did not improve. So the constraint is not headcount. It is that nobody has redesigned how the work gets done, which is exactly McKinsey's conclusion and the sentence in their analysis I keep coming back to: the industry is automating inefficiency faster than it is eliminating it.

More technology, more people, less output per unit of labor. That is what $150 billion a year of healthcare IT bought.

Now the part that should change how you sell. McKinsey puts labor cost improvement above 20% for organizations that genuinely redesign a care model, and attaches a warning worth more to us than the number is. Those gains only appear when workflow, technology and staffing are redesigned together. Adding a point solution without redesigning the care process or the staffing model can reduce productivity, increase caregiver burden and reverse the benefit you sold.

Read that as a commercial instruction. A technology dropped into an unchanged workflow can make your customer measurably worse off, and they will attribute that to your product. Which puts the implementation conversation inside the value case, rather than in a service handoff after signature.

2. The payment floor is moving, and your business case is standing on it

Take a device whose economics assume a hospital outpatient facility fee. Most of the interventional and imaging portfolio in this industry was priced against that assumption at some point.

Now move the same procedure into a physician office or an ambulatory surgery center at office rates. Your customer's revenue per case falls. Your price did not. The value case you wrote in 2023 is now arguing for a bigger share of a smaller number, and you will find out about it as a stalled renewal rather than as a policy briefing.

The site of care is being repriced underneath the product, and the product's business case was built when the site was worth more.

Chapter and verse.

This is finished rulemaking, not a proposal, and CMS moved three levers inside eight weeks.

On October 31 the CY2026 Physician Fee Schedule final rule applied a 2.5% efficiency adjustment to work RVUs and the intraservice physician time behind them, for non-time-based services. Procedures, radiology and diagnostic tests. Time-based codes are exempt, so evaluation and management, care management and behavioral health walk away clean and the procedural side absorbs it. CMS derived the 2.5% from five years of cumulative MEI productivity adjustment, noted that more recent BLS data would have justified 3.6%, and said further efficiency adjustments should be expected roughly every three years.

Read that last part twice. This is a standing haircut on procedural work, not a one-time event.

The same rule cut the facility practice expense RVUs allocated on work RVUs to half the non-facility amount, and CMS was explicit about why: it wants to recognize greater indirect costs for office-based practitioners than for facility-based ones.

Then on November 21 the CY2026 OPPS and ASC final rule did two things that matter more to a device business than anything in the PFS.

It extended site-neutral payment to drug administration services in excepted off-campus provider-based departments, paying them the Physician Fee Schedule equivalent from January 1. Clinic visits in those departments have been site-neutral since 2019. Drug administration is the new one, and CMS put the cut at $290 million in CY2026, $220 million out of Medicare and $70 million out of what beneficiaries pay in coinsurance. Sole community hospitals are exempt.

The same rule resumed the full phase-out of the Inpatient Only list, starting with 285 procedures removed for CY2026.

CMS declined to extend site-neutrality to on-campus clinic visits this year and issued a request for information instead. In the rule it named where it intends to look next: other APC families, specifically imaging without contrast, and other settings, specifically on-campus outpatient clinic visits.

So the agency has told you, in writing, which revenue pool it is coming for after this one.

Site-neutral is also one of the few things both parties describe as obviously correct while describing it in opposite language. Republicans call it eliminating government waste. Democrats call it stopping hospital consolidation from extracting rents. Same policy. Two constituencies. That is what makes it likely rather than merely sensible.

The commercial move here is unglamorous and it is worth real money. Stress-test every active business case against ASC and office reimbursement instead of hospital outpatient reimbursement, and check whether any of your procedures sit in the 285 coming off the Inpatient Only list. If the case only clears at the facility rate, you have a case with an expiration date on it that your customer's finance team can already see.

3. Premium is getting redefined, and your category lands on one side of it

Here is the same technology described two ways.

Digital pathology, old version: better visualization, sharper images, remote consultation. Digital pathology, new version: more cases read per pathologist in a market where you cannot hire a pathologist.

Nothing about the product changed. The sentence changed, and the sentence is what gets funded. Run that swap across a portfolio and it looks like this.

So the premium category of the next decade is the one that produces the best clinical outcome per unit of scarce labor, and every existing portfolio splits against that test.

Chapter and verse for why this becomes policy and not only procurement.

The FDA already runs a dual track and has for a while. Low-risk digital health, non-device clinical decision support and remote patient monitoring get enforcement discretion and a fast lane. High-risk hardware, PMA devices and complex AI software as a medical device get longer reviews, heavier post-market surveillance and a standing expectation of real-world evidence. The center of gravity moved. A 2027 commercial plan built on a 2021 regulatory template is going to be late.

There is a wrinkle inside that worth naming, because the regulator has a labor problem of its own.

In February 2025 about 700 FDA employees were cut, roughly 220 of them from CDRH. That April another 250 or so CDRH staff got reduction-in-force notices and about 30 were called back, which is another net 220 against a center that started the year around 2,260 people. Across the agency, FDA lost about 21% of its workforce, more than 4,400 people, between September 2024 and January 2026.

Nobody knows the current CDRH number, including, by several accounts, CDRH. The center normally publishes its headcount in the annual report. The 2025 report left it out, and HHS has declined to give one.

Here is the number that made it into daylight anyway. In the first quarter of 2025 CDRH granted three de novo marketing authorizations. In the same quarter of 2024 it granted ten.

The scarce specialties inside the agency are the same ones scarce everywhere else: AI and machine learning, wireless coexistence, interoperability, cybersecurity. Washington simultaneously wants a faster FDA, more AI clearances, breakthrough access, better post-market surveillance and real-world evidence programs, all of which run on exactly that labor.

Which makes MDUFA VI partly a workforce negotiation, and puts industry in an unusual position. We have a direct financial interest in a well-staffed regulator, because regulatory predictability has a dollar value on our side of the table.

4. 2027 and 2028 are a deal window, and medtech is not on the list yet

Everybody I talk to assumes divided government means nothing happens. I think the opposite is true here, and the reason is arithmetic.

CBO published the arithmetic in July. Federal subsidies for health insurance run about $2.4 trillion in 2026, which is 7.4% of GDP, and grow 65% to $3.9 trillion by 2036, which is 8.4% of GDP. That is $33.6 trillion across the decade, of which Medicare is $16.1 trillion, just under half. Subsidy growth outruns projected GDP growth of 46%.

Now the number that turns a budget line into a campaign.

Across that same decade the uninsured population rises from 30 million to 37 million. Subsidies up 65%, and seven million more Americans without coverage at the end of it.

Nobody can run on that. Whoever holds the majority in 2028 has to explain why spending two thirds more bought worse coverage, and no version of that answer is "send the existing system more money." They need pay-fors, and pay-fors require identifiable cost pools that poll badly.

So the pressure here is fiscal rather than ideological, and fiscal pressure produces deals between people who agree on nothing else.

Chapter and verse, and this is the part that surprised me.

Look at what the Ways and Means Committee actually passed on July 15. Seven health bills, most of them out of committee without a single opposing vote, from a Republican-led committee in a year everybody wrote off as gridlock:

Price transparency reaching hospitals, laboratories, imaging providers and ambulatory surgery centers. Medicare Advantage prior authorization reform. A bill making Medicare Advantage plans publish how premium revenue is actually spent. Enhanced Medicare reimbursement for remote physiologic monitoring in rural and underserved areas. Rural anesthesia. Long-term acute care. Family caregiver access.

Medical loss ratio disclosure and prior authorization reform are things you would expect to read in a progressive platform. They came out of Ways and Means, and the committee's own release describes the package as holding health care empires accountable.

Energy and Commerce advanced the Lower Costs, More Transparency Act in the same window. Senate HELP advanced the Patients Deserve Price Tags Act, which reaches further than either House bill and names imaging providers explicitly.

Three committees, two chambers, both parties, one direction of travel.

Five places a deal gets made in 2027 and 2028: PBM and vertical integration reform, Medicare Advantage coding and prior authorization, site-neutral payment, price transparency, and technology and productivity incentives. There is a sixth that both parties need politically, which is rural stabilization, and it is already arriving attached to technology rather than to bigger checks. H.R. 3108 is a remote patient monitoring bill with the word rural in the title. That is the shape of the bargain, and it is a medtech bill whether or not anybody in medtech noticed it pass.

Medtech sits off the target list in all of that. Devices are a modest share of total health expenditure, the clinical innovation is visible, the jobs are American and the patient stories are sympathetic. That protection is real and I would treat it as temporary. Once the frame is affordability rather than coverage, Congress starts walking the value chain asking where the money goes. Hospitals. Insurers. Medicare Advantage. PBMs. Pharma. Physician consolidation. Private equity. Administrative overhead. Then medical technology.

The order matters less than the direction. When the question arrives, the answer that protects a category is a productivity answer, and you either have the data or you do not.

5. Fee-for-widget is the business model with the most exposure

Picture a capital placement that pays for itself on a ten-year proprietary consumable stream and a service contract only the manufacturer can perform. That has been one of the most reliable structures in this industry.

Now put three things next to it. Reference pricing benchmarks for mature high-volume device categories. Antitrust and transparency pressure on bundling capital equipment with procedure fees and locking systems into long-term proprietary consumable contracts. Right to repair rules requiring manufacturers to give hospital biomed departments and independent service organizations the manuals, diagnostic software and parts.

Each one on its own is survivable. Together they take apart the financial logic of the placement.

So a business model that earns its return from the lock rather than from the outcome is carrying policy risk that nothing in the product plan addresses.

Chapter and verse.

The coverage side moved this summer, and almost nobody in commercial noticed.

Transitional Coverage for Emerging Technologies, the breakthrough-device pathway finalized in August 2024, is paused for new candidates as of August 11. CMS said so in the Federal Register notice establishing its replacement, the RAPID coverage pathway, in the plainest sentence a federal agency writes: the TCET pathway will be paused for new candidates upon publication of this notice.

RAPID is a better idea. A proposed national coverage determination issues the same day FDA authorizes the device, final roughly 60 days later for Class II and 90 for Class III. Set that against the roughly five years a device has historically waited for national coverage after authorization, and it is the most consequential thing to happen to medtech market access in a decade.

Read the eligibility before you celebrate. In vitro diagnostics are excluded outright. Devices already inside a pivotal IDE study do not qualify, because RAPID works by aligning CMS with FDA before market authorization. Which means the pathway rewards companies that engaged CMS years earlier than their commercial plan told them to, and offers nothing to anyone already at the finish line.

TCET is the cautionary tale here. It was capped at five devices a year against roughly eight expected nominations. It ran two years. CMS has never published how many it accepted, and exactly one manufacturer publicly announced getting in.

Now put the third thing next to it, because it lands on the same P&L. CMS finalized the repeal of the alternative NTAP and OPPS device pass-through pathways for breakthrough devices. Those let a breakthrough-designated device qualify for separate payment without demonstrating substantial clinical improvement. Devices designated by September 30, 2026 are grandfathered. Everyone else has to prove substantial clinical improvement to get paid separately at all.

So the door to faster coverage opened and the shortcut to separate payment closed, in the same season. Both changes reward the same thing, which is evidence generated early, and both punish the same thing, which is treating evidence as a launch expense. Companies already running RWE infrastructure get years of coverage while everyone else builds registries.

Comments on RAPID close October 13. If your market access team has not read the RAPID notice, that is this week's assignment.

On the state side, the friction is already here and it is specific. California SB 1120 bars health plans from basing medical-necessity decisions solely on AI. Texas SB 1188 requires a practitioner to personally review AI output before a clinical decision. Texas HB 149 requires providers to disclose AI use in treatment. Illinois HB 1806 carries penalties up to $10,000 per violation. Utah and Indiana added insurer-side AI disclosure and anti-downcoding rules effective this year and next. Trackers count roughly 240 health-AI bills across 43 states in 2026.

Add extended producer responsibility packaging laws, PFAS restrictions and right to repair moving in parallel, and a national manufacturer gets a patchwork rather than one rule, which is worse than either a federal standard or no standard at all.

6. The medtech labor market is going the other way, and this part is personal

Two labor markets, moving in opposite directions, inside the same industry.

Healthcare delivery has too few people, so it automates to create capacity. Medtech has too many legacy commercial functions and too few specialized technical ones, so it restructures on one end while bidding aggressively on the other.

Karl Storz is the cleanest example I have seen this year. On June 16 it filed a WARN notice in North Carolina for 108 positions at the Morrisville headquarters of Asensus Surgical, the robotics business it bought in 2024. It is ending development of the Luna platform, winding down Senhance and retiring the Asensus brand.

Now read what it kept. In its own statement the company said the relevant technologies, data and expertise move into the wider portfolio, and about 30 of the 108 roles were engineers. So a hardware program died and most of the cut fell outside engineering, while the software, clinical data and AI capability got absorbed rather than released.

That is the barbell in one filing. The platform bet closed. The capability underneath it got more valuable.

So the middle is where the risk sits.

The left column of that chart is compression rather than elimination. One commercial person with agents, account intelligence and automated customer analysis manages a great deal more economic activity than they did three years ago, and the org chart adjusts to that whether or not anyone announces it.

I would put one more category at the top of the right column. Commercial people who can build and defend an economic transformation case in front of a CFO. There are not many of them, most organizations have three or four, and everyone I know who has them is protecting them.

If you are a rep reading this, that is the whole career instruction. The part of your job that consists of coverage, relationship maintenance and product explanation is getting compressed. The part that consists of constructing an argument about somebody else's operating model is getting more valuable every quarter. Move your hours.

---

What this does to COF

Clinical, Operational and Financial has been our frame for a long time, and it works because it maps to how a hospital actually decides. Clinical value determines whether physicians want it. Operational value determines whether the system can implement it. Financial value determines whether a constrained purchaser will fund it.

The affordability turn adds a fourth column, and it is the one that determines whether Washington protects your category or walks toward it.

That fourth row is one number and it is harder to produce than it looks, because it has to survive a finance review. Here is the chain it has to run.

Labor hours released, times the relevant clinician category, gives you FTE capacity. FTE capacity gives you additional procedures or studies or visits. Those give you incremental revenue, avoided recruitment cost, reduced agency labor, lower overtime, and turnover you did not have to replace. Then the access effects, which are faster time to treatment and lower cost per episode.

That is an ROI model a CFO can use without translating it first. It is also, and I say this as somebody who builds these for a living, about three weeks of work per category and worth every hour, because the same number answers the procurement question, the operational question and the political one.

In DealSMARTS℠ terms the labor question belongs in Discovery, in the first or second call, rather than in Commitment where most teams put it. Asking a CFO how many clinician hours a year they lose to a given workflow is a discovery question. Handing them your calculation of it inside the final business case is a closing argument nobody had a chance to argue with, which is why those decks stall.

---

The Move

This week. Not next planning cycle.

Build the labor line for your top category. One number, one page. How many RN, physician, technologist, physicist, pathologist, coder or scheduler hours does this release per thousand patients, and what does your customer pay for those hours today. If you cannot produce it by Friday, that is the finding.

Rewrite one value proposition in output-per-clinician terms. Take your strongest product and rewrite your opening sentence so the unit is a scarce person rather than a clinical delta. Run both versions at two accounts this month and watch which one gets a second meeting. It costs nothing and it settles an argument your marketing team has been having for a year.

Stress-test every open business case at ASC and office rates, and check whether any of your procedures sit in the 285 coming off the Inpatient Only list. If a case only clears at the hospital outpatient rate, rebuild it now. Your customer's finance team is already modeling the other number.

Ask the workflow question before you ask for the order. Automating 20% of a task removes nothing from the cost line unless the operating model changes with it, and a point solution dropped into an unchanged workflow can leave your customer worse off. So ask what has to change in the schedule, the staffing model and the handoffs for the savings to appear, then put the answer in your proposal.

Find out who in your account owns the labor budget. It is usually not the person you call on. Agency spend, overtime and recruitment sit with a CNO, a COO or a chief human resources officer, while the technology that saves those dollars gets bought out of a capital budget owned by somebody else. Name every person who has to act, mark who owns each name, and the unowned rows are your exposure.

Leaders, count your bench. How many people in your commercial organization can walk into a CFO's office and build a labor and capacity argument without help. If the answer is under five, you have a hiring problem and a development problem at once, and the development one is faster to fix.

One for the next twelve months. Start collecting the real-world evidence a coverage bargain will require before it is struck. If a statutory breakthrough pathway arrives with mandatory RWE attached, companies already running that infrastructure get years of Medicare coverage while everyone else builds registries.

---

American healthcare has a productivity problem, and the cost problem everyone argues about is its shadow. Twenty years, negative productivity in clinical care, $150 billion a year of IT spending, 55 points of gain everywhere else in the service economy.

That gap is now large enough and expensive enough to become an election issue, and when it does, every technology in this industry gets sorted into one of two piles. Things that help produce a successful episode of care with fewer scarce resources. Things that add a line to the bill.

Medtech is the one part of the healthcare value chain that can genuinely land in the first pile. Almost nobody in medtech is currently measuring themselves that way.

Your competitor is going to figure this out. The only real question is whether it happens before or after your next contract cycle.

---

Dr. Gunter Wessels is the founder of LiquidSMARTS℠, a commercial engineering firm for medical technology companies. LiquidSMARTS℠ guarantees 10% pipeline velocity improvement in 90 days.

Originally published on LinkedIn.