Workforce Intelligence SaaS for Nursing Homes After the Staffing Mandate Died

The Problem

Ask a nursing home owner what keeps them up at night. It is not the roof. It is not the kitchen. It is labor: the nurses who may or may not show up for the night shift, and what that uncertainty costs. Direct-care nursing expense across skilled nursing hit $49.2 billion in 2023, according to Ankura's September 2026 analysis, making it simultaneously the largest cost line, the largest compliance surface, and the largest driver of the quality scores that determine referral flow and reimbursement, which means the labor decision is the business decision. Who wins in post-acute care is, to a first approximation, the question of who manages labor better, and almost nobody manages it with data.

The agency trap is the mechanism. When a shift goes unfilled, the facility calls a staffing agency and pays a premium of roughly 55 to 70 percent over employed staff. Every purchased hour makes the next one more likely. Facilities in the top quartile of agency use show 7.7 percentage points higher permanent RN turnover than their peers, so the purchased hour degrades the workforce that would have prevented the next purchase. Contract labor usage rose 12.2 percent from 2019 levels specifically among facilities with operating margins of negative 4 percent or worse. Weak margins force agency use. Agency use weakens margins further. Ankura calls it a doom loop, and it is the mechanism behind a meaningful share of closures.

Here is where quality closes the circle: Covr's August 2026 analysis of CMS Payroll Based Journal data found that facilities holding 3.75 to 4.5 nursing hours per resident day see consistent quality outcomes while star ratings trend downward as agency use increases. National agency use averages 5.7 percent, more than ten times the median because a small group of high-agency facilities skews the mean, while high performers run under 1 percent. An unfilled shift becomes a survey citation, becomes a rating, becomes a lost referral, becomes an empty bed, becomes the margin that funded the staffing in the first place.

Here is the absurd part: every hour of this story is already recorded. Since 2016, every Medicare and Medicaid certified nursing home has submitted daily staffing hours to CMS through the Payroll Based Journal system, payroll-verified and auditable, which means the single most important number in the business has been collected, standardized, and published for a decade while almost nobody in the industry has looked at it. CMS publishes the data quarterly on data.cms.gov. Anyone can download it. No cost center in American health care is more measured than the industry's $49.2 billion labor line. And the least analyzed.

The Mandate Died. The Math Didn't.

An obvious objection: the federal staffing mandate would have forced all of this. It is gone. True, and the timeline matters: the mandate's twenty-two-month life is what cleared the field for the economics to take over. CMS finalized the rule in April 2024: 3.48 total nursing hours per resident day, including 0.55 RN and 2.45 nurse aide hours, plus an RN on site around the clock. Industry analysis put the price at $6.5 billion a year and 102,000 additional clinicians, with 79 percent of facilities needing to increase staffing, four in five unable to meet the around-the-clock RN requirement, and roughly 290,000 residents at risk of displacement if buildings closed beds they could not staff. In April 2025 a federal judge in Texas vacated the core provisions. An Iowa judge followed in June. Signed July 4, 2025, the One Big Beautiful Bill Act barred CMS from implementing or enforcing the staffing provisions until September 30, 2034, a ten-year moratorium that outlasts most administrators' careers and effectively repealed the rule years before CMS admitted it. HHS withdrew its appeals that September, and CMS published a repeal rule in December 2025, effective February 2, 2026: the mandate is dead, twenty-two months after its final rule.

What survived the repeal is more durable than the mandate was. Fines. Stars. Withholds. State minimums are still law and still enforced: New York requires 3.5 hours per resident day and computes compliance quarterly straight from PBJ data, fining violators up to $2,000 per day. California requires 3.5 hours with fines up to $50,000 and has cited more than 400 homes since 2021. CMS's Five-Star staffing domain still grades every facility on six PBJ-derived measures, and staffing stars move the overall rating that discharge planners read, which means the public dataset still prices every building's reputation every quarter whether the building looks at it or not. Skilled Nursing Value-Based Purchasing still withholds 2 percent of Medicare Part A payments, and starting with the FY2026 program year the score includes total nurse staffing hours and turnover, both drawn from PBJ. That same reconciliation law also cut federal Medicaid spending by more than $9 billion through 2035, per KFF's December 2025 analysis, and Medicaid is the primary payer for 63 percent of residents. Reimbursement is being squeezed while labor stays fixed. At least the mandate told facilities how many hours to buy; the surviving system punishes them financially for buying those hours badly.

The Gap in the Market

Software for nursing home staffing exists. Software that turns the public staffing record into a profit-and-loss instrument does not, and the gap between those two facts is the entire business.

Company	What They Do	What's Missing
SimpleLTC / SimplePBJ (Netsmart)	The incumbent PBJ tool: assembles payroll and timekeeping files, validates against CMS rules, submits to CMS, predicts Staffing Five-Star ratings, runs what-if scenarios. Offers a free benchmarking tier (SimplePBJ Essentials) as a funnel into the paid submission product.	It is a filing tool for the administrator, not a P&L tool for the owner: benchmarking is a lead magnet for the filing product rather than a P&L instrument, and there is no labor-cost-per-resident-day economics, no agency-premium leakage math, no state fine-exposure computation, no acquisition diligence mode, and no VBP dollar modeling. SimpleLTC's free tier proves the dataset is accessible; it does not monetize the decisions the dataset enables.
OnShift, SmartLinx	Workforce management, scheduling, timekeeping, attendance, and PBJ export for nursing homes, embedded in daily operations.	They see one facility's schedule, not the market, so there is no cross-facility benchmarking from public PBJ and no agency-spend analytics against county percentiles; their roadmap serves the scheduler, not the CFO.
ShiftKey, IntelyCare, Clipboard Health	Marketplace platforms that fill open shifts with gig nurses and aides: fast-growing, well-funded, and genuinely useful in a shortage.	They profit from the agency premium this product helps eliminate, so a marketplace will never tell an operator that 9 percent agency usage costs $180,000 a year in premium or that converting half of it to employed staff pays back in seven months: a structural conflict of interest.
Covr	Scheduling plus workforce analytics; published the August 2026 PBJ benchmark report (3.75-4.5 HPRD quality band, agency thresholds, overtime-turnover link) from 380 customer facilities.	A scheduling vendor with a research arm rather than a standalone intelligence product: their benchmarks describe their own customers, and the opportunity is the independent, all-facility version sold to owners, lenders, and acquirers who will never buy a scheduler.
PointClickCare	The dominant SNF EHR: clinical documentation, billing, and analytics modules across most of the industry.	It is the clinical system of record, not a labor-economics platform, with no PBJ-derived market benchmarking, no fine-exposure engine, and no diligence scoring; it sells to operators, not to the capital side of the market.
CMS Care Compare / Provider Data Catalog	Publishes the raw PBJ files, Five-Star ratings, and penalty data quarterly: free and authoritative.	Raw CSVs rather than answers: a 237-megabyte quarterly staffing file with 1.3 million rows is not a product, and nobody at a 120-bed facility in Fresno is joining it to BLS wage data to compute their agency leakage.

Across the market, the pattern is the same: vendors sell filing to administrators (SimpleLTC), scheduling to schedulers (OnShift, SmartLinx, Covr), shifts to desperate directors of nursing (the marketplaces), or charts to executives (PointClickCare), and every one of those products answers a different question than the one the owner asks at the P&L review. Nobody sells the owner's question. Is the $49.2 billion being spent well? What answers it is a dataset that is public, quarterly, and machine-readable. That combination, a public dataset plus an unserved buyer, is the entire opening.

The Solution

A workforce intelligence platform built on public PBJ data, sold to the people who own the P&L rather than the people who file the reports: multi-facility operators, independent owners, private-equity acquirers, REITs, and lenders, organized around six components:

1. Labor P&L benchmark: Every facility's PBJ hours by role, multiplied by county-level wage data, producing a true labor cost per resident day, benchmarked against county, state, and peer percentiles. For the first time, the operator sees what their labor actually costs per unit of output. And where they sit in the distribution. Most administrators know their hours; almost none know their cost per resident day relative to the building across town.

2. Agency leakage calculator: Agency hours as a share of total hours, times the 55 to 70 percent market premium, equals dollars walking out the door every year. Next, the module prices the conversion: what it costs to hire the FTEs that would replace the agency hours (wages, benefits, recruiting) against the premium currently being paid, with a payback period. Covr's benchmarks supply the targets: under 5 percent agency is competent, under 1 percent is elite, and the national mean of 5.7 percent is ten times the median because the tail is so bad.

3. Five-Star and VBP impact modeler: CMS's staffing star methodology is public, so the platform predicts next quarter's staffing stars from current PBJ data before CMS publishes them, and models what-if scenarios in dollars. Hire three CNAs and the model shows the expected star movement, the VBP multiplier effect on the 2 percent Medicare withhold, and the referral value of the rating change. CMS has already published the FY2028 VBP performance standards (3.29 threshold, 5.87 benchmark on total nurse staffing), so the targets are known two years out, which is an extraordinary thing in health-care regulation: a payment formula whose finish line is visible before the race starts, and nobody has productized the training plan.

4. State fine-exposure engine: New York determines staffing-law compliance quarterly using PBJ data, which means a facility's fine exposure is computable from the same public file the state uses, before the state sends the letter. Under every state regime with a staffing minimum and a fine schedule, currently led by New York's $2,000-a-day penalties and California's $50,000 fines, liability comes pre-computed from the same public files the states use, and the 50-state mandate database tracks legislatures as they move. After the federal repeal, the states are where enforcement lives.

5. PBJ audit and rating recovery: Facilities routinely under-report hours they actually worked: a McKnight's review of PBJ filings prepared for legal defense found paid, worked hours missing from submissions as a matter of course, including corporate staff on site in reportable roles, salaried overtime never logged, and contract hours the vendor never sent. Every missing hour lowers the staffing stars that drive referrals. Payroll gets reconciled against PBJ submissions by the audit module, which finds the missing hours and supports corrected filings. This is found revenue: the hours were already paid for.

6. Diligence mode: For acquirers and lenders, a staffing-risk score on any facility or portfolio: turnover trends, agency share, weekend staffing drop-off, the gap between reported hours and star-implied hours, and citation history correlated with staffing levels. In a consolidating market where staffing predicts survey risk and survey risk predicts cash flow, the diligence memo writes itself from public data, and the firm that can score a twelve-facility portfolio before the LOI has an edge no broker's offering memorandum can match.

The Original Calculation: What the Agency Premium Actually Costs

Take a representative 100-bed facility at 85 percent occupancy. That is 85 residents. At 3.8 total nursing hours per resident day, the building burns through roughly 118,000 nursing hours a year. Every year. Put agency usage at 8 percent of hours, worse than the 5.7 percent national mean but common in the struggling tail: about 9,440 agency hours a year.

Price the premium at 60 percent over a $32 blended employed hourly cost, the midpoint of Ankura's 55 to 70 percent range. Above what employed staff would have cost for the same hours, the agency premium alone runs roughly $181,000 a year, which is more than the salary of the director of nursing the building cannot afford to hire because it is spending the money on agency instead. Convert half those agency hours to employed FTEs and the building keeps about $90,000 annually, every year, while also cutting the turnover driver that Covr's data ties to agency reliance and weakening the doom loop that Ankura identifies behind a meaningful share of closures. Recruiting and signing bonuses for five or six FTEs pay back in months. Not years.

Now the penalty math: New York's $2,000-a-day fine over a 90-day quarter is $180,000 of exposure per quarter for a noncompliant building, and the state's actual enforcement averaged about $178,000 per fined home across the $8 million assessed to date, which is not a theoretical number but last quarter's invoices.

And the Medicare math: a facility with $4 million in annual Medicare Part A revenue has $80,000 withheld under VBP's 2 percent take. How much comes back is determined by the staffing and turnover measures now in the score; CMS keeps roughly 40 percent of the total withhold pool regardless, so the program is a net tax on the industry that only the better-staffed buildings partially recover. On the commercial side, the Five-Star staffing rating feeds the overall stars, and discharge planners in narrowing networks steer toward four- and five-star buildings, so a one-star staffing deficit becomes an occupancy deficit with a lag, and occupancy is the revenue line the whole P&L rests on.

Revenue Model

Revenue Stream	Amount	Notes
Operator SaaS (per facility, monthly)	$450-650	Tiered by bed count with an annual prepay discount, covering the labor P&L benchmark, agency leakage calculator, Five-Star predictor, state fine-exposure monitor, and quarterly PBJ refresh; the buyer is the owner or regional director, not the administrator.
Portfolio / enterprise (annual)	$35,000-60,000	Multi-facility operators: cross-building benchmarking, portfolio agency-spend rollup, acquisition target screening, board-ready quarterly labor reports. Priced above the per-facility floor so it never undercuts the standard tiers.
Diligence seats (per firm, monthly)	$2,000	PE firms, REITs, lenders, brokers: staffing-risk scoring on any facility or portfolio, comp sets, turnover and agency trend screens for LOI-stage underwriting. High willingness to pay; the alternative is a consultant at $400 an hour.
PBJ audit and rating recovery (one-time)	$7,500	Payroll-to-PBJ reconciliation, missing-hours recovery, and corrected filing support, priced against found revenue since recovered stars move referrals within two quarters, and a natural land-and-expand into the SaaS subscription.
State association white-label (annual)	$25,000-50,000	State health care associations license a branded benchmark portal for members. Distribution channel disguised as a revenue line: every member login is a qualified SaaS lead.

Unit economics on a 30-facility regional operator: 30 facilities at $550 a month is $198,000 a year in SaaS. CAC runs $20,000 to $30,000: the owner holds the budget but buys on annual cycles, and the sale needs the CFO, not just the administrator, so cycles run 9 to 12 months. At 36-month retention the LTV is $594,000 against a $20,000 to $30,000 CAC, an LTV to CAC ratio above 20x on 85 percent gross margins, because the core dataset is free and the compute is a quarterly batch job rather than a real-time stream.

Market Size

TAM: Start with the certified base: 14,693 facilities participating in Medicare or Medicaid as of July 2025, roughly 15,000 in HHS's current count, which makes this one of the rare health-care markets where the entire customer universe is enumerated, addressable, and already filing the data the product runs on. At a $6,000 average annual contract (blended across single-facility SaaS and portfolio pricing), operator software is $88M a year. Add diligence seats: roughly 250 active PE firms, REITs, lenders, and brokerages touching SNF assets at $36,000 a year is $9M, because every one of them currently pays consultants $400 an hour for worse versions of this analysis built in Excel from the same public files. Add state association white-labels at 30 states times $40,000, another $1.2M, plus the PBJ audit product at 800 audits a year times $7,500, $6M, for a total of approximately $104M a year.

SAM: The serviceable market is facilities where staffing already carries a price tag: states with enforced minimums and fine schedules (New York, California, Massachusetts, New Jersey, Illinois, Pennsylvania and a half-dozen more), plus multi-facility operators and the capital side everywhere. Roughly 5,000 facilities at $6,000 is $30M, plus diligence and audit concentrated in the same footprint at $6M: $36M a year.

SOM (year 3): 450 operator facilities at $5,400 average is $2.43M. Thirty diligence seats at $24,000 is $720,000, for a total of approximately $3.15M ARR: under 10 percent of the SAM facility count, achievable through state association partnerships and the audit product's land-and-expand motion.

Why Now

The repeal removed the deadline, not the cost. February 2026 ended the compliance conversation when the federal mandate died, and left the economics untouched: $49.2 billion in direct-care labor, a 55 to 70 percent agency premium, and a turnover doom loop that survives any administration. Meanwhile the law that killed the mandate cut more than $9 billion in federal Medicaid through 2035, per KFF's December 2025 analysis, and Medicaid is the primary payer for 63 percent of residents, which means reimbursement is tightening around the one cost line that cannot be cut without cutting care. Reimbursement is tightening while labor is fixed, which makes labor efficiency the only controllable margin lever. That is a better forcing function than a mandate, because it never expires.

Staffing now moves Medicare dollars directly: the VBP program's 2 percent withhold is not a rating, it is cash, and the FY2026 program year put total nurse staffing hours and turnover into the score. With FY2028 performance standards already published, operators can see the targets two years out. Every facility's PBJ data determines its earn-back. Nobody has productized the what-if math.

States are fining real money computed from public data: New York's $8 million in assessed penalties came from the same PBJ files anyone can download, which means fine exposure is a calculation, not a surprise. California's $50,000 fines and 400-plus citations since 2021 say the same thing. Post-repeal, enforcement migrated to the states, and the states publish their inputs.

Finally, the benchmarks exist: Covr's August 2026 analysis turned PBJ data into an operating playbook with numbers: 3.75 to 4.5 HPRD for quality, under 5 percent agency for competence, under 1 percent for elite, 3 to 7.5 percent overtime, under 50 percent turnover. A year ago the industry argued about philosophy; now it can argue about percentiles, which is what software is for.

Public, quarterly, and machine-readable: CMS publishes daily nurse staffing every quarter through a stable API and bulk CSV, so the moat is not access to the data but the modeling layer: case-mix replication, wage joining, fine-schedule encoding, VBP simulation. That layer is buildable by a ten-person startup and ownable before the incumbents notice, because the incumbents sell filing and scheduling rather than economics, and enterprise health-care software ships analytics in quarters rather than weeks, which turns a two-year head start on the modeling layer into a genuine lead.

SNF assets keep trading. Every buyer underwrites staffing risk, whether they model it or not. Consolidation needs diligence. Turnover, agency share, and weekend drop-off predict citations; citations predict CMPs and star declines; star declines predict occupancy. A diligence product that scores all of it from public data, before the LOI. It sells into every deal.

Startup Costs

Category	Cost	Notes
PBJ data pipeline + ETL	$120K	Quarterly ingestion of daily staffing files, provider info, penalties, and Five-Star crosswalks; facility identity resolution across CMS file vintages; historical backfill to 2017 for trend features, staffed by two data engineers over four months.
Five-Star and VBP model replication	$80K	Reimplement CMS's published case-mix adjustment and staffing-star cut points, plus VBP scoring simulation against the published FY2028 standards, because the model must reproduce CMS's own stars within rounding before it is allowed to predict anything. Validated against published ratings on historical data.
State mandate and fine database	$45K	50-state staffing minimums, fine schedules, waiver provisions, and enforcement history, kept current by a part-time regulatory researcher. New York and California first; the rest follow the enforcement money.
Application engineering (6 months)	$180K	Six months of work by two full-stack engineers covering facility scorecards, the agency leakage calculator, the what-if modeler, diligence screens, and portfolio rollups; the dataset is a quarterly batch, so the app is standard SaaS.
Pilot program (20 facilities)	$35K	Free platform plus PBJ audits for 20 buildings across 4 states. Goal: recovered-star case studies and measured agency-conversion ROI the sales deck can quote.
Go-to-market (year 1)	$40K	State association conferences (AHCA/LeadingAge state chapters), the annual "State of SNF Labor" report built from PBJ data, PE/REIT/BD outreach through diligence seats. No field sales team in year one. One enterprise AE adds ~$160K fully loaded.
Legal and insurance	$25K	Data-licensing review (CMS data is public but the terms need reading), E&O for a product whose scores appear in acquisition memos, state fine-schedule legal verification.
Cloud infrastructure and operating buffer (12 months)	$30K	Quarterly batch compute over ~1.3M-row staffing files is trivial load, covering hosting, monitoring, and support tooling.
Total	$555K

Post-launch burn runs roughly $38,000 to $45,000 a month. On a lean team of five. Against the year-3 SOM of $3.15M ARR at 85 percent gross margin, break-even lands around month 32, assuming the owner-level sale closes as modeled, the audit product feeds the subscription pipeline, and the 9-to-12-month sales cycle does not stretch further under Medicaid pressure.

Limitations

PBJ data lags reality by a quarter because CMS publishes each quarter's staffing files months after the quarter closes, so the platform is a strategic instrument rather than a real-time alerting system. Facilities managing week-to-week crises still need their scheduler. This product tells them whether the crisis is structural.

Replicating CMS's case-mix adjustment is approximate: the published methodology is detailed, but the resident-level PDPM inputs are not all public, so predicted stars will deviate from published stars at the margins. As a validation gate, the model must reproduce CMS's own ratings within rounding on historical data, which bounds the error, but a prediction is still not a publication, and the product must say so wherever the two could be confused.

Labor cost modeling uses county wage data, not the facility's actual payroll. BLS occupational wages by metro are a good proxy for the distribution but a specific building's costs will differ, which means the leakage calculator is directional until the operator connects actual payroll, an integration worth building in year two.

State fine regimes have waiver valves: California granted 236 workforce-shortage waivers, and New York's governor declared shortage conditions that let homes petition for reduced fines, so fine exposure is real but not automatic. A product that cries $180,000 at every building will be ignored, which is why the engine must model waiver likelihood rather than just the statute: a fine the state will waive is not a fine.

Price sensitivity cuts exactly the wrong way here. Independent operators run thin margins. They have been burned by software that promised savings. Part of the audit product's job is solving this: $7,500 against found revenue is an easier yes than $550 a month against projected savings, and every audit converts at a high rate once the missing hours appear.

SimpleLTC's free tier sets the price anchor at zero for benchmarking. This product must stay clearly on the other side of the filing-versus-economics line: SimpleLTC answers "did we file correctly," this answers "are we spending well." If the two blur in the buyer's mind, the free tier wins.

Strongest Counterargument

The mandate's death removed the only forcing function this market ever had. Without it, nobody buys. Nursing home administrators are drowning in surveys, turnover, and census pressure. They do not wake up wanting a labor P&L dashboard. Two layers removed from the data sits the economic buyer, the owner or CFO, who historically buys software the way they buy everything else: whatever the administrator already uses. SimpleLTC's free benchmarking plus the existing scheduler covers 80 percent of the perceived need at zero marginal cost, and "workforce intelligence" is a feature that OnShift or PointClickCare ships in a quarter, not a company. Post-repeal, with Medicaid facing $9 billion in cuts, operators will cut software. Before they buy it.

The uncomfortable version: this is a vitamin dressed as a painkiller. When staffing economics get hard, the repeal proved, the industry's response is litigation and waivers, not optimization software. New York fined 45 homes $8 million. Half the state's buildings still miss the standard. If $2,000-a-day fines do not change behavior, a dashboard will not either.

Start the honest rebuttal by conceding the administrator sale is lost. Administrators buy filing tools; this product is not for them, and pitching it as one repeats SimpleLTC's game on their home field. Picture the buyer: the owner looking at a $181,000 annual agency premium on one building, the regional CFO rolling up thirty, and the PE associate underwriting a twelve-facility acquisition. Those buyers already pay consultants $400 an hour for worse versions of this analysis built in Excel from the same public files. Forcing functions here are not mandates but money: VBP withholds that show up as reduced Medicare payments, NY fines that arrive as invoices, star ratings that move referrals within two quarters of publication. None of them require a federal mandate to bite. As for the incumbents, filing vendors optimize for submission accuracy and scheduling vendors optimize for shift coverage; neither has ever sold a dollar-denominated answer to an owner, because their buyers never asked the question. Feature risk is real but slow: enterprise health-care software ships analytics quarters, not weeks, and a two-year head start on the modeling layer plus the state fine database is a genuine lead. Against the SaaS tier the vitamin objection has the most force, and the least against the audit and diligence products, which is why the business leads with found revenue and deal underwriting and lets the subscription follow the proof.

What You Can Do

If you operate nursing homes: Download your own PBJ file from data.cms.gov and compute two numbers: agency hours as a share of total hours, and total HPRD against Covr's 3.75 to 4.5 quality band. If agency is above 5 percent, multiply the agency hours by 60 percent of your blended wage. That is your leakage. One multiplication. No consultant required. Then check whether your submitted hours include corporate staff time on site and salaried overtime; McKnight's legal-defense review suggests they often do not, and every missing hour is a star you already paid for.

If you own, lend to, or are buying facilities: Screen every target's PBJ trend before the LOI: three quarters of rising agency share plus falling weekend HPRD is the signature of a building heading for a citation cycle. Price the staffing risk into the bid the way you price deferred maintenance, because it behaves the same way.

If you're a builder evaluating this space: The wedge is the New York fine-exposure calculator. It is computable entirely from public data, it names a dollar figure the state will eventually name anyway, and it sells to owners in New York, where the fines are already flowing. Build that, prove the model reproduces published stars, then expand to the P&L layer. SimpleLTC proved the dataset is accessible; nobody has proved the economics are sellable, and that proof is the company.

The Bottom Line

The federal staffing mandate lasted twenty-two months from final rule to repeal, and its death changed nothing about the underlying business. Fifteen thousand nursing homes still spend $49.2 billion a year on direct-care labor, still pay a 55 to 70 percent premium for agency hours, still lose staff to the turnover the premium causes, and still get graded, fined, and paid on staffing numbers the government publishes for free. New York turned the public dataset into $8 million of fines and Medicare turned it into a 2 percent withhold, while the industry's response so far has been to file the data and ignore what it says. Every hour is recorded, every quarter, for every building in America, and the layer that turns those hours into dollars does not exist yet. The mandate told facilities what to spend. Somebody still has to tell them whether it worked.