How an independent specialty group walks into a payer negotiation
Twelve surgeons. Eight quarters of remits. Three CPTs that pay for the building. A walkthrough of the four prompts an independent orthopedic group runs to answer the three questions every CEO has, and the negotiation memo that comes out the other side.
8.4 pp
Drift caught on one CPT
2x
Same-payer hospital spread, 8 miles apart
2 targets
Self-funded employers in service area
28 min
Open file to negotiation memo
The three questions
Every independent specialty CEO is wrestling with the same three questions right now. They show up almost word-for-word in conversations with orthopedic groups, GI groups, cardiology groups, women's health groups, ophthalmology groups. The framing is consistent because the economics are consistent.
What every independent specialty CEO is asking
What should our payers be paying us?
Given our specific surgeons, our specialty, and our market. Not a national average. Not what the broker said last year.
Are they paying what they said they would?
Against either the contracted rate or a defensible benchmark. Is any payer silently walking the rate down.
Is there a carve-out worth having?
Who are the self-funded employers in our service area, what are they paying their carrier and their broker for the network we're already in.
The case study below is a 12-physician orthopedic group with three offices on the central California coast. Five joint-replacement surgeons, three subspecialty surgeons (hand, foot and ankle, spine), and the rest split across sports and shoulder. The largest independent ortho group in its region. The questions they asked are the universal independent-CEO questions. The numbers are theirs.
How it's done today
The independent group's answer to all three questions today runs through a handful of tools that sit at very different price points. Each one answers part of the question. None of them answers all of it. None of them talks to the others. The lower-cost rows are what an independent group actually uses. The higher-cost rows are what payers, large systems, and PE-owned MSOs use, and they sit out of reach for most groups.
Broker market check
Free, but anchoredYour benefits broker pulls a market sample on request. The sample is anchored to the broker's own book of business, which is mostly employer-side, not provider-side. Useful as a directional input, not as a benchmark.
AAOS or specialty-society comp survey
~$1K-$3K / yearNational, self-reported, lagging by 18 to 24 months. Tells you what surgeons of similar tenure earn. Does not tell you what your payers should be paying you.
PMS variance reports
Bundled in PMSMost modern PMS platforms surface contracted-versus-allowed at the line level. Useful for catching short-pays. Does not detect rate-walk-down across quarters because no PMS report shows percent-of-current-Medicare by quarter.
Hospital MRFs (price transparency files)
Free, but rawEvery hospital has been required to publish negotiated rates with every payer since 2021. Most publish in a buried JSON file. Reading them at scale is a data engineering project.
MRF aggregator SaaS (Turquoise, Serif, Trilliant)
$30K-$150K / yearA handful of vendors have built dashboards on top of the hospital MRFs. They normalize the JSON and expose negotiated rates by payer and CPT through a UI or API. Pricing is enterprise. Sold to payers, large health systems, and PE-owned MSOs. Independent groups rarely buy it because the seat cost exceeds the recovery on a single negotiation cycle.
Form 5500 filings
Free, but scatteredEvery self-funded employer files annually with DOL. Schedule A names the carrier and broker compensation. Schedule C names directly-paid service providers. The data exists, but no one outside the benefits consulting world reads it.
Form 5500 prospecting tools (Judy Diamond, Larkspur, BrightScope)
$5K-$15K / seat / yearTools built for benefit brokers. They surface Schedule A and C filings as searchable employer leads, with carrier history and broker comp. Strong for outbound to self-funded employers. Not built for the provider-side question of which payer rates to anchor against, and they do not sit alongside hospital MRF data.
Outside consultant
$15K-$60K / engagementA managed care consultant pulls remit data, builds a benchmark deck, charges by the engagement. Good output, slow turnaround, hard to iterate, expensive enough that most groups do it once every few years rather than continuously.
The realistic workflow for an independent group is a broker market check before every contract cycle, a PMS variance report when AR slips, and a consultant engagement once every three to five years when the contract is up for major rework. Most independent groups never read a hospital MRF and have never opened a Form 5500 filing. The SaaS dashboards that aggregate them sit in payer and large-system budgets at five-figure annual contracts, not in group-practice budgets.
Question 1: What should our payers be paying us
Two prompts answer this one. The surgeon side first. The hospital side second. Both fit on a page.
The surgeon side
You type
“What should Aetna be paying my surgeons for a knee replacement? This is my roster on the csv.”
What Medistill does under the hood
- Reads the 12 NPIs out of the attached roster.
- Routes to the physician fee benchmark tool, which wraps the CMS Medicare Part B Public Use File. Per-NPI allowed amount on CPT 27447, locality-adjusted, case-mix-adjusted to each surgeon's actual practice.
- Computes the group-weighted Medicare anchor across the surgeons who clear the CMS reporting threshold of 11 cases per year.
- Overlays the published commercial band for orthopedic surgery (130% to 180% of Medicare).
Output for this group, on CPT 27447 (total knee arthroplasty):
Surgeon-side benchmark, CPT 27447
The number you want in your head walking into the conversation is $1,952 as the ask, and $1,691 as the floor you do not go below. Anchored to your surgeons' actual Medicare allowables, not a guess and not what the broker said last year.
The hospital side
Same payer, same procedure, every hospital in the service area. The hospital MRFs answer this directly. The query is one line:
You type
“What does Aetna pay hospitals around San Luis Obispo and Santa Maria for a knee replacement? Show the range, low to high and median.”
Five hospitals come back. Three on case rate, two on per diem (DRG 470 inpatient, not directly comparable):
| Hospital | System | Aetna rate | Setting |
|---|---|---|---|
| French Hospital Medical Center | Dignity | $6,602–$7,087 | Outpatient |
| Marian Regional Medical Center | Dignity | $13,747 | Outpatient |
| Arroyo Grande Community Hospital | Dignity | $13,747 | Outpatient |
| Sierra Vista Regional Medical Center | Tenet | $26,311 | Inpatient (DRG 470) |
| Twin Cities Community | Tenet | $43,901 | Per diem |
The headline read: Aetna pays French Hospital $6,602 for a knee. Aetna pays Marian Regional $13,747 for the same knee. Same payer, same procedure, both Dignity Health, eight miles apart on Highway 101. That is a 2x spread on the facility line of the all-in episode.
The total-cost-of-care framing matters here. Aetna pays the hospital between $6,602 and $26,311 to host a single knee replacement in this service area. Aetna pays the surgeon performing that knee approximately $1,250. Even at the lowest facility rate, the facility-to-professional ratio is 5:1. At the highest, it is over 21:1. A 10 to 15 percent professional fee increase is a rounding error in Aetna's per-episode cost. That is the right framing for the negotiating table.
Question 2: Are they actually paying it
This is the question most vendors get wrong, because to answer it you need to look at the practice's claims data, and the practice's claims data is PHI. Most vendors solve that by signing a Business Associate Agreement and ingesting the data. Medistill solves it by not taking the data.
The architecture
Your PMS
Athena, ModMed, NextGen, etc. Five clicks for the claims export.
Scrubbed data
Strips 18 Safe Harbor identifiers per 45 CFR 164.514(b). Runs locally on the practice's laptop.
Claude Desktop
Holds the de-identified file. Runs the variance math locally. Never sends your file out.
Medistill MCP
Returns market benchmarks, Medicare anchors, payer behavior. Never sees your file.
Once the file has been de-identified per Safe Harbor it is not PHI under federal law. HIPAA stops applying. That is the legal mechanism that makes the workflow clean. No BAA, no data egress, no procurement cycle. The practice scrubs locally and runs the export anytime.
You type
“For the attached file, flag any payer × CPT where the rate is trending down quarter-over-quarter from 2024Q1 to 2025Q4. Use Medicare as the benchmark.”
The agent loads the 24 months of remits, computes average allowed amount per payer per CPT per quarter, pulls the Medicare published rate per CPT through the connector, runs a linear regression on each payer-CPT pair across the eight quarters, and flags any pair with statistically significant negative slope and end-to-end decline above the noise floor.
Of 94 payer-CPT pairs that met the volume threshold, exactly three flagged. All three are Aetna. All three are major joint replacements:
| Pair | 2024Q1 | 2025Q4 | Δ | R² | % MC drift |
|---|---|---|---|---|---|
| Aetna × 27447 (knee) | $1,351 | $1,250 | −7.5% | 0.999 | 122% → 113% |
| Aetna × 27130 (hip) | $1,270 | $1,173 | −7.7% | 0.999 | 118% → 109% |
| Aetna × 23472 (shoulder) | $1,439 | $1,391 | −3.3% | 0.987 | 122% → 118% |
R-squared above 0.98 is the part that should not be glossed over. The rate is moving in a near-perfectly straight line down, quarter after quarter, by approximately the same dollar amount each time. That is not normal variation, payer-mix shift, or modifier drift. It is an automated fee schedule decrement.
Cross-payer comparison across the same three CPTs and the same eight quarters makes the contrast unambiguous:
| Payer | CPT 27447 Δ | CPT 27130 Δ | CPT 23472 Δ |
|---|---|---|---|
| Aetna | −7.5% | −7.7% | −3.3% |
| Anthem Blue Cross CA | +0.1% | +0.8% | +0.1% |
| Blue Shield of California | −0.1% | −0.8% | −0.2% |
| UnitedHealthcare | −0.4% | −0.5% | +0.3% |
| Cigna | −0.5% | +1.2% | +0.4% |
| Medicare FFS (benchmark) | 0.0% | 0.0% | 0.0% |
CMS did not change the fee schedule for these three codes in this period. None of the practice's other commercial payers reduced rates. The cuts are Aetna-specific, and they target precisely the three codes with the highest dollar exposure for the practice. Not the E&M codes (99213/99214). Not the joint injections (20610). Not the minor surgical codes.
Annualized financial exposure
| CPT | 2025 Aetna cases | Per-case erosion | Annualized |
|---|---|---|---|
| 27447 (knee) | 98 | $101 | $9,901 |
| 27130 (hip) | 53 | $97 | $5,154 |
| 23472 (shoulder) | 52 | $48 | $2,502 |
| Total (3 codes) | 203 | $17,557 |
$17,557 is the run-rate annual hit already absorbed in 2025 from these three codes alone. If the slope continues through 2026Q4, the group's rates land at 4 to 8 percentage points above Medicare on its three core surgical codes. That is a position no orthopedic group of this scale and quality should accept, and it is not what they signed up for at contract anniversary.
Question 3: Can we go around the network
Direct-to-employer contracting is the third question every independent specialty CEO is asking. The reason most groups never pursue it is they do not know which employers in their service area are self-funded, who their carrier is, or who their broker is. That information is filed annually with DOL on Form 5500 and its schedules. Almost nobody outside benefits consulting reads it.
You type
“Who are the self-funded employers in my service area, who's their carrier, who's their broker, and who do I call?”
The agent reads three different Form 5500 schedules. Schedule A has the carrier and any premiums paid. Schedule C has service providers, including brokers paid directly by the plan. The main filing has the plan administrator, which is the fiduciary contact (with phone). The agent joins all three on plan ID, filters for funding flags that indicate self-funded, restricts to service area zip codes, and ranks by participant count.
Two real targets at the 100-life threshold within service area:
| Sponsor type | H&W lives | Carrier | Broker comp (Sch A) |
|---|---|---|---|
| Local credit union (SLO) | 326 | Anthem Blue Cross L&H | $244K |
| Regional petroleum company (Santa Maria) | 363 | Blue Cross of California PPO | $124K |
Two targets is not a flood. It is two more than most independent ortho groups are working today. The right framing is depth (carve-out partnership on these two) rather than breadth.
Two honest notes about the data. First, broker name is not always recoverable. Form 5500 Schedule C only names brokers paid directly by the plan. When the broker is paid via commission embedded in the carrier's premium (the standard arrangement), the broker's name lives on the carrier's filing under the carrier's EIN, not the sponsor's. The compensation total ($244K, $124K) proves the broker exists and is well-paid. The brokerage name comes via a phone call to HR.
Second, “self-funded” in Form 5500 is a noisy flag. Both sponsors above show the carrier with substantial premium activity and no stop-loss flag, which signals ASO or level-funded rather than pure self-insured. That changes the pitch. ASO employers carve out specific procedures (an ortho center of excellence) rather than replacing the network. The conversation is still real. The framing is “carve-out partner” not “replace the network.”
The synthesis chain
Four prompts answered the three questions. The synthesis chain that follows is the part no other vendor in this category does. Other tools sell you the data and stop. Three more prompts, in the same chat window, take the data and produce the deliverables.
The one move
“Given everything you just showed me across the four prompts, what's the one move I make this week that has the highest dollar impact? One sentence.”
Agent reads back over its own outputs, weights each finding by dollar impact, picks the top one. Returns one sentence.
Draft the negotiation memo
“Now draft the one-page Aetna negotiation memo for that move. Use the Medicare anchor from Prompt 1a, the hospital spread from Prompt 1b, and the 8-quarter walk-down from Prompt 2. Target rate, three justification points, walk-away. Leave [bracket] placeholders where you don't have my data.”
Agent reasons over the specific dollar figures from each prior result and writes prose that ties them together. Not a template. The dollar figures, percentage points, and hospital names came out of the conversation.
The 90-day calendar
“Now show me the 90-day calendar. Aetna meeting, employer outreach, monitoring loop. Order by dollar impact, not by ease. One line per item, owner, target date, expected outcome.”
Agent sequences the moves. Aetna conversation moves the most money but is the slowest (renewal cycle). Employer outreach is faster but smaller. Returns a calendar, ordered correctly.
The chain works because the agent retains conversation context. The fortieth prompt next Tuesday at 6 a.m., when the CEO is prepping for the actual Aetna call, is what they are paying for. The four prompts above are the demo.
The memo it produces
Closing prompt 2 produced an eight-page negotiation memo, structured as an evidence pack with executive summary, findings, proposal, leverage analysis, anticipated payer counter-arguments, and a week-by-week plan. The headers and key passages, anonymized:
Confidential — Internal Negotiation Memorandum
Aetna Contract Renegotiation: Evidence Pack and Negotiating Position
1. Executive Summary
Aetna has been silently reducing the group's professional fees on three high-value joint replacement codes for eight consecutive quarters, while every other commercial payer (Anthem Blue Cross, Blue Shield of California, UnitedHealthcare, Cigna) has held rates flat across the same period. The reductions are not random. They are linear, programmatic, and concentrated on the codes with the highest dollar value per case.
Across the 24 months examined, Aetna has moved the group from 122% of Medicare toward 113% of Medicare on total knee arthroplasty (CPT 27447), with similar trajectories on total hip and total shoulder. If the trend continues unchecked through 2026, the group will be paid at near-Medicare parity on its three most economically important surgical procedures.
Headline asks
- Stop the rate erosion. Restore CPT 27447, 27130, and 23472 to 2024Q1 levels at minimum, with a contractually fixed annual update mechanism going forward.
- Reset the ortho surgical fee schedule to 150% of Medicare. The midpoint of a defensible 130–180% commercial band for orthopedic surgery. Not aggressive, not generous, just market.
- Eliminate the silent step-down clause. Replace any conversion-factor or fee-schedule-vintage language with a transparent, Medicare-indexed update.
- Term: three years, with no further unilateral fee schedule changes mid-term.
Sections 2 through 8 follow with quantified findings, proposed rate structure, leverage analysis, anticipated counter-arguments with rebuttals, and a week-by-week negotiation plan. Methodology and per-surgeon Medicare benchmark in the appendices.
The rebuttal section
The strongest part of the memo is Section 6: anticipated Aetna arguments and the data-backed rebuttals. This is what the practice walks into the room with.
“Your rates are consistent with our regional fee schedule.”
Then explain the eight-quarter linear decline that no other payer in the practice's book matches. The data does not show consistency. It shows programmatic reduction.
“Medicare benchmarks aren't how we contract.”
They are how the rest of the commercial market contracts, including Aetna in other states and other groups. They are also how Aetna's own analytics teams benchmark internally. The Medicare anchor is not optional. It is the only common reference point.
“We can't move on rates this cycle.”
Then the practice will issue a 90-day termination notice and complete the cycle out of network. Interim status quo is acceptable only with written commitment to a true-up at restoration.
“What about facility costs?”
Aetna pays facilities in this service area $6,602 to $26,311 per knee replacement. The professional fee is approximately $1,250. Doubling the professional fee would raise Aetna's per-episode cost by less than 10%. Reducing the facility cost by 10% would save 5 to 20 times more. The professional fee is not the cost driver. It is the easiest line item to underpay.
Three minutes earlier the practice did not have that memo. They have it now. Their Aetna network rep has not seen a memo this tight from a 12-physician group.
What this changes for an independent group
The same three questions, today versus through Medistill:
Today
With Medistill
The unlock is not the cost line. It is iteration speed. A group that can re-benchmark every quarter, re-detect rate-walk-down on every payer, and refresh the direct-to-employer pipeline whenever a new Form 5500 cycle posts, stops being surprised. The Aetna walk-down in this case study had been running for two years before the practice saw it. The next walk-down, by the next payer, gets caught in the quarter it starts.
That is the difference between defending the rate you already have and walking into the next conversation with real numbers.
FAQ
Questions independent specialty CEOs ask before running their own claims file through the workflow.
How do independent specialty groups benchmark payer rates today?+
Most rely on a broker's market check (anchored to that broker's book), an AAOS or specialty-society compensation survey (national, lagging, not payer-specific), or the billing team's manual reconciliation against contracted rates. None of those answer the question of what the payer should be paying given the group's specific surgeons, locality, and case mix. A Medicare-anchored benchmark, computed per NPI from the CMS Part B Public Use File and overlaid with the published commercial band for the specialty, is the only defensible per-group answer.
What is the difference between a Medicare-anchored benchmark and a broker market survey?+
A Medicare anchor is computed from CMS-published per-provider allowed amounts. It is locality-adjusted, case-mix-adjusted to that surgeon's actual practice, and reproducible. A broker market survey is a sample of the contracts that broker has visibility into, which usually means employer-side benefits arrangements, not provider-side payer agreements. Brokers see the buy-side, not the sell-side. Their benchmark is also subject to selection bias toward their own book.
Can a practice analyze its own claims data through Medistill without sharing PHI?+
Yes. The PMS export is scrubbed locally on the practice's own laptop before any data leaves. The 18 Safe Harbor identifiers per 45 CFR 164.514(b) are stripped. The de-identified file is loaded into Claude Desktop on the same laptop. The Medistill connector returns market benchmarks, Medicare anchors, and payer behavior reference points. Claude does the variance math against the file locally. The file never leaves the practice.
What does Form 5500 actually disclose about a self-funded employer?+
Form 5500 main filing names the plan sponsor, plan administrator, and fiduciary contact (with phone and address), and the participant count by plan type (health and welfare versus pension). Schedule A names every insurance carrier the plan paid premiums to, plus broker compensation totals reported by the carrier. Schedule C names service providers paid directly by the plan, including brokers paid that way. Brokers paid via commission embedded in the carrier premium do not appear named on Schedule C, only as a compensation total on Schedule A. The carrier and the broker compensation level are usually enough to start the outreach conversation.
Is detecting rate-erosion against a Medicare benchmark the same as detecting actual-versus-contract variance?+
No. Actual-versus-contract requires the practice's signed fee schedule. The Medicare-anchor approach detects something stronger: programmatic rate-walk-down that the contract permits but the practice is not auditing. A linear, R-squared near 1.0 decline in percent-of-Medicare across eight quarters, on the high-dollar codes only, while every other commercial payer holds flat, is a contract clause being executed as written. The contract-versus-actual view catches the same exposure plus any single-claim underpayments. Both views are one prompt away once the de-identified file is in the room.
How long does the full negotiation prep workflow take?+
Twenty-eight minutes for the four prompts: the per-NPI Medicare benchmark, the local hospital facility-rate spread, the rate-erosion detection across eight quarters, and the self-funded-employer pipeline. The synthesis prompts that follow (one-sentence headline move, polished negotiation memo, ninety-day calendar) take another ninety seconds. Total time from open laptop to defensible deliverable: under thirty minutes.
Run your own file
The best test of any payer-side benchmarking tool is a twenty-eight-minute walk through your own four prompts. Pull the export, scrub it locally, drop the de-identified file in. Ask what the payer should be paying. Ask whether they actually are. Ask who the local self-funded employers are. See whether the memo that comes out the other side is one you would put in front of your network rep.
Free 50 free credits, full access. If the first prompt doesn't change how you walk into your next contract cycle, don't subscribe.