Getting a surgeon’s income off the clock
Why “more volume” is not scale, which four rules actually decide it, and the one calculation that settles the whole question.
The problem, stated as an equation
A surgeon is paid for presence at a discrete event. A procedure performed. An encounter conducted. So income takes one shape:
Events are capped by hours — operating room time, clinic slots, and a body that ages. Rate is capped by a fee schedule that falls in real terms nearly every year.
Which means orthopaedic income is linear in personally-delivered hours, with a falling slope. Every conventional growth strategy — another clinic day, more advanced-practice providers, a second site — raises the number of events. None of them changes the exponent. You are still selling hours; you have just bought some from other people and kept the margin.
Ownership does change the exponent — surgery-centre equity, real estate, a device royalty. It works, and it is why so many surgeons pursue it. But it is a capital answer, not a clinical one, and some of its neighbours (physician-owned distributorships in particular) carry serious regulatory risk. It is not available to most people reading this.
What actually scales, precisely
Sort the things a surgeon does by how their cost grows with the number of patients.
Forming a fresh judgement on each patient — also O(n) in surgeon-hours. This is the one people mistake for scalable because it feels like thinking rather than labour. It is not.
Writing the rule once, and signing the exceptions — O(1) for the rule, plus O(exceptions). This is the only one that scales.
So the question “how does a surgeon earn while asleep” has exactly one non-capital answer: your judgement has to be encoded once and executed by others — people or software — with you accountable for the standard and for the cases that fall outside it.
That is not a technology opinion. It is the shape of the only structure that has ever let expertise cover a population.
Four rules decide whether you can actually do it
Remote therapeutic monitoring is the one place in the fee schedule where a clinician is paid per patient per month rather than per visit. Revenue accrues while time passes. That is literally income earned while asleep, and it exists today.
Four Medicare rules govern it, and most models we see are wrong before they start because they miss one.
You are receiving it. So the obvious plan — monitor my own post-operative patients and get paid monthly — is blocked for the entire global period. On a 90-day global, your monitoring revenue on your own surgical patient starts at day 91.
Combine rules 1 and 2 and the consequence is sharp: on your own surgical patients you are locked out for ninety days, and at day 91 the slot may already be occupied by somebody who was not.
General, not direct. The billing practitioner need not be present. This is the only line in the rules that lets one clinician’s accountability cover a population instead of a schedule.
The calculation that settles it
The management code is twenty minutes of interactive communication, per calendar month, per patient. And the published reality is that this line is captured in roughly 53% of monitored months — because it needs a real human interaction that often does not happen.
So take a 300-patient panel and do the arithmetic:
= 53 hours per month of interactive time
Fifty-three hours a month is impossible for a surgeon and entirely ordinary for delegated clinical staff. There is no version of this where the surgeon personally does the talking and the panel is large enough to matter.
1. There is now a ten-minute management code. CMS finalised 98979 (RTM) and 99470 (RPM) for 10–19 minutes in a calendar month, valued at 0.31 work RVUs — half of the twenty-minute codes. A panel run on the short code is roughly half the time burden.
2. The live interaction need not fill the whole time. CMS adopted the CPT language verbatim: these codes “require a live, interactive communication with the patient/caregiver. The interactive communication contributes to the total time, but it does not need to represent the entire cumulative reported time of the treatment management service.”
So the honest figure is not fifty-three hours of live conversation. It is fifty-three hours of billable management time, only part of which must be live — and less than that if the ten-minute code fits. The direction of the argument is unchanged: the constraint is still hours, and it still cannot be the surgeon’s hours. Its size was overstated.
So rule 4 is not a detail. It is the entire mechanism. General supervision is what converts a monitoring panel from a schedule-bound service into something that grows independently of your hours — and it is why the interesting engineering is in preparing and evidencing that interaction, not in the monitoring device.
The inversion
Now put rules 1, 2 and 3 together and the strategy inverts in a way most people find uncomfortable.
They have no global period blocking you. They need no established relationship. Nobody else is holding their monthly slot. And today they are treated as the low-value half of the clinic — an encounter that generated one evaluation-and-management code and nothing else.
That is the reinvention, and it is not a marketing reframe — it falls directly out of the rules. The conservatively-managed patient stops being the consolation prize for a clinic slot and becomes a recurring, monthly, staff-delivered, surgeon-attested relationship that consumes no operating-room time at all.
It also happens to be where the clinical harm accumulates. Delay to physical therapy in knee osteoarthritis is associated with a graded increase in opioid exposure — adjusted relative risks rising from about 1.25 to 2.50 as delay lengthens, with physical therapy itself at 0.77 (Kumar et al., BJSM 2023, n=67,245). Graded, not exponential; association, not causation. The population that is cheapest to neglect is the one where neglect costs the most.
Two clocks, two different economics
The result is that a practice runs two distinct models at once, and confusing them is fatal to the maths.
Day 91 onward, and every non-operative patient. Monitoring is revenue — monthly, per patient, staff-delivered under general supervision, surgeon-attested.
One instrument, two economic logics. The first protects a payment you already receive; the second creates one you currently do not.
The honest arithmetic, including what I will not tell you
The standard monthly musculoskeletal stack prices at about $105.55 per patient per month for CY2026. Every model you will be shown multiplies that by twelve.
Do not. The device-supply line and the management line have different capture rates, and the management line lands in roughly 53% of monitored months. A twelve-times model is out by something close to half.
Pick your own realised figure, hold it to your own capture rate, and note that the binding constraint is not the dollar amount at all — it is the fifty-three hours.
And then there is the other rail, which may be the better one
Everything above is remote therapeutic monitoring. There is a second rail on the same population, and it did not exist in usable form until this summer.
The ACCESS Model (Advancing Chronic Care with Effective, Scalable Solutions) is a CMS Innovation Center model under §1115A. It began 5 July 2026 and runs ten years. Four clinical tracks, and one of them is chronic musculoskeletal pain.
Which means no twenty-minute interactive-communication gate. The fifty-three hours above is an RTM constraint. It is not an ACCESS constraint.
And there is no global period to work around, because the model is not attached to a surgical episode at all.
The payer is populating the outcomes column itself. Publicly, risk-adjusted, for musculoskeletal care. That is the column a routing agent could not read — the reason it sorts on price is that price was the only field with data in it.
Entry also requires designating a physician Clinical Director accountable for clinical oversight and compliance. A named accountable clinician stops being a differentiator and becomes a condition of admission.
That is a much lower barrier than participation, and for most practices it is the first door, not the last one. It also means a surgeon can be paid inside ACCESS while somebody else carries the organisational burden.
So this is not additive to the monitoring model on this page. It is an alternative, chosen per patient. Any plan that runs both on the same beneficiary is wrong before it starts, and that is the single most expensive mistake available here.
Source: cms.gov/priorities/innovation/innovation-models/access, read 10 September 2026; that page was last modified 12 August 2026 and lists rolling start dates that fall after its own last-modified date, so ⚠ re-verify current application dates against CMS before acting on them. Participants must be Medicare Part B-enrolled organisations. Not billing or legal advice.
The threat, which is the same shape as the opportunity
If income comes from encoded judgement executed by others, the only question that matters is whose signature — and the billing practitioner for RTM does not have to be a surgeon. A physical therapist can hold that slot. So can a nurse practitioner. And they have more time than you do.
So “scale your judgement” and “commoditise your judgement” are the same sentence, read at different distances. The defensible position is not I can monitor too. It is narrower and much stronger:
Conservative management of a surgical candidate is a surgical judgement. Nobody else can hold it, and it is exactly the population rules 1 to 3 leave open.
The architecture already exists, and it is federally funded
None of this is speculative design. ARPA-H’s ADVOCATE programme — $62.7M over four years, building toward the first FDA-authorised clinical agentic AI (the goal; nothing is authorised today) — specifies exactly this decomposition. One performer describes its system as separating “conversational intelligence from clinical authority by a clinician-built rules system that validates every proposed action against approved protocols before execution.”
The clinician writes the rule once. The agent executes against it. The clinician holds accountability and takes the exceptions. A separate supervisory system watches the whole thing, cheap checks first and expensive ones only on what survives.
That is O(1) plus O(exceptions), specified in a federal contract. General supervision is the legal vehicle that already exists to run it in a practice. The longer argument about both agents is here.
What is still open, named rather than glossed
- Does AI-conducted interaction count toward the management minutes? Largely answered, and not the way I guessed. Commenters on the CY2026 rule asked CMS directly whether secure messaging, asynchronous chat, automated bi-directional messaging and AI prompts count. CMS adopted the CPT language and declined to specify further exclusions — “so long as they meet the CPT specifications.” The test is therefore whether a given exchange is a live, interactive communication with the patient or caregiver under the CPT manual, not whether a human typed it. ⚠ Note this sits awkwardly against the 2021 rule’s “real-time synchronous, two-way audio” language, so it is still a compliance question — but it is a much narrower one than “probably not.”
- The code-level split of $105.55. Unverified here, and load-bearing for any revenue model.
- Whether the slot is already taken in your market. Rule 2 makes this a land-grab. It is checkable and nobody checks it.
Disclosure. SurgeonValue builds physician-attestation and remote-monitoring infrastructure for orthopaedic practices. This argument favours the category we work in. Discount it accordingly, and check the rules yourself — they are public.
Not billing advice. Rules 1–4 quoted from telehealth.hhs.gov, “Billing for remote patient monitoring”, page last updated 17 January 2025, verified live 10 September 2026. Coding guidance changes; payment varies by locality and year. Confirm against the current Physician Fee Schedule and your own compliance review. Nothing here is risk-adjusted and nothing here ranks any provider.
The rail itself · Practice structure and ownership · The whole model · blaine@co-op.care