Where the puck is going

…and only to allow payment for RPM or RTM services when performed by clinical staff employed by the practice and not when those services are delivered by contractors.

Since 2020, CMS has allowed leased and contracted clinical staff to furnish remote physiologic monitoring. Nearly every RPM program in the country is built on that permission. Remove it and the volume goes with it.

Four changes. Any one of them hurts.

All effective January 1, 2027 if finalized. Comments close September 14, 2026.

CMS is also asking whether to collapse all 17 RPM and RTM codes into 4 bundled G-codes, where you bill nothing unless every component is delivered every calendar month.

You get the thing you paid for.

Forty-three percent of enrollees not receiving treatment management is not a fraud statistic. It is a design statistic. When the device line and the management line are billed separately, a rational operator optimizes the one that scales and lets the other go. Nobody has to break a rule. The rule paid for shipping a box, so boxes got shipped.

The 43% figure is OIG's. The improvement counts are illustrative, chosen to show the mechanism, and are not drawn from published outcome data.

Read past the RPM section and CMS gets loud.

The same proposed rule that squeezes remote monitoring pays a premium for accountability. G2211 becomes a modifier worth 16% on the associated E/M code. A second modifier, available only inside a Shared Savings Program or LEAD Model ACO, is worth 32%. Same visit. Same patient. Double the premium. The only difference is whether you carry total cost of care.

Both conversion factors fall. The dominant term is the expiring one-year 2.5% increase from the Working Families Tax Cut, partly offset by the statutory updates and a budget-neutrality adjustment for work RVU changes. The gap between them is the message: a +0.75% statutory update if you are in an advanced APM, +0.25% if you are not.

You receive a recurring payment for managing a patient, and the full payment only if enough of your patients cross a CMS-defined improvement threshold. The model runs July 5, 2026 through June 30, 2036.

In the musculoskeletal track the targets are specific and they are all patient-reported. A site-appropriate instrument has to move: PROMIS PF up two points and PROMIS PI down two on the T-score, or eight points on the Oswestry, or ten on KOOS JR or HOOS JR. Pain intensity is required for every beneficiary and may not rise more than two points. A global impression of change is required at the end. Partial attainment does not count. And unlike the other tracks, MSK has no control targets at all, only minimum improvement, because the track is built to reward getting someone better rather than keeping them where they are.

Two things fall out of those numbers. The MSK amount is small, which is the point: this is deflationary by construction, not a new margin pool. And the withhold is organization-level, so one disengaged patient does not sink you, while a program that never moves function does.

There is a detail in the payment guidance worth noticing if you sell remote monitoring. RTM patient education and device setup, 98975, appears on the list of substitute services that can reduce an ACCESS payment when another provider initiates it during the care period. CMS is not only paying differently. It is naming activity billing as leakage.

Notice what ACCESS does not do. It does not name a device. It does not tell you to staff a monitoring center, or not to. It does not care whether your nurse is a W-2 or a 1099, because it is not paying you to have a nurse. Under that design nobody needs to write a rule banning contractors: a contractor who does not move the outcome does not get paid, and one who does is worth every dollar.

RTM is a good model. That is the uncomfortable part.

I want to separate two things the proposed rule runs together: whether remote therapeutic monitoring is clinically worth doing, and whether the way it is currently staffed invites abuse. My answer is yes to both.

Therapist-led RTM is one of the better ideas in the code set. A physical therapist who already owns the plan of care, watching adherence and function between visits and adjusting, is doing real clinical work. That is not a call center reading a dashboard. It is the therapist extending their own treatment into the twenty-nine days a month the patient is not in the clinic, which is exactly where musculoskeletal recovery is won or lost.

And the implementation problem is real. Standing up monitoring means devices, transmission, integration, escalation pathways, and someone actually watching. Most practices cannot build that, and telling them to hire their own staff does not conjure the staff. Strip out vendor clinical support and a lot of legitimate programs stop, including good ones. The proposal treats a staffing arrangement as though it were the fraud, when the fraud was billing for management nobody performed.

Personally I am a fan of a simple co-management structure. Name the accountable clinician. Let them delegate the watching to people who are good at watching. Keep the relationship, the plan, and the responsibility with the person who owns the outcome. That is a cleaner fix than an employment test, and it is closer to how surgical care already works.

If there is a path that preserves the staffing model while meaningfully tying reimbursement to outcomes instead of activity, I think that path is viable, and I think it is better than either extreme on the table.

That is a comment worth writing, and it is more useful than defending the status quo. The status quo produced a 43% completion rate. Asking CMS to keep paying for it is not a serious position. Asking CMS to let a therapist keep their vendor support, on the condition that payment follows a PROM rather than a transmitted-days log, is.