Buried in a 1,500-page proposed rule is a sentence that would end a venture-funded industry.

On July 14, 2026, CMS released its Calendar Year 2027 Physician Fee Schedule proposed rule, and for remote patient monitoring it is, in one vendor's assessment, the most consequential rulemaking since the RPM codes were created in 2019. The direction has reversed. Where recent years brought expansion, new device codes, a shorter management code, higher rates, this rule brings guardrails.

The central proposal: Medicare would allow payment for remote physiologic monitoring and remote therapeutic monitoring only when clinical staff employed by the billing practice furnish the services, not contracted third-party companies. Staff need not be on-site, but the employment relationship must run through the practice.

That single requirement is existential for a whole category of companies whose entire business is providing monitoring as an outsourced service to physician practices.

What RPM is and why it grew

Remote patient monitoring lets a practice track a patient's physiologic data, blood pressure, glucose, weight, oxygen saturation, from home, with clinical staff reviewing readings and intervening when something looks wrong. Medicare began paying for it in 2019, and the clinical logic is sound: catching a heart-failure patient's weight gain early is far cheaper than the hospitalization it prevents.

The economics created an industry. Most physician practices lack spare clinical staff to watch data streams from hundreds of patients, so companies emerged to do it for them, supplying devices, staffing the monitoring, and handling logistics, while the practice bills Medicare and shares revenue with the vendor. That outsourced model is what CMS now proposes to defund.

Alongside the third-party ban, the rule would require a separately reportable initiating visit at the onset of monitoring, restrict RTM to established patients, and recalculate payment. CMS is also soliciting comments on collapsing the current RPM billing codes into four new codes entirely, which signals more disruption ahead.

The fraud was real

It would be easy to write this as regulators strangling innovation, and that would be dishonest, because the problem CMS is responding to is documented.

The proposal follows OIG reports that flagged widespread fraud in the space. The abuse patterns in outsourced RPM have been well catalogued: vendors enrolling patients who never needed monitoring, billing for data review that amounted to an automated dashboard nobody read, boiler-room operations signing up beneficiaries by phone, and revenue-share arrangements that gave marketers a direct financial incentive to maximize enrollment rather than to identify patients who would benefit.

The structural vulnerability is obvious once you see it. When the entity generating the billing is a vendor paid per enrolled patient, and the physician whose name goes on the claim is several steps removed from the actual monitoring, the incentive to enroll aggressively is enormous and the clinical check is weak. CMS's requirements, an initiating visit, an established-patient relationship, staff employed by the biller, are each aimed at that gap: they force a real clinical relationship and real accountability into a chain where both had become optional.

But the remedy targets structure, not conduct

Here is the objection worth taking seriously, and it is the crux of the coming fight.

CMS is not proposing to punish vendors who defrauded Medicare, or to tighten documentation standards, or to audit the outliers. It is proposing to eliminate an entire delivery structure because that structure proved susceptible to abuse. Legitimate operators and fraudulent ones lose the same way.

And the practices that depend on outsourcing are not, on the whole, the sophisticated ones. Large health systems have the staff to run monitoring in house, which is why one vendor noted the rule largely validates practices that run their own programs. Small, rural, and independent practices are precisely the ones that cannot hire dedicated monitoring staff for a few dozen patients, and so rely on partners. The industry's warning is that providers will exit remote monitoring first in exactly the underserved markets CMS most wants to reach, and that reduced monitoring would drive the avoidable acute care that costs Medicare far more than the monitoring does.

That argument is self-interested, coming from companies whose revenue is at stake, and should be read with that in mind. It is also probably right about the distributional effect. An anti-fraud rule that works by requiring in-house staffing capacity is, functionally, a rule that favors large organizations, whatever its intent.

The squeeze underneath

The RPM changes land on top of broader payment pressure, which is what turns a difficult rule into a potentially decisive one.

CMS proposes conversion factors of $33.17 for qualifying alternative payment model participants, a 1.19% decrease from 2026, and $32.84 for non-qualifying providers, a 1.68% decrease. Physician payment is going down across the board, partly because the one-year 2026 conversion factor increase is expiring.

So a practice facing a general payment cut is simultaneously being told that a supplemental revenue stream now requires hiring staff it cannot afford. For a marginal independent practice, the rational response is not to build an in-house monitoring program. It is to stop offering monitoring.

The AI half of the rule

The same rule advances Medicare's slow attempt to figure out how to pay for clinical software, which is the larger and less-covered story.

CMS proposes two new technology-focused MIPS improvement activities and continues developing a framework for Software as a Medical Service, introduced at length weeks earlier in the outpatient rule.

That framework matters more than its dull name suggests. Medicare's payment architecture was built to compensate people for time and procedures. Clinical AI does not fit: an algorithm that reads an image or flags a deteriorating patient consumes almost no clinician time but may deliver substantial value. Until CMS decides whether such software is a billable service, a bundled practice expense, or nothing at all, the business case for clinical AI in Medicare remains unresolved. The agency is edging toward an answer, and the RPM crackdown suggests the mood in which it will be written: cautious, and skeptical of arrangements where a vendor sits between the patient and the biller.

Notably, the rule also includes a request for information about the CPT coding system and the AMA's role in it, asking whether CMS is adequately considering harms from the licensing costs and innovation barriers of a privately controlled code set. That is a striking question for the agency to pose about the vocabulary it uses to pay for everything, and it signals that the plumbing of medical billing is genuinely in play.

What happens next

Nothing is final. This is a proposed rule with a comment period, and the industry response is already organized: vendors are submitting cost-composition data CMS explicitly requested and documenting clinical workflows to rebut the suggestion that these services do not involve clinical staff. The telehealth trade group called many changes constructive while flagging that certain provisions may raise concerns or have unintended consequences. Proposed rules do get softened; this one may be.

The question CMS has to answer is genuinely hard, and worth stating plainly rather than resolving glibly. Outsourced RPM produced real fraud and real clinical value, often through the same companies. A rule permissive enough to preserve access will preserve some abuse; a rule strict enough to end the abuse will end some access. CMS has proposed erring hard toward the second, and the cost of that choice will fall disproportionately on small practices and the patients they serve.

Outsourced RPM produced real fraud and real clinical value, often through the same companies. This is not a necessary crackdown or a war on innovation. It is both, and the argument is about the ratio.

Further reading