Medicare's 2027 Remote Patient Monitoring (RPM) Proposal: Employed Staff Only, an Initiating Visit and Lower Payments

CMS wants remote patient monitoring to be care management run by a doctor’s own practice, not a service bought from a vendor. The CY 2027 Medicare Physician Fee Schedule proposed rule, issued on 14 July 2026, would pay for remote patient monitoring (RPM) and remote therapeutic monitoring (RTM) only when the clinical staff doing the work are employed by the billing practice. It would also require a separately billable initiating visit, limit RTM to established patients, revalue the codes downward, and asks whether to collapse 17 codes into four [1][2].

A year ago the direction was the opposite. The 2026 fee schedule made RPM easier to bill, adding codes for as few as two days of data and ten minutes of management time [3]. The 2027 proposal answers years of findings from the HHS Office of Inspector General that much of what Medicare paid for was not real monitoring [4][5].

The rule is not final. Comments closed on 14 September and the proposed effective date is 1 January 2027 [2]. This post covers each proposal, the evidence behind it, who it affects, and what practices and vendors should do while they wait.

What exactly is proposed?

1. Employed staff only. This is the proposal with the biggest commercial impact. CMS says outsourcing “can fragment care, lead to insufficient involvement and oversight of the billing practitioner” [6]. Staff would not need to be in the same building, but they would need to be the practice’s employees. Buying devices, software and automation from vendors would remain allowed [8].

2. An initiating visit. Monitoring would have to start from a separately billable visit in which the clinician discusses the monitoring, decides it is appropriate and records consent. A code that does not involve a face-to-face encounter, or is not separately payable under Medicare, would not count [2][6].

3. RTM for established patients only. RPM already requires an established relationship. CMS proposes the same for RTM [1].

4. Lower values. CMS says devices “may be available at a reduced cost compared to our initial estimates” and proposes revaluing the services [1]. Device and setup codes would carry less practice expense, and treatment management codes would lose practice expense entirely, leaving them valued on physician work [7]. The fact sheet gives no per-code figures, so the size of the cut depends on the final rule.

5. Four G-codes, maybe. CMS did not formally propose them, but it asked for comment on replacing the 17 existing RPM and RTM CPT codes with four HCPCS codes [1][6]:

Why is CMS doing this?

Because remote monitoring grew faster than anyone could check it.

• 2024 report. OIG found Medicare RPM enrollees grew from about 55,000 in 2019 to about 570,000 in 2022, and payments from $15 million to more than $300 million. About 43% of enrollees did not receive all three components of the service: education and setup, device supply, and treatment management [4][6].
• 2025 report. RPM payments exceeded $500 million in 2024. OIG flagged month-over-month enrollment spikes of more than 150% at some practices and 34 practices billing overlapping claims for the same beneficiary [5][6].
• Fraud warnings. OIG issued a consumer alert in 2023 about cold calls and enrollment in monitoring that patients had not discussed with their own clinician [8].

CMS’s position is that some OIG recommendations “cannot be fully resolved with the current coding structure” [1]. The employed staff and initiating visit requirements target the pattern OIG described: monitoring sold to patients by third parties, with little involvement from the clinician whose name is on the claim.

Which codes are in use today?

The 2026 fee schedule is the baseline the proposal would change [3]:

Published figures for 99445 and 99454 vary between about $47 and $52 depending on source and locality [3][9]. The important point is that the two-day threshold introduced for 2026 survives in the G-code idea: GRPM2 would require only two days of data. If CMS goes ahead, intermittent monitoring stays billable, but only inside a managed, practice-run service.

We covered how the 2026 codes changed what counts as clinically useful wearable data in wearable data as a regulated input.

Who is affected?

• Practices running RPM themselves with employed nurses or medical assistants: mostly paperwork. Add the initiating visit to the workflow and expect lower payment per patient.
• Practices that bought RPM as a turnkey service, where a vendor’s staff call patients and log the minutes: the most exposed. If the proposal is finalized, that time stops being billable on 1 January 2027 unless the staff become practice employees [2][8].
• RPM vendors that supply clinical staff: need a new model. Goodwin’s summary notes the proposal leaves room for vendors that sell software and automation supporting a practice’s own staff [8].
• Device makers: still needed, at a lower implied device value [1][7].
• Rural practices: CMS itself asked how the changes would affect access where practices cannot replace outsourced staff with their own [8].

Where do wearables fit?

Less than the name suggests. Medicare requires an RPM device to meet the FDA’s definition of a medical device and to collect and transmit physiologic data automatically, not from readings the patient types in [10]. CMS does not itself require FDA clearance, though some payers do. In practice most programs use connected blood pressure cuffs, scales, glucose meters and pulse oximeters, not general consumer wearables.

Consumer wearables matter around the edges: for context between readings, for spotting who has stopped engaging, and for the growing number of patients who already wear one. We looked at how far clinicians actually use that data in do doctors use wearable data? The short answer is that most physicians have looked at it and almost none have it in their workflow, which the 2027 proposal does nothing to change.

What should practices and vendors do now?

1. Map who performs every billable minute. If any are contractors, decide now whether to employ them, move the work in house or stop billing it from 1 January 2027 [2][8].
2. Design the initiating visit into enrollment: a separately billable encounter with documented discussion, clinical determination and consent [2][6].
3. Re-forecast revenue per patient on the assumption that device, setup and management codes will be worth less [1][7].
4. Check RTM patient lists for patients without an established relationship [1].
5. Plan for a two-code world in case the G-code idea is adopted later: one setup code and one monthly code with a two-day, 20-minute, one-interaction minimum [1][6].
6. Watch for the final rule, which can change any of this, and re-read it rather than a summary before acting [2].

The short version

The CY 2027 proposed rule would pay for RPM and RTM only when performed by a practice’s own employed clinical staff, require a separately billable initiating visit, limit RTM to established patients and revalue the codes downward, and it floats replacing 17 codes with four [1][2][6][7]. It responds to OIG findings that RPM payments grew from $15 million in 2019 to more than $500 million in 2024 while 43% of enrollees missed required components [4][5]. Nothing is final until CMS publishes the final rule; the proposed effective date is 1 January 2027 [2].