Medicare PT rate -1.68%, but RVUs rise 1-3%

Medicare's proposed 2027 conversion factor for a non-qualifying practice is $32.8409, down 1.68% (Holland & Knight, July 2026), and the same proposed rule would raise relative value units (RVUs) on common physical therapy codes by roughly 1% to 3% (APTA, July 2026). If you bill Medicare, those two numbers pull in opposite directions, and the gap between them is small enough that one extra hour of recovered clinician time per week can outweigh it. That is the part of the 2027 fee schedule worth acting on before the final rule lands.

The two numbers and how they net out

CMS released the CY 2027 Physician Fee Schedule proposed rule on July 14, 2026 (AOTA, July 2026). The conversion factor, the dollar multiplier applied to every code's RVUs, falls 1.68% to $32.8409 for practices outside an advanced alternative payment model, which AOTA attributes mainly to the expiry of a temporary 2.5% increase that applied in 2026. Against that, the proposal changes how practice expense RVUs are calculated, moving from a specialty-wide indirect cost average to a service-specific method. Physical therapy has low indirect costs, so the shift favors the profession. APTA estimates an overall payment increase of 1% to 3% for physical therapy services (APTA, July 2026), and its later modelling projects a 35% to 45% rise in practice expense RVUs for the most commonly used codes over eight years, or 1% to 3% a year (APTA, September 2026).

Here is the arithmetic, and the caveat. If the 1% to 3% is an RVU change that still has to be multiplied by the lower conversion factor, the net effect on your Medicare fee schedule income runs from about -0.7% (1.01 x 0.9832) to +1.3% (1.03 x 0.9832). If APTA's figure is already net of the conversion factor cut, as its wording ("despite the conversion factor decrease") suggests, the range is +1% to +3%. We could not confirm which reading is correct from the sources available to us, and APTA is an advocacy body with its own interest in the framing. The Federal Register impact table and your own billing mix are the only reliable answer, so run your top ten CPT codes through the proposed rule before you budget.

On an illustrative $250,000 a year of Medicare fee schedule revenue, the spread across both readings is roughly a $1,750 loss to a $7,500 gain. APTA itself warns that "not all providers will expect a similar increase" (APTA, September 2026), so treat the upper end as a ceiling, not a forecast.

So what for you: the proposed rule moves your Medicare income by a few thousand dollars either way, which is not enough to build a 2027 plan around. Budget for flat, and treat any increase as a margin cushion rather than growth.

The line that matters more for AI: remote therapeutic monitoring

Buried in the same proposal is a change that bites harder than the conversion factor if you use software-led home programs. CMS proposes three new conditions of payment for remote therapeutic monitoring (RTM), effective January 1, 2027 if finalized (MedBridge, 2026). RTM could only be furnished to established patients. It would need a separately billable face-to-face visit, in person or by telehealth, where the billing practitioner discusses RTM and obtains consent. And payment would be made only when the monitoring is done by clinical staff employed by the billing practitioner or practice, which rules out third-party monitoring companies (MedBridge, 2026). Staff need not be on site, and general supervision still applies.

CMS also asks for comment on collapsing six RTM codes into two: one set-up and education code, and one monthly code requiring at least two days of data transmission, 20 minutes of treatment management, and one real-time interactive communication (MedBridge, 2026). The 10-minute management tier and separate device supply billing would go. Holland & Knight's summary reports the same three conditions (July 2026), and AOTA notes CMS would remove practice expense inputs for treatment management codes 98980 and 98981 (AOTA, July 2026).

The AI relevance is direct. Many exercise-adherence and movement-tracking platforms use automated flagging, and some bundle a vendor-employed team that reviews the data and contacts patients. Under the proposal, Medicare would not pay for monitoring performed by that vendor team. If your RTM billing depends on it today, your 2027 revenue from those codes depends on a final rule due, per MedBridge, in early November 2026. The comment period closed on September 14, 2026, so the text is now in CMS's hands.

So what for you: this week, find out who actually does the monitoring behind every RTM claim you submit. If the answer is a vendor's staff, ask the vendor in writing how they will restructure for January 1, and price a fallback in which your own employed staff do the 20 minutes.

What recovered time is worth, set against a fee schedule swing

A fee schedule change of a few thousand dollars is useful context for what time savings are worth. Take an illustrative case, with every input an assumption you should replace with your own: a three-clinician practice where AI-assisted documentation gives each clinician back two hours a week, and that time is redeployed into billable care. Over 46 working weeks, two hours a week is 92 hours per clinician. At 45 minutes a visit that is about 123 visits, and at an assumed $90 per visit it is around $11,000 per clinician, or $33,000 across the practice.

That is four to five times the top of the fee schedule range above, and it does not depend on a rule being finalized in your favour. The condition is real: it only works if you have waiting-list demand to fill the slots, and if the tool genuinely saves the time. We have shown elsewhere that most AI scribe vendors do not publish independent accuracy data (see the accuracy-testing gap in AI scribes), so measure your own before counting the saving. The same logic applies to claims work, where an hour a week lost to prior authorization paperwork is an hour not billed (what prior authorization costs a physio practice).

So what for you: the lever you control is clinician hours, not the rate. Pick one workflow, track minutes before and after for 30 days, and only then multiply by your own revenue per visit.

Compliance points that move with the rule

Two smaller items sit alongside the headline numbers. Holland & Knight lists the KX modifier threshold moving from $2,480 to $2,540 for 2027 (July 2026), so check whether your billing software or any AI claim-scrubbing tool has the new figure loaded. And if any vendor handles patient data on your behalf, including RTM platforms, a HIPAA business associate agreement should already be in place, whatever the payment rules end up saying.

The rule is still a proposal. Final figures can differ, and nothing here should be read as billing, legal or financial advice. Confirm every code-level number against the final rule and your Medicare Administrative Contractor before you change a fee or a contract.

So what for you: update the KX threshold in your systems when the final rule publishes, and keep a dated record of which tools touch billing data.

What the evidence points toward

Frankly, if you are weighing an AI purchase against a 2027 budget, the maths favors spending on time recovery and ignoring the fee schedule noise. The conversion factor and RVU changes are close to a wash on one reading and a modest gain on the other. The RTM proposal is the item with a hard January 1 date and a real chance of removing revenue you currently count on. Order your decisions accordingly: first the RTM contract check, then one measured time-saving trial, then the budget.

If you want a ranked view of which workflows in your practice are worth automating first, an AI Opportunity and Growth Assessment produces that list from your own numbers, or you can book a free 20-minute call to talk it through.

Related: What prior authorization costs a physio practice  ·  5,944 denials: Medicare's AI pilot exposed  ·  Billing automation for OT and SLT

Figures are from the CY 2027 Medicare Physician Fee Schedule proposed rule (released July 14, 2026; Federal Register July 16, 2026) as summarized by Holland & Knight (July 2026), AOTA (July 2026), APTA (July and September 2026) and MedBridge (2026). The net-effect range, the $250,000 revenue example and the $90 per visit time-value example are our own illustrative arithmetic, not CMS or APTA figures. Whether APTA's 1% to 3% estimate is before or after the conversion factor change could not be confirmed. The rule is proposed, not final. This article is for informational purposes only and does not constitute clinical, legal, billing or financial advice.