CMS released the final 2027 physician fee schedule on November 4, and the headline everyone grabbed was the pay cut. It got worse, not better. The conversion factor for most physicians drops 2.31% to $32.63, a deeper cut than the 1.68% CMS floated in the summer. If you are a qualifying participant in an advanced risk model you come off a little easier, down 1.83% to $32.95, which is the two-track payment story I wrote about last week and will not relitigate here. The cut is real and it stings. It is also not the most important thing in this rule.
The most important thing is buried in the remote monitoring section, and it is the clearest signal CMS has sent in years about where this is going. Starting in 2027 you need a face-to-face initiating visit before you can bill remote monitoring, and remote therapeutic monitoring gets limited to established patients. Fine. But the one that matters lands in 2029. From that date, you can only bill remote physiologic and remote therapeutic monitoring when the clinical staff delivering the service are your own employees. Not a third party's. Not a vendor's. Yours.
Read what CMS left untouched, because that is the tell. You can still buy the device from an outside company. The software, the connectivity, the data hosting, the cybersecurity, the technical support, all of it can still come from a vendor. The only thing Medicare is going to make you bring in house is the person doing the clinical work. CMS just drew a bright line between the box you ship and the care you deliver, and told you the box was never what they were paying for.
I have been saying a version of this for years, so let me be direct about it. The value of remote monitoring has never been the ability to mail someone a cellular blood pressure cuff. Remote physiologic monitoring payments jumped 31% in a single year, to $536 million in 2024, and a lot of that money flowed to businesses whose actual core competency is logistics. Ship the device, confirm it connects, generate the minimum number of billable days, drop the claim. That is not care. That is arbitrage on a billing code with a shipping department attached. It worked because Medicare paid for the inputs it could count instead of the outcome it actually wanted. This rule starts pricing that arbitrage toward zero, and it should.
You can see the same intent everywhere else in the rule if you look for it. CMS is giving clinicians until 2030 to move off traditional MIPS and onto value pathways or advanced risk models, and it built new pathways aimed squarely at conditions like diabetes and hypertension. It closed a loophole in the advanced risk bonus that it estimates will keep $2.38 billion from going out the door over the next decade. The through line is not subtle. Medicare is slowly rebuilding its payment logic around specific conditions and specific results, and backing away from paying for generic activity. The remote monitoring staffing rule is that same move pointed at a different part of the problem.
Here is where I part ways with the people applauding. I am glad CMS saw the problem clearly enough to act. But look at what they actually did. They proposed the own-staff requirement for 2027, took the comments, and then pushed it all the way to 2029. Commenters asked for 2028. CMS gave them an extra year on top of that. You do not delay the central fix by two years if you believe it is the central fix. The delay tells you CMS knows that owning the staff is not really the lever. It is a proxy. Employment status is an easy thing to write into a rule and a hard thing to argue with. What it is standing in for is the thing nobody can regulate cleanly, which is whether someone accountable actually reaches the patient and changes what happens to them. CMS reached for the proxy because the proxy is governable, and then bought two years because some part of the building knows the proxy is not the real thing.
So let me tell you what I think happens before 2029 even arrives. The automation around programs like ACCESS and APCM gets much broader, much faster than most people in this debate are planning for. The work of staying in contact with a patient between visits, the routine outreach, the first-pass triage, the documentation, the follow-up that falls through the cracks today, a large share of that stops being a pure headcount problem. Not because the humans disappear. Because the output per human climbs steeply when an agent handles the parts that never needed clinical judgment in the first place.
Picture what a mature ACCESS or APCM operation looks like in 2028. An agent runs the daily check on every patient whose readings came in, flags the ones drifting the wrong way, drafts the outreach, and hands your nurse a ranked list of the fifteen people who actually need a human today instead of the two hundred names she would have had to sort through by hand. The nurse spends her whole day on clinical judgment because the routine work cleared itself. One nurse now carries a panel that used to take three. So ask the 2029 question against that picture. Is the nurse's employer delivering the service, or is it the company that built the agent that did most of the reaching. You are forced to answer, and the honest answer is that it does not matter, because the only thing the patient experiences is whether someone caught the problem in time.
That is the question that survives all of this. Not who employs the team. Not whether a person or an agent sent the message or checked the reading. Whether the patient got better. Did the A1c come down. Did the blood pressure hold. Did the admission that used to happen simply not happen this time. The people I know who are actually moving outcomes are already agnostic about who, or what, performs each individual touch, as long as a human owns the accountability for that patient and the patient is measurably healthier at the end of it. The labor model is becoming a detail. The outcome is the whole thing.
Which is why I hope this rule is a step toward CMS getting out of the business of counting minutes and staff and devices altogether. Time-based codes made sense when you could not measure outcomes at scale. You can now. Every one of these fights, the sixteen days of data, the twenty minutes of clinical time, your employee and not theirs, exists because Medicare pays for inputs it can audit instead of the result it is actually trying to buy. Pay for the result and the fights evaporate. Nobody argues about whether a nurse or an agent made the call when the contract pays on whether the population got healthier. Tie the money to the outcome and the market sorts out the staffing, the automation, and the devices on its own, faster and better than any rule can.
So here is what I would do if you run one of these programs. If your remote monitoring operation is honestly a device-shipping business with a billing engine bolted onto it, you have until 2029 to become something else, and you should not take the full two years. Build the thing that actually moves the number. Accountable contact between visits, measured against outcomes, staffed by whatever mix of people and automation gets the patient better. We run our own care management at roughly one nurse to two hundred patients, with automation carrying more of the routine load every quarter, and when it works we see A1c fall 20 to 30% and systolic pressure come down and hold inside 90 days. Those are our numbers, from our panels, and you should pressure test them against your own claims data before you believe me or anyone else. The direction, though, is not in question.
And if you are a provider buying remote monitoring from a vendor, ask them the only thing worth asking. Show me your outcomes, not your shipment volume. If the best they can do is tell you how many devices went out the door and how many days of data came back, you already know what 2029 is going to do to them.
CMS spent this rule drawing a line between the box you ship and the care you deliver. That is the right line. Now they should finish the job and pay for the only side of it that was ever the point.
