Pay up, productivity down: What MGMA's Andy Swanson sees for practices in 2026

For the first time in years, physician pay and productivity have split, with compensation rising even as encounter volume slipped, according to the latest MGMA provider compensation and productivity data report. The shift arrives alongside a new Medicare efficiency adjustment that, effective Jan. 1, trims the work RVU value of roughly 7,700 procedural codes, leaving practice leaders to rethink how they benchmark, schedule and recruit.

Physicians Practice spoke with Andy Swanson, chief customer success officer at the Medical Group Management Association, about what is driving the split, how the efficiency adjustment will land on different specialties, where burnout stands and the one number he thinks every practice leader should track this year. The conversation has been edited for length and clarity.

What is driving the 2026 split between physician pay and productivity?

Andy Swanson: I would say two things. First, the encounter level is slightly down, and our ratio of encounters to RVUs is actually staying flat. So while encounters are down, those patients that doctors are seeing actually have higher acuity and a little bit more complexity. What's going on is that we are moving lower acuity patients to our APPs and non-physician providers, offloading some of the lower burdensome caseload to the right level of care, and physicians are seeing the right level of care.

Compensation slightly went up this year, actually a little less than CPI would tell you, but still year over year, anywhere between one and a half and 3 percent for primary care, non-surgical specialty and surgical specialty care. So compensation continues to rise while our RVUs are going down slightly. Pay has to keep up with the times, so as the consumer price index goes up by 3 percent, we're only paying doctors about one and a half to two and a half percent more. Pay is going up disproportionate to our RVU production, but not as much as inflation.

So those are the drivers of the higher expense and quote unquote lower production. But I warn about the lower production, because while case volume might have slipped slightly in encounters, the complexity still remains high for physician providers.

Is rising physician pay with falling productivity sustainable?

Swanson: There are two long-term unsustainable factors here. The first is that, in theory, pay can't keep going up and up and up while production goes down. It's one year, so I want to be cautious about that.

The second thing causing pay to go up is the cost of running a practice, both in salary and expense tied to physician costs and other clinician costs, but also all the other costs associated with medical groups. So if cost continues to go up, including salaries, but reimbursement continues to stay flat or negative compared to consumer price index increases and inflation, then no, things are not sustainable. Lowering costs in physician groups is going to have to be imperative, which includes salaries. But there's going to be a breaking point on reimbursement, and as reimbursement decreases or goes flat, it's not sustainable. Those two things together are a recipe for long-term problems.

What do practice administrators need to know about the 2026 Medicare efficiency adjustment?

Swanson: The cut is two and a half percent on about 7,700 codes, disproportionately affecting procedural volume and not necessarily time-based procedures. For primary care it's largely safe, relatively unchanged. For non-surgical providers and surgical providers, it's going to be a big deal.

All benchmarks in this coming calendar year are going to have to start with the understanding of procedural volume mix, and where that two and a half percent cut is going to be felt in full extent and where it's going to be felt in a nominal extent. Almost all non-surgical providers, like dermatologists and oncologists, and surgical providers, like orthopedic, neurological and pediatric surgeons, should walk in thinking, I'm going to see somewhere between a one and a half and two and a half percent cut in RVUs, not tied to the provider's actual performance in encounters, but just on paper. That's what's going to happen to those codes.

So how do we combat that? That becomes the next logical question, and the answer is we have to watch our encounters and RVU mix. We make sure the patients we're seeing, if possible, increase that number slightly per provider, and make sure the high acuity patients find their way to those specialists and they're not being seen by APPs. We do a good job of the bifurcation of caseload mix, making sure the procedural volume is going where it needs to go and the visit volume is going where it needs to go, locking down those schedules based on the acuity of anticipated patients, and maximizing our investment in the physicians we have. That's how we combat this arbitrary decrease in reimbursement.

How should practices prepare for the efficiency adjustment in hard-to-recruit specialties?

Swanson: There are two things. First, if you're searching for a physician, do not think you can arbitrarily cut your starting salary by two and a half percent and assume every physician is going to be happy to receive slightly less pay than they would have gotten in 2025. MGMA data would suggest that the significant march up in starting salary that began in COVID has now plateaued or peaked. While we see some starting salary coming down slightly, we are not anticipating significant decreases. We've reached the peak, and now we're probably at a plateau. We have to maintain where we are with compensation, because if you're recruiting one of those hard-to-recruit specialties, decreasing that starting salary is probably not going to land you the top quality candidate.

So then it becomes exactly what I suggested before. It's schedule management, watching the patient load on those new providers, making sure they're seeing a few more patients and not disproportionately using APPs while we get those physicians ramped up, and getting the most out of our investment in them. As people staff their NPs or PAs alongside these doctors, maybe the ratio of starting physicians is a little bit lower, and we're using that specialist's hands on more patients out of the gate, then thinking about ramping up ancillary support downline, not in the first or second year. But again, it's watching encounter mix and CPT code mix in the patients those physicians are seeing, making sure we're getting not just production, not just a number of patient encounters, but the relative acuity the provider is seeing from the patients going into those encounters.

Is physician burnout now a baseline condition?

Swanson: To a certain degree I'd say yes, but to a large degree I'd say no. We've reached a new baseline, where coming out of the pandemic we were reaching peaks that were untenable and unsustainable. A lot of groups have done a good job of addressing core burnout issues, and the production conversation we're having right now is corollary to this. So yes, we've reached maybe a new baseline where one in three doctors are suggesting they're leaving practice because of burnout-related issues. That too is unsustainable, especially with the physician crunch we're in today.

The groups that are doing a good job of schedule management, decreasing untenable production levels on their physicians, along with the longstanding issues of technology utilization and making sure administrative burden does not reach the bedside as much as it has in years past, those are the good fights that groups are still putting up on behalf of their physicians. Physician leaders are making wise investments in technology that are really going to decrease the things that can be most burdensome and most burnout-driving.

There's certainly more room to decrease physician burnout. We're not there yet, and I don't think the industry at large is suggesting we've reached a steady state and this is just going to be the way it is. The gains felt over the past few years are a little bit harder to see today, and perhaps with the advent of AI we're going to get there, but the returns just aren't there yet.

Why has the productivity payoff from AI been so uneven?

Swanson: The places where the gains have been felt the most are largely in the AI being used at the bedside to capture notes and documentation, things that have been cumbersome between patients and providers. The good news is we've pointed AI to help the provider where it matters most, with the patient relationship. There isn't a physician using an AI scribe who doesn't love the technology and isn't a big fan of the burden it has reduced.

What we are missing is where the next few unlocks in efficiency gains and operational effectiveness can be augmented with AI. I would caution everybody, we have short attention spans. It's the middle of 2026. A year ago, half of the people using AI scribes were not using AI scribes. Our industry is slow to adopt technology, especially outside of clinical advancements, so this has actually gone pretty well and pretty quick. I'd urge people to have a little patience. Over the next 12 months we've made some pretty significant gains already, and I think there's going to be another set of significant gains coming.

To think we all adopted online utilization back in the late '90s and early 2000s at the same rate is a flawed notion, and to think we gained all of those gains in the course of 18 to 24 months is also flawed thinking. We just have to give it some time. The technology is there. Our use cases are still developing, and as we explore, test and get some pilot projects going, I think we're going to see some of that manifest into true return. We're just not quite there yet.

What number should practice leaders track beyond work RVUs in 2026?

Swanson: It's funny, I've been at MGMA 11 years, and it's been a minute since we started talking about total visit volume. It sounds weird, because that's pre-RVU stuff, but as we moved all in on RVUs and production per provider, I think we lost sight of the total populations we're serving and how we're serving them, encounters, patient visits, whatever your organization is thinking about. That number becomes really important when you look at the division of those encounters across the different clinicians who work in your group, whether it's a physician, an APP, a care navigator or whoever that happens to be.

It's almost a little back to the future, thinking about what was good a long time ago may be a better way to start setting baselines now. Then we dive into, on a per encounter basis, are we getting the most out of our providers on those encounters? So I'd suggest we go upstream a little bit, think about our total patient panel size, and then think about how we're serving that panel size by each provider. That might get us to a new normal on what we're expecting our providers at different levels of acuity to handle for patient volumes, and what that means for benchmarks and pay downline.

I do think it merits an examination of compensation models again. What was old may be new again. We'll be starting to look at base salary models blended with production and quality metrics. Reexamining that base and what it looks like on a per encounter and RVU basis starts to make more sense now, especially as APPs come in to help us manage all the volume we're experiencing. We're in a physician shortage, and maybe times have changed, and now APPs become the central point of contact, and we're using physicians for the highest acuity patients and care. Maybe we've finally achieved what we've been talking about for a long time, which is the right level of care for what the patients need, and it may not always be physicians.

What should medical groups watch for in 2026?

Swanson: Stay tuned for 2026. As I mentioned with the RVU adjustment, baselines and metrics for 2026 are going to be a little wonky. As we launch our 2026 data set, I'm already looking forward to 2027. It's always going to be interesting to watch the baselines and the metrics move across the year.