Medicine began at the bedside, often the patient’s own. The doctor made house calls, knew the family, and remembered who had been sick last winter and whose father had the same problem. Care felt personal because the doctor had the time and the relationship to make it so.
That isn’t the medicine most surgeons practice today. And it isn’t because surgeons care less.
The math has changed
Physician payments for joint replacement have been declining in real terms for decades. Between 2000 and 2024, inflation-adjusted Medicare surgeon reimbursement fell 56.9% for primary total knee arthroplasty and 54.3% for primary total hip arthroplasty, the steepest declines among the high-volume inpatient procedures studied (Palmer et al., 2025).
The pressure continues. Under the proposed 2027 Medicare Physician Fee Schedule, the work RVU for a total knee replacement would fall from 19.11 to 15.94, a 16.6% reduction. For total hip replacement, it would fall from 19.11 to 15.37, a 19.6% reduction (CMS, 2026). Work RVUs are only one component of the total professional payment, but the direction is hard to miss.
Surgeons tend to respond by just doing more. Over my 25 years in medicine, I’ve watched that play out. I know surgeons who see 50 or more patients in a clinic day, perform a dozen joint replacements in a single OR day, or complete 20 or more sports medicine or hand procedures. Most of us would gladly see fewer patients and have more time with each one. But for many practices, that model no longer pays the bills.
At the same time, much of a physician’s week is spent away from the patient. In 2024, physicians reported working an average of 57.8 hours per week. Only 27.2 hours were spent in direct patient care; another 13 hours were spent on indirect patient care, such as documentation, orders, test results, and referrals, and 7.3 hours were spent on administrative work (AMA, 2025a). Prior authorization alone adds a substantial burden. In a separate AMA survey, practices completed an average of 39 prior authorizations per physician each week, consuming about 13 hours of physician and staff time (AMA, 2025b).
Add claim denials, payer audits, quality reporting, bundled payment programs, inbox messages, leadership responsibilities, research, and teaching. Each may be reasonable on its own, but together they compete for the same finite resource: time.
Patients arrive differently, too. They arrive with more medical information than they once did, and increasingly some of it comes from generative AI. In a 2024 survey, 21.5% of respondents reported using ChatGPT for health information. Among those users, more than two-thirds subsequently turned to a physician, suggesting that AI-generated information often becomes part of the clinical conversation rather than replacing it (Ayo-Ajibola et al., 2024).
Some of that information is helpful. Some isn’t. And almost none of it knows how a particular surgeon actually practices. So part of the modern visit is spent not only educating patients, but sorting through what they’ve already learned and translating generic information into a specific plan.
What hasn’t changed
Patients still want to feel cared for. They want a surgeon who understands what matters to them, answers their questions, and is available when something doesn’t go as expected.
For many surgeons, that relationship is still one of the best parts of the job. It’s also one of the first things to get squeezed when schedules tighten.
I felt that tension in my own practice. There were days when I knew exactly what conversation a patient needed, but I was also aware of a full clinic outside the door, messages waiting in the inbox, and other demands competing for the same few minutes. That constant pressure creates anxiety and takes some of the satisfaction out of the work. The problem was never that I wanted to give patients less. There was simply more work competing for the time I had to give them.
We’re not going back to house calls and ten-patient clinic days—at least not within the traditional payment model. Some physicians are responding by opting out of Medicare, charging cash for certain services, or moving toward concierge and direct-pay practices that trade volume for more time with patients. Those models may work for some physicians and patients, but they aren’t a solution for most surgical practices or health systems. The more useful question is how to preserve the parts of care that a person needs most when there’s less time available for everything else.
We can’t solve every problem by adding another person
Healthcare has traditionally addressed growing complexity by adding staff or shifting work from one team member to another. Often, that’s exactly the right answer. Nurses, APPs, medical assistants, schedulers, and administrative staff are essential to modern surgical care.
But there are limits. Labor is expensive, good people are hard to recruit and retain, and adding another employee to an inefficient process doesn’t make the process efficient. At some point, practices have to look at the work itself and ask which parts truly require human time.
Some tasks clearly do. Making a diagnosis, weighing treatment options, counseling a worried patient, recognizing that a recovery isn’t following the expected course—those require judgment, experience, and often a relationship.
Other work is much more predictable. Scheduling reminders, routine preoperative checklists, collecting standard information, documenting a straightforward encounter, checking whether expected milestones have been met, or delivering information that every patient needs can increasingly be standardized or automated.
The goal isn’t to replace people. It’s to use their time more deliberately.
Where technology can help now
This is no longer theoretical. Several technologies are already changing where practices spend their time.
Documentation. Ambient AI scribes are one of the clearest examples. In a randomized trial involving 238 outpatient physicians across 14 specialties, one of two systems reduced time spent writing notes by 9.5%, while the other did not produce a statistically significant reduction. Both showed improvements in some measures of physician workload and well-being, but clinically significant inaccuracies still occurred occasionally, so physician review remains necessary (Lukac et al., 2025). The value isn’t eliminating the physician from documentation; it’s reducing how much physician attention documentation consumes.
Administrative workflow. Appointment reminders, preoperative checklists, referral tracking, routine scheduling communication, and other predictable processes don’t necessarily require a staff member to initiate every step manually. Automating some of those steps can reduce the steady stream of small tasks and interruptions that consume a workday.
Routine follow-up. Much of postoperative follow-up is designed to determine whether recovery is proceeding normally. Digital check-ins can collect routine information and flag patients whose responses warrant attention. This lets the team spend more time with patients who need help rather than contacting everyone to identify the few who do.
The revenue cycle. Missing or inaccurate claim information and authorization problems remain important contributors to claim denials (Experian Health, 2025). Automated eligibility checks, authorization workflows, and pre-submission claim review can catch some of those problems earlier, before they become another task for someone to fix later.
Patient preparation. Much of what patients need to know before and after surgery is predictable: what the procedure involves, how to prepare, what recovery generally looks like, and when to call for help. Delivering that information consistently before the visit doesn’t replace the surgeon-patient conversation. It can make that conversation better by leaving more time for the patient’s individual concerns, expectations, and decisions. I looked at what it costs to repeat this information, one patient at a time, in The Hidden Cost of Repeating the Basics.
None of these tools is useful simply because it’s digital or uses AI. Poor technology can create as much work as it removes. The test should be straightforward: does it reliably take work away from the people using it while preserving—or improving—the patient’s experience?
Protect the time that matters
The scarce resource in a surgical practice is human attention.
A nurse can spend ten minutes making a routine call to a patient who is doing exactly as expected, or spend those ten minutes helping the patient whose recovery is going poorly. A surgeon can spend part of a short visit reviewing information that could have been learned beforehand, or use that time to understand the patient’s goals and whether surgery is actually the right decision.
I don’t think patients necessarily need more minutes with every member of the team. They need the right person at the right moments, with enough time and attention to matter. The challenge is designing care so those human minutes are spent where they matter most.
Less time doesn’t have to mean less care
The pressures on surgical practice are unlikely to reverse. Reimbursement will remain constrained, administrative requirements will continue to evolve, and patients will continue to expect more information and responsiveness.
We can’t meet every new demand simply by asking clinicians to work faster or by hiring another person. We have to become more deliberate about where people’s time goes.
Use technology for the predictable work. Use people for the work that requires judgment, expertise, and connection. And protect the moments that patients actually remember: the difficult decision, the reassurance before surgery, the unexpected problem, and the conversation when someone needs more than an algorithm can give.
Selected references
American Medical Association. (2025a). Doctors work fewer hours, but the EHR still follows them home: 2024 national physician comparison data.
American Medical Association. (2025b). 2024 AMA Prior Authorization Physician Survey.
Ayo-Ajibola, O., Davis, R. J., Lin, M. E., Riddell, J., & Kravitz, R. L. (2024). Characterizing the adoption and experiences of users of artificial intelligence-generated health information in the United States: Cross-sectional questionnaire study. Journal of Medical Internet Research, 26, e55138. https://doi.org/10.2196/55138
Centers for Medicare & Medicaid Services. (2026). Calendar Year 2027 Medicare Physician Fee Schedule Proposed Rule.
Experian Health. (2025). State of Claims 2025.
Lukac, P. J., Turner, W., Vangala, S., et al. (2025). Ambient AI scribes in clinical practice: A randomized trial. NEJM AI, 2(12). https://doi.org/10.1056/AIoa2501000
Palmer, R. C., Elmenawi, K. A., Hannon, C. P., Lieberman, J. R., & Heckmann, N. D. (2025). Medicare reimbursement for primary hip and knee arthroplasty is disproportionately decreasing relative to other high-volume inpatient procedures: Leader of the pack. Journal of Arthroplasty, 40(9 Suppl 1), S105–S109. https://doi.org/10.1016/j.arth.2025.04.088
