Dr. Michael Suk, CEO of physician advocacy platform MDEnvoy, can discuss the link between physician burnout and turnover, and what this could mean for struggling practices amid growing doctor shortages.
The doctor shortage gets the headlines. Burnout is what is actually driving it.
Go deeper with GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
The Association of American Medical Colleges projects a shortfall of up to 86,000 physicians by 2036. The Health Resources and Services Administration projects the gap will exceed 187,000 by 2037. Both forecasts assume that physicians will continue working roughly as they do today. Neither holds if older physicians retire earlier than expected, and survey after survey says they will. More than two in five active US physicians are already 65 or within a decade of it. Around 40% reported weekly burnout before the pandemic. The number has not improved since.
Hospitals have responded as most institutions do to uncomfortable workforce data. They have stood up wellness programs, resilience training, peer support apps, and chief wellness officers. Some of that work is useful. None of it is a strategy.
This is now well documented. The largest study to date, William Fleming’s 2024 analysis of 46,336 workers across 233 organizations published in the Industrial Relations Journal, found no evidence that individual-level wellness interventions, mindfulness, resilience training, stress management classes, and wellbeing apps produced any measurable improvement in employee mental health. The author concluded that organisations must change the workplace, not the worker. A Cornell study published this March put a sharper word on the rest of the apparatus. Researchers in the College of Arts and Sciences described what they called corporate bullshit, a category of organizational communication that sounds important but means nothing, and showed that workplaces structurally reward it. In healthcare, the parallel is what some have started calling ‘carewashing’. Pizza parties for nurses working unsafe ratios. Mindfulness apps for physicians processing prior authorisations at 11pm. Recognition is presented as a substitute for the structural change that recognition cannot deliver.
The newest example is artificial intelligence. Ambient AI scribes were sold as the answer to the documentation burden, and hospitals quickly moved to them. Then, in April, STAT reported the largest study to date, 1,800 clinicians across five academic medical centers over two years. The scribes saved physicians sixteen minutes of documentation time per eight-hour shift. Sixteen minutes. The number is nothing, but it is not a solution either, and it certainly does not justify the implementation costs being absorbed across health systems on the assumption that the problem was about to be automated away. A separate rapid review concluded that ambient scribes reduce self-reported documentation time, lengthen notes, leave standardized burnout scores unchanged, and do not affect productivity. The pattern is familiar. The technology is real. The structural problem is not the technology.
Burnout is not a personal deficiency in a workforce that has historically tolerated extraordinary demands. It is what happens when the operating model steadily moves administrative load, financial risk, and decision authority away from the people who carry the clinical consequences. Treat the symptom, and you keep the disease.
Look at how the work has actually changed. A practicing physician today spends roughly two hours on documentation and inbox management for every hour of direct patient care. The electronic health record, sold as efficiency, became a billing instrument. Prior authorization, designed to reduce waste, became a parallel clinical workflow that physicians staff without compensation. Quality reporting, intended to improve care, became an industry in its own right. None of this came from clinical need. Each layer was added by parties who benefit from it and do not absorb its cost.
The hospital executive sits at the center of this. Not as the villain in the story, but as the only actor positioned to rewrite it. Payers will not. Vendors will not. Federal regulators will, eventually, but slowly. Hospital leadership controls the levers that determine whether a physician spends the next ten years practicing medicine or processing it.
Start with the administrative load. Every hospital can measure, today, how many hours its physicians spend on tasks that do not require a medical degree. Most do not measure it. The ones that do are usually surprised by what they find. The fix is not another portal or another committee. It is a deliberate redesign that pulls non-clinical work off physicians and onto staff, or through software or scope-of-practice changes that match what the work actually requires. Systems that have done this seriously, usually led by physician executives with real operational authority, see retention improve before wellness scores do. That order matters.
Then there is the question of decision authority. Physicians who can shape the conditions of their work tolerate higher workloads than those who cannot. This is not abstract. When a hospital removes physicians from staffing decisions, scheduling design, technology selection, and care pathway development, it produces a workforce that sees its own job as something that happens to it. That experience is what burnout surveys are measuring. Restoring authority does not mean returning to private practice autonomy. It means designing governance in which clinicians have a real, not ceremonial, role in the decisions that shape their day.
The third issue is the economic relationship. Employed physicians now make up the majority of the US workforce. Most are paid through productivity formulas that reward volume and quietly ignore the unpaid administrative load. A physician who spends an hour fighting a denial is not generating revenue under that formula, but the hour is real, and the denial work is required. Hospitals that adjust compensation to recognize the full scope of the work, including what payers and vendors have offloaded onto clinicians, recover loyalty that wellness programmes cannot buy.
There is one more shift coming, and it is the one most hospitals are least prepared for. As AI takes over more of the documentation, coding, and pattern recognition work that physicians used to spend their time on, what remains is judgment. The space between what an algorithm produces and what a clinician must decide is not narrowing. It is widening. AI delivers a probable diagnosis. Whether to act on it, in this patient, with this history, in this institution, under this insurance plan, with this family at the bedside, is judgment work that does not transfer. Hospitals that imagine they are buying labour replacement are buying the wrong product.
Judgment is not a single skill, and this is where most physician roles will need to change. The clinician who can move between clinical reasoning, operational logic, financial pressure, and policy context is the clinician an AI-augmented hospital actually needs. Cross-pollination is no longer a niche capability of the physician executive. It is becoming the architecture of the physician role itself. The institutions that recognize this and build for it, with deliberate exposure across domains, governance authority that crosses silos, and compensation structures that recognise integrative work, will find themselves on the right side of the workforce curve. The institutions that keep narrowing physicians into ever-thinner specialty channels while AI eats the documentation around them will find their physicians replaceable in the wrong direction first.
None of this is a soft intervention. It requires operational discipline and a willingness to renegotiate relationships with vendors and payers that have benefited from the current arrangement. It also requires a different kind of physician leadership. A chief medical officer who functions as an internal communicator for executive decisions is doing a different job than a physician executive who has authority over the operating model. Hospitals that have not made the second role real are unlikely to solve this problem with the first.
Healthcare leadership often treats the workforce question as separate from the strategic question. It is not. A hospital that cannot retain its physicians cannot deliver on anything else it has promised, whether that is growth, quality, or cost. The financial models that look attractive on paper assume a physician workforce that is increasingly unwilling to deliver them.
What does this look like as an executive committee agenda item next quarter?
Start with measurement. Most hospital systems have never run a real documentation audit on their own physicians. Pick a representative panel across primary care, medicine, and surgery, and measure, in hours per week, how much time they spend on tasks that do not require a medical degree. Prior authorisation work. Inbox triage. Form completion. Coding queries. The number usually falls between 15 and 25 hours per physician per week, and most leadership teams are surprised by the result. Publish it internally. The audit itself changes the conversation, because once the number is on a slide, it becomes harder to keep treating the workforce question as a wellness question.
The next move is subtraction. Pick three administrative requirements currently imposed on physicians that the institution itself imposes, not the payer or the regulator, and take them off. Duplicative training modules. Internal forms no one reads. Approval workflows for decisions that physicians are already credentialled to make. Whether a hospital can identify three things to remove is itself diagnostic. The ones that cannot are usually the ones that have not actually looked.
Then restructure one governance committee, so physicians have a vote rather than a voice. Schedule design, technology selection, care pathway approval, and capital allocation. Any of these will do. Move clinicians from advisory to decision-making in one of them and watch what happens to retention in that service line over the following year. The signal is usually clear within twelve months.
And then there is compensation, which is where most of these conversations stall. Employed physicians are paid through productivity formulas that reward volume and ignore the unpaid administrative work the institution itself requires of them. Adjust one formula to credit some portion of that work, denial appeals, or care coordination time, or inbox management above a defined baseline, and the message lands. The dollar amount matters less than the fact that the institution has finally acknowledged the hour exists.
None of this is a transformation plan. It is what a hospital executive committee can authorise in a single quarter without new capital, new vendors, or new regulations. The systems that are still presenting burnout dashboards to their boards in 2028 without having done any of this will be the systems explaining why their physician vacancy rate keeps getting worse, and their growth projections keep getting revised down.
Wellness is not the answer. Neither is the next AI procurement. The answer is to stop producing the conditions that require either of them, and that work starts with whatever the executive committee chooses to put on its agenda this month.
