Most conversations about physician burnout start with hours, and hours matter. But the more useful question is what those hours are spent on. A growing body of work points to a specific culprit: the administrative and documentation load that sits on top of clinical care, much of it invisible on the schedule and a lot of it done after everyone else has gone home.
This piece looks at burnout through that lens, leans on one well-documented finding rather than a pile of invented statistics, and tries to be honest about where technology actually moves the needle and where it does not.
The clearest number we have: roughly 2:1
The single most-cited measurement of this imbalance comes from Sinsky et al., published in Annals of Internal Medicine in 2016. Using direct observation of ambulatory physicians, the authors found that for every hour of direct clinical face time with patients, physicians spent close to two additional hours on electronic health record and desk work during the clinic day. The ratio was roughly 2:1 in favour of the screen, not the patient. That finding has stuck around because it reframes the problem: the fatigue many clinicians describe is not simply about seeing too many patients. It is about the work that wraps around each encounter — the clicking, the order entry, the documentation, the inbox. The patient in front of you is the part that felt like medicine; the rest is what made the day long.
"Pyjama time" is a documented phenomenon, not a metaphor
The Sinsky study also noted that documentation does not stop when the clinic closes. The after-hours EHR work that spills into evenings and weekends has acquired its own informal name in the literature and among clinicians: "pyjama time" — the stretch of charting done at home, after dinner, in front of a laptop. It is widely discussed precisely because it is so common, and because it quietly erodes the boundary between work and recovery that protects against burnout in the first place.
What the burnout data does and doesn't say
Here it is worth being careful. Physician burnout is extensively documented by organisations that survey clinicians year after year — the American Medical Association and Medscape's annual reports are among the most widely referenced — and the qualitative picture is consistent: large proportions of physicians report exhaustion, and administrative burden ranks at or near the top of the contributing factors they name. What you should be wary of is any single, confidently quoted burnout percentage, because the figures move by year, specialty, country, and how each survey defines burnout. The responsible claim is the directional one: burnout is prevalent, it is repeatedly measured, and clerical load is one of its most consistently cited drivers. If you need a precise number, take it from a named report and a specific year rather than from a blog.
The takeaway
The Sinsky 2016 ~2:1 ratio of EHR and desk work to direct patient care, plus the well-documented spread of after-hours "pyjama time" charting, tells the core story: a large share of burnout traces back to administrative burden, not face time with patients. That is the part technology can realistically attack — and the part it cannot.
Where technology genuinely helps
If documentation is a measurable share of the burden, then anything that reduces the keystrokes per encounter is attacking the right target. This is the honest case for clinical tooling — not that it fixes burnout, but that it can shrink one of its biggest measurable inputs. We've written more on this in the hidden cost of clinical documentation.
- Ambient and voice-to-note documentation. A scribe that turns the spoken consultation into a structured draft removes the after-visit retyping that fuels pyjama time. In Shifaa AI, the voice-to-SOAP scribe transcribes with OpenAI's Whisper and drafts with Anthropic's Claude — and it fills empty fields only, never overwriting what the doctor wrote.
- Surfacing instead of searching. Pulling drug-interaction, allergy, and dosing checks, or ranked differentials with citations, into the moment of care saves the clerical detour of looking things up across tabs. The doctor still decides; the tool just shortens the path to the evidence.
- Reducing context-switching. Records, vitals, queue, and notes living in one place cuts the swivel-chair tax of moving between systems — a small per-task cost that compounds across a full clinic day.
Where it does not — and shouldn't pretend to
Software can lower documentation load. It cannot, by itself, fix the reasons that load exists. Many of the deepest drivers are systemic, and naming them honestly matters more than overselling a feature.
- Documentation requirements are externally imposed. Billing rules, regulatory checkboxes, and medico-legal defensiveness create much of the charting volume. A better editor does not change what you are obliged to record.
- Throughput pressure is organisational. Short appointment slots and understaffed clinics produce burnout regardless of how good the note-taking gets. Faster notes can quietly become an excuse to book more patients.
- Inbox and message volume keeps growing. Patient portals, results, and refills generate work that no scribe touches.
- Autonomy, culture, and workload sit upstream of any app. Tools are a lever, not a cure.
So the realistic framing is this: technology that genuinely reduces documentation burden is worth pursuing, because the Sinsky data tells us that burden is large and that after-hours charting is real. But a clinical assistant earns trust by being modest about its scope — it shortens the clerical tail of each encounter and leaves the diagnosis, and the decision, with the doctor. The systemic causes of burnout need policy, staffing, and workflow change. The best a well-built tool can do is stop adding to the pile, and take a measurable bite out of the part that was always avoidable.