There is a term that has quietly become a fixture in conversations about physician burnout: pajama time. It refers to the hours a doctor spends finishing clinical notes, processing inbox messages, and completing documentation after the workday has technically ended. Not at a desk in a clinic. At home, often in the evening, on a laptop that never quite leaves the bedroom.
For most Canadian primary care physicians, pajama time is not an occasional inconvenience. It is a structural feature of the job. And the data measuring it has become precise enough to demand a serious response.
The hidden clock: quantifying the primary care charting crisis
Primary care physicians now spend more time on documentation than on the patient visits that generate it. AMA 2024 data shows that of a 57.8-hour physician workweek, only 27.2 hours go to direct patient care. Canadian physicians average 10.4 hours per week on administration, rising to 19 hours per week for Ontario family physicians. A landmark study measured 36.2 minutes of EHR time per 30-minute primary care visit, including 6.2 minutes of after-hours pajama time. Ambient AI documentation tools are now measured to save Canadian physicians an average of 64 minutes per day.
The documentation burden in primary care has been studied extensively, and the numbers that have emerged are striking in their consistency. The American Medical Association's 2024 physician data found that of a 57.8-hour average physician workweek, only 27.2 hours go to direct patient care. The remainder is absorbed by documentation, inbox management, order entry, and administrative tasks that electronic health record systems have redistributed onto physicians over the past two decades, steadily eroding the electronic health record efficiency that these systems were supposed to deliver.
Canadian physicians face the same structural reality. The CMA's 2025 National Physician Health Survey found that Canadian physicians spend an average of 10.4 hours per week on administrative tasks. In Ontario, the Ontario College of Family Physicians found family physicians spending 19 hours per week on administration, nearly half a full working week consumed before a single clinical encounter begins.
Breaking down the 36.2-minute EHR visit metric
A cross-sectional study of 307 primary care physicians at Massachusetts General Hospital and Brigham and Women's Hospital produced one of the most cited figures in the documentation burden literature. Primary care visits are scheduled for 30 minutes. Physicians in the study spent an average of 36.2 minutes on the EHR per visit, with a third quartile reaching 45.7 minutes.
That 36.2-minute figure included 6.2 minutes of pajama time per visit, charting completed outside clinical hours. Industry analysis from healthcare technology research puts the ratio even more starkly: for every 15 minutes spent with a patient, physicians spend an average of nine minutes on documentation. The research team at MGH and Brigham and Women's concluded directly that EHR time burden represents a serious threat to the primary care physician workforce.
The arithmetic is straightforward: documentation is consuming more time than the visit that generates it.
The hidden toll of after-hours charting beyond the clinic day
Research published in the Journal of Internal Medicine found that physicians accumulate an average of 1.2 hours of pajama time on clinic days and 1.3 hours on weekends. The AMA's own survey data shows that 20.9 percent of physicians spend more than eight hours per week on the EHR outside normal working hours, a figure that has not improved meaningfully since 2022.
For a physician seeing patients five days a week, the weekly pajama time total frequently exceeds two hours. For those carrying heavier patient panels, it climbs further. This is time subtracted directly from recovery, family, and personal life; the same time that, when consistently eroded, accumulates into the chronic depletion that drives physicians toward early retirement or reduced clinical hours.
In Canada, where 54 percent of physicians are already considering reducing their clinical hours in the next 24 months, this is not a theoretical risk. It is an active withdrawal from a system that cannot absorb it.
The varied spectrum of ambient AI impact: what the data shows
Ambient AI documentation works by capturing clinical encounters and generating structured draft notes that physicians review, edit, and approve before they enter the record. Approaches vary across platforms: some tools transcribe conversations in real time during the visit, while others process a recording the physician makes immediately after the encounter. In both models, no note enters the chart without physician sign-off. The goal is the same in either case: eliminate the documentation tail that follows clinicians home at the end of the day.
The evidence base for medical charting automation has grown substantially since 2023, spanning controlled trials, large-scale health system evaluations, and multi-centre observational studies. The findings are directionally consistent and worth examining in detail, because the variance across studies carries important implications for how these tools perform in practice.
| Study and source | Scope | Measured outcome |
|---|---|---|
| Kaiser Permanente, Tierney et al. (2023–2024) | Large integrated health system, 63-week evaluation | 15,700 physician hours saved in one year; significant reductions in note-taking time and pajama time |
| UW Health Randomised Trial (2025) | Operational RCT, Aug 2024 to Mar 2025 | 30 minutes saved per day per provider; clinically meaningful reduction in burnout scores |
| UCLA Study, Lukac et al., NEJM AI (2025) | 238 physicians, 14 specialties, two-month trial | 9.5 percent statistically significant reduction in time-in-note for Nabla users versus control |
| Multi-study synthesis, Penn and Sutter Health | Multiple outpatient implementations | 2.5 to 3 hours reduction in after-hours documentation per week |
| Multi-centre study, STAT News (2026) | 1,800 clinicians, five academic medical centres, 2023 to 2025 | 16 minutes saved per 8-hour shift; primary care clinicians benefited most |
| Doctors of BC Pilot (2025) | 30 community-based physicians, 7,000 appointments | Physicians described it as practice-changing; significant reduction in administrative burden reported |
The Canadian context adds a specific data point: the CMA and CFIB national survey found that 28 percent of Canadian physicians already use an AI scribe, saving an average of 64 minutes per day. A further 42 percent expressed interest in adopting one.
Why implementation quality dictates automation outcomes
The variance in the data table is not random. It reflects something important about how clinical workflow optimization through ambient AI performs in practice.
The 16-minute-per-shift finding from the large multi-centre study represents average usage across 1,800 clinicians, many of whom used the tools inconsistently. The 30-minutes-per-day and 64-minutes-per-day figures come from settings where tools were integrated deliberately into clinical workflows and where adoption was actively supported. A tool that is available but not embedded in how a clinic actually operates produces marginal gains. A tool configured to match existing rhythms and introduced with structured support produces substantially larger ones.
The implementation question is one our clinic adoption guide addresses in full, and it is worth reading before any evaluation of ambient documentation begins.
The Doctors of BC pilot illustrates the point well. Community-based primary care physicians in a supported trial described the experience as practice-changing, a conclusion that reflects the difference between a tool handed over and a tool properly adopted.
Joud Health AI: a platform built for both layers
Most discussions of AI in primary care treat the documentation problem and the operational problem as separate challenges requiring separate solutions. Joud Health AI is built around the recognition that they are the same problem: physician and staff time consumed by work that does not require human judgment.
Joud operates through two dedicated portals, each designed for a distinct role in the clinic.
The administrative staff portal handles the operational layer. The AI voice agent answers inbound calls around the clock, covering bookings, rescheduling, lab result routing, and general inquiries without a hold queue. Digital intake forms collect patient information before the appointment begins, eliminating the re-entry cycle that consumes front-desk time. Automated check-in removes another manual interruption from the clinic morning. All of this syncs natively with the clinic's Canadian EMR, meaning appointments, intake records, and check-in data flow directly into the system the clinical team already uses, without manual transfer or duplicate entry.
The physician portal is a different kind of tool entirely, and worth describing with precision.
When a physician opens the portal at the start of the day, they see their full appointment schedule alongside each patient's check-in and arrival status, updated in real time through the EMR sync. Before entering the exam room, they have access to an AI-curated patient profile: a summary of previous visits, current medications, recent labs, and the reason for today's appointment, drawn from the EMR and presented in a single clean view. The fragmented chart-hunting that consumes the first minutes of too many appointments is replaced by a prepared, consolidated picture of the patient before the encounter begins.
During or immediately after the visit, the physician clicks to record. Joud transcribes the recording and surfaces the full transcript as soon as recording ends. The physician then generates a SOAP note from the transcript with a single click. That draft can be reviewed, edited, and signed within the portal, and synced directly to the EMR. The note does not follow the physician home. It closes in the clinic.
The physician portal is currently in active beta with select Canadian clinics, developed alongside the operational suite so that the EMR integration, compliance framework, and data infrastructure are shared across both portals from the ground up.
When both layers function together, the effect is compounding. A physician who enters each appointment already briefed, documents immediately after each encounter, and whose clinical team is no longer fractured between phone calls and in-person patients operates in a structurally different environment than one where only a single bottleneck has been removed.
Moving from documentation fatigue to practice liberty
The recovery available to Canadian primary care is not hypothetical. It is already being measured.
Physicians using ambient AI documentation consistently report not just efficiency gains, but a qualitative shift in how their workday feels and how present they are able to be with patients. The note that once followed them home closes before they leave the clinic. The weekend charting that used to consume Sunday afternoon begins to disappear.
The 64 minutes per day that the CMA/CFIB survey identifies as the average Canadian AI scribe saving compounds across a working year into something significant. Across a 46-week working year, 64 minutes per day returns roughly 245 hours of physician time. That is more than six full working weeks reclaimed from documentation that should never have followed anyone home.
The documentation burden in Canadian primary care is real, it is measured, and it is solvable with tools that now carry a substantial clinical evidence base. The question for clinic owners and medical directors is not whether to engage with this. It is how soon.
Elevation Labs builds clinical-grade operational infrastructure for Canadian primary care. Book a demo to learn more.