ACGME limits resident clinical and educational work through several separate tests: weekly hours, days free, time after a 24-hour period, maximum continuous assignments, call frequency, and moonlighting. A schedule meets the rules only when the correct averaging period, specialty requirements, leave, work from home, and actual post-publication assignments are included.
Key takeaways
- The 2026 Common Program Requirements place the core work-hour requirements in sections 6.20 through 6.28; older policies may still cite the former VI.F numbering.
- A single week above 80 hours is not automatically a violation. For a full four-week rotation with no leave, the limit is 320 hours. A calendar month, shorter rotation, or leave changes the calculation under the ACGME FAQ.
- Programs should structure schedules for eight hours between work periods. Fourteen hours free after a 24-hour clinical and educational work period is mandatory and begins when the resident leaves the hospital.
- In a 2012 national survey of general surgery residents, 62.1 percent reported under-recording hours at least once and 14.6 percent did so often or always (Drolet et al., JAMA Surgery, 2013). That finding should prompt local validation, not be treated as a current estimate for every specialty.
Where do the 2026 ACGME work-hour rules live?
The governing baseline is the ACGME Common Program Requirements (Residency), 2026 edition. The work-hour rules sit in Section 6 under “Clinical Experience and Education.” Specialty Program Requirements can add constraints, so the common document is the starting point rather than the complete policy for every program.
The italicized “Background and Intent” passages in the Common Program Requirements are explanatory and cannot be cited by a Review Committee. Only the numbered requirements can. If your program policy quotes Background and Intent text as though it were binding, that is worth correcting.
| Rule | What it requires | 2026 | Legacy |
|---|---|---|---|
| 80-hour week | 80 hours maximum, averaged over four weeks, including all in-house work, clinical work from home, and all moonlighting | 6.20 | VI.F.1 |
| One day in seven | Minimum one day in seven free of clinical work and required education, averaged over four weeks | 6.21 | VI.F.2 |
| Rest between periods | Should have 8 hours off between periods; must have at least 14 hours free after 24 hours of in-house call | 6.21.a | VI.F.2 |
| Maximum work period | 24 hours of continuous scheduled clinical assignments, plus up to 4 additional hours for transitions and education | 6.22 | VI.F.3 |
| In-house call frequency | No more often than every third night, averaged over four weeks | 6.27 | VI.F.4 |
| Moonlighting | Internal and external moonlighting count toward the 80 hours; PGY-1 residents may not moonlight | 6.25 | VI.F.5 |
| Rotation exception | Text permits up to 10 percent or 88 hours with Review Committee approval; enforcement of 6.24 and 6.24.a is suspended | 6.24 | VI.G |
The former 16-hour maximum for PGY-1 residents was removed in 2017. PGY-1 residents are now subject to the same 24-hour maximum scheduled clinical and educational work period, with up to four additional hours for specified activities, as other residents. The separate 80-hour weekly average applies to everyone. Internal and external moonlighting count toward it, and PGY-1 residents may not moonlight.
How does the averaging period work?
For a complete four-week rotation with no leave, the 80-hour weekly average permits no more than 320 hours. An 88-hour week can still meet that single requirement if lighter weeks keep the average at or below 80. It does not establish overall compliance because maximum periods, days free, call frequency, specialty rules, or an approved exception may change the result.
Averaging occurs by rotation over four weeks, a calendar month, or the full rotation when it is shorter than four weeks. ACGME does not permit a rolling average across high- and low-intensity rotations. Vacation and approved leave come out of both numerator and denominator. Clinical work from home, including EHR work and patient calls, counts; reading, studying, and research done at home do not.
Programs should calculate each applicable test separately and then check the current Specialty Program Requirements and FAQ. A four-week average that passes the 80-hour test does not cure a maximum-period or specialty-specific violation.
What do 24 plus 4 and the rest rules permit?
The maximum scheduled clinical and educational work period is 24 hours. A resident may remain on site for up to four additional hours for safe transition of care, continuity of existing patient care, and education. During that period the resident cannot care for new patients, attend an outpatient clinic, or begin a new procedure. All 24 plus 4 hours still count toward the weekly average.
Programs should structure schedules to give residents eight hours off between work periods. The FAQ recognizes that a resident may voluntarily remain or return in defined circumstances, so eight hours is not expressed as the same hard numeric prohibition as the next rule. After 24 hours of clinical and educational work, at least 14 hours free is mandatory. That clock begins when the resident actually leaves the hospital, including after any permitted additional time.
For at-home call, clinical work performed from home and time spent back in the hospital count toward the 80-hour average; passive availability does not. The every-third-night limit does not apply, but at-home call cannot be assigned during the required day free. Its frequency and intensity must not preclude rest or reasonable personal time, and the program director must monitor both.
What changed in 2026?
The current edition uses the 6.2x numbering rather than the former VI.F labels. Programs updating policies should map both labels during the transition rather than assume an old citation changes the underlying rule.
Effective February 9, 2026, ACGME suspended enforcement of several requirements pending a major revision, including 6.24 and 6.24.a, which describe the rotation-specific exception up to 88 hours (ACGME suspended-requirements list). Because 6.24 is permissive rather than a standard prohibition, the published wording alone does not explain how a new or renewed request will be handled. A program planning above 80 hours should obtain current direction from its Review Committee instead of relying on the unchanged text.
How reliable are reported hours?
Reported hours are necessary but imperfect. In a 2012 survey of 1,013 general surgery residents, 62.1 percent reported under-recording hours at least once (Drolet et al., JAMA Surgery, 2013). In one pediatric program, an EHR-derived measure identified more violations than self-report in 36.8 percent of resident-blocks (Dziorny et al., PLOS ONE, 2019).
Neither study provides a current national rate across specialties. EHR activity is also a proxy: it misses non-EHR work and can capture activity outside a scheduled period. Programs should compare schedules, resident reports, and available operational records, then investigate discrepancies without assuming intent.
What should be checked after publication?
A posted schedule is only one input to compliance. Swaps, illness, leave coverage, clinical work from home, and actual departure times can change the final calculation. No reliable published benchmark establishes that swaps are the primary source of violations, so programs should measure their own changes rather than assume where the risk sits. Attending schedules require different controls, such as the circadian checks in the emergency department scheduling guide.
Required conferences count as educational work and must be included. A conference after a 24-hour period can breach the mandatory 14-hour interval; a shorter interval after an ordinary shift may conflict with the eight-hour expectation or a stricter local policy, but it is not the same ACGME test. Rotation boundaries and coverage changes should be checked with the surrounding assignments visible.
What can scheduling software establish?
Software can apply explicit rules consistently and retain an audit trail, but automation itself does not prove fairness or compliance. At one internal medicine program, an optimized model increased first-choice rotation assignments from 30.5 percent to 80.5 percent and reduced modeled night-to-day conflicts from 0.7 to 0.3 per intern (Howard et al., PLOS ONE, 2020). Only 22 of 82 residents completed the fairness survey, and the single-program result does not establish a universal effect.
A manual process can also be auditable when the rules, inputs, approvals, and results are visible. Software does not validate self-reported work or replace Review Committee interpretation. It can make it easier to re-run constraints after a change and to connect those changes with the cost of clinical coverage. Distribution metrics also need different denominators outside training, as the anesthesia coverage guide shows for FTE, contract, site, and voluntary assignments.
How is compliance monitored?
Programs monitor clinical and educational work through resident reporting and their own processes; ACGME also uses the annual Resident/Fellow Survey as part of accreditation oversight. Survey results are an indicator, not a substitute for reviewing actual assignments, leave, home work, and specialty-specific requirements.
Sponsoring Institutions may set internal review triggers that are stricter than ACGME’s published requirements. Those local thresholds should be labeled as institutional policy rather than presented as national accreditation cutoffs.
Frequently asked questions
Do shift swaps break compliance? They can. A swap may affect the 80-hour average, days free, call frequency, maximum period, or the mandatory 14-hour interval. Its effect must be calculated with the resident’s surrounding assignments and actual work.
Does moonlighting count toward the 80 hours? Yes, internal and external both. PGY-1 residents may not moonlight under any circumstances. The claim that moonlighting is excluded is outdated.
Can we still apply for the 88-hour exception? ACGME suspended enforcement of requirements 6.24 and 6.24.a on February 9, 2026, while leaving their text in the 2026 requirements. Confirm the current request and renewal process with the applicable Review Committee before planning a rotation above 80 hours.
Where to start
Map the current Common and Specialty Program Requirements into separate tests. Calculate each test over the rotation’s permitted averaging period, removing leave correctly, and include required education, moonlighting, and clinical work from home.
Review conference calendars, rotation boundaries, and actual departure times alongside scheduled assignments. Then record every post-publication change and re-run the affected tests before approval. After several rotations, the program will know whether its main gap is schedule design, reporting, leave coverage, or change control.
YouShift builds resident schedules from a program’s own rules, so duty-hour limits are encoded as hard constraints before a schedule is published, and every swap, absence, and time-off request is re-checked against those same rules afterward. The four-week block average, the 14-hour post-call window, and call frequency are tracked as the schedule is worked, not reconstructed at survey time. More than a thousand physicians schedule on it across the United States and Europe.
See what your next block looks like generated. Tell us your program’s rules and we will build it.
Sources
- ACGME, Common Program Requirements (Residency), 2026 edition
- Drolet BC, Sangisetty S, Tracy TF, Cioffi WG, Surgical Residents’ Perceptions of 2011 Accreditation Council for Graduate Medical Education Duty Hour Regulations, JAMA Surgery, 2013
- Dziorny AC et al., Pediatric trainees systematically under-report duty hour violations, PLOS ONE, 2019
- Howard FM et al., Implementation of an automated scheduling tool, PLOS ONE, 2020
- ACGME, Common Program Requirements Frequently Asked Questions, 2026
- ACGME, List of Suspended Common Program Requirements, February 2026