Why Your 80-Hour Duty-Hour Tracker Gets Emergency Medicine Wrong

1. The one-day-off-in-seven that can't be averaged

Under the Common Program Requirements, a resident must average one day off in seven over four weeks (§6.21.b). “Averaged” is the operative word: three days off in week four can cover a week with none.

Emergency Medicine removes the averaging:

§6.17.a.4 — Emergency medicine residents must have a minimum of one day (24-hour period) free per each seven-day period. This cannot be averaged over a four-week period.

Every single week must contain a 24-hour day off. A schedule that's perfectly legal for an internal-medicine resident — front-loaded weeks, days off clustered at the end of a block — is a violation in EM. A tool that checks the four-week average passes it anyway. That's not a rounding error; it's a structurally different constraint.

2. Rest measured against the previous shift, not a fixed number

The common requirements suggest eight hours off between work periods. Many tools hard-code “8 hours” (or “10”) as the rest threshold. EM's rule is relational:

§6.17.a.2 — There must be at least one equivalent period of continuous time off between scheduled work periods.

Rest must be at least as long as the shift that preceded it. After a 12-hour shift, eight hours off isn't enough — you need twelve. The classic failure is the “clopen”: a night shift ending at 07:00 followed by an evening shift at 15:00 the same day. Eight hours of rest, where the rule required at least as much as the night shift. A fixed-threshold checker waves it through.

And rest runs from the end of all scheduled activity — including conference. A resident who finishes a clinical shift and then sits in two hours of required didactics starts the rest clock when didactics end, not when the shift ended. Tools that only look at clinical shifts miscount.

3. Two weekly caps the 80-hour number hides

EM rotations carry two limits the generic 80-hour view never surfaces (§6.17.a.3): no more than 60 scheduled hours/week seeing ED patients, and no more than 72 total hours/week on EM rotations. A resident can be under 80 hours overall and still violate the 72-hour EM cap once conferences are added in. If your tool only tracks the 80, you won't see it coming.

Why this is hard to bolt on

These aren't settings you toggle on a generic engine. “No averaging” changes the shape of the constraint — it has to be checked every week, not once per block. Relational rest depends on the preceding shift's length, computed from the end of all activity. The two weekly caps interact with how you classify each hour. Get the model wrong and the tool is confidently, invisibly incorrect — which is worse than no tool, because it tells a coordinator a broken schedule is fine.

SchedulerRX was built for EM from the constraint up. The no-averaging day-off rule is a hard constraint in the solver — a schedule that violates it is infeasible, not a warning you can click past. Rest is computed against the preceding shift, including conference time. The point isn't a longer feature list; it's that the EM rules generic tools get wrong are the ones the engine treats as non-negotiable.

FAQ: EM Duty-Hour Rules Generic Tools Miss