Rostering an Emergency Department: Coverage That Cannot Fail

Most rosters answer one question: is there someone on the shift? An emergency department roster has to answer a harder one: is there someone on the shift who can do each specific thing the shift might demand — run a resus, hold triage, supervise an intern, take charge? Headcount is the easy half. The hard half is the mix.

That difference drives everything unusual about ED rostering — why every shift looks the same on paper and no two are interchangeable, why swaps that would be trivial in any other workplace need sign-off, and why the roster is owned by someone senior rather than delegated to whoever has time.

The 24/7 baseline

An ED cannot close, so the roster is three shifts, seven days, no gaps — 21 shift instances a week before anyone takes leave. The pattern below is a common mid-size shape.

The overnight thinning is deliberate, not budgetary. Presentations drop after midnight, but note what does not thin proportionally: registrars. Nights run with fewer juniors per senior because supervision is the scarce resource at 03:00 — an intern with a question at 3am needs an answer within minutes, and there is no ward round coming to catch a mistake at nine.

Roster the skill mix, not the headcount

“Six RNs on days” is not a rosterable requirement. The real requirement hiding inside it looks more like: at least one triage-qualified nurse at the front at all times, at least two staff current in resuscitation, at least one who can access the medication room unsupervised, and a team leader.

This is why an EN cannot simply fill an RN vacancy even when the count says the shift is covered. Enrolled and registered nurses have different scopes of practice — an EN works under RN supervision and cannot hold certain responsibilities alone. A shift with the right number of bodies and the wrong mix is uncovered in every way that matters, while looking fine on the spreadsheet.

Why every swap needs sign-off — and why that is not bureaucracy

In a café, two staff swapping a shift is their business. In an ED it cannot be, because the two people swapping do not have the information needed to know whether the swap is safe. Each sees their own qualifications. Neither sees the whole shift.

A concrete failure: two RNs swap a Tuesday night. Both are competent, both are senior, the swap looks utterly harmless. But the one leaving was the only triage-qualified nurse on that night, and the one arriving is not. The shift is now non-compliant, and nobody involved did anything careless — the information needed to catch it simply did not live with either of them.

So the rule in an ED is the inverse of the rule everywhere else: every trade goes through the nurse unit manager or the consultant who owns the medical roster, because they are the only ones holding the full coverage picture. The approval is not a trust judgement about the people. It is a coverage check that only one seat can do.

What makes this bearable is speed. Staff tolerate an approval step that answers in hours; they stop requesting swaps — and start burning leave or calling in sick instead — when it takes days. If you gate everything, you owe the team a fast gate.

Handover is rostered time

The table shows clean boundaries — 07:00, 15:00, 23:00 — and no ED actually works that way. Clinical handover takes 15 to 30 minutes of structured overlap, and it is the highest-risk moment of the shift: the departing team knows things about every patient in the department that exist nowhere in writing at the level of detail the next hour needs.

Roster the overlap explicitly — incoming staff start 15 minutes before the boundary, or outgoing staff are paid 15 past it. An unrostered handover is one that gets squeezed the moment the department is busy, which is precisely when it matters most.

Nights: rotate slow, recover properly

Night coverage is the part of the roster that damages people when done casually. Two principles cover most of it:

• Block nights together. Three to four consecutive nights lets the body adapt once, rather than whiplashing between schedules. Alternating single nights and days is the most fatiguing pattern there is.
• Protect the recovery. The days after a night block are not free days — the first is physiological recovery, and rostering someone for an evening shift the day after their last night is rostering them impaired.

Fatigue in an ED is not a wellbeing line-item; it is a clinical risk with the same standing as the skill-mix rules. The roster is where it gets managed, because by the time it is visible on the floor it has already cost something.

Plan the gap before it happens

Sick leave in a hospital does not wait for convenient shifts, and an ED cannot run short the way a shop can. Three layers, in the order you should exhaust them:

• A published escalation order — who is called first for an overnight RN gap, and who decides between a casual-pool fill and stretching the team. Decided in daylight, executed at midnight.
• A casual or float pool of staff already credentialed for your department. The middle of a staffing crisis is the wrong time to check whether someone has done a resus course.
• Named on-call seniors — not “someone will come in” but a person, per night, who knows it is them.

Putting it together

Three shifts with deliberate overnight thinning; coverage defined as a skill mix with explicit per-shift minimums, not a headcount; every swap approved by the one seat that can see the whole picture — quickly; handover overlap rostered as real time; nights blocked and recovery protected; and the escalation path written before it is needed.

None of it is exotic. Every functioning ED does most of this already — the difference between the ones that run smoothly and the ones held together by their NUM’s memory is whether the rules live in a system anyone can check, or in a head that eventually takes leave.