Nearly every rota that does not work in a round-the-clock setting has the same origin, and it is not the rotation: the staffing sum was never done. Somebody started from the people available and tried to fit the cover around them, rather than the other way round.
It is worth redoing, because it is a two-minute calculation that settles years of argument.
The sum
Covering one post twenty-four hours a day, seven days a week, is 168 hours a week. In eight-hour shifts that is twenty-one shifts.
On a 37.5-hour contract that needs 4.5 full-time people for that single post. On a 40-hour contract, 4.2.
And that is before annual leave, sickness, mandatory training and parental leave. Add a realistic allowance for absence and you land comfortably at five and a half people per continuous post.
The practical point: if you have a post to cover round the clock and five people, you do not have a rotation problem. You have headcount at the edge, where any absence becomes overtime, and no shift pattern can change that.
Posts, not units
The second mistake is thinking in units rather than posts. "The unit needs six people" says nothing useful: six at the same time, or six across the day?
Coverage has to be declared per band and per role:
- night: how many present, and of which profile;
- morning, nearly always the heaviest band because of personal care and medication rounds;
- afternoon;
- the handover overlaps, which are real work and should be counted as such.
Handover is the item that vanishes from the count most often. Fifteen minutes between two shifts, three times a day, is more than five hours a week per post. If it is not in the plan, those are hours somebody works unpaid or does not work at all.
Staffing standards are not a rota
Many regions and regulators set minimum care hours per resident. Those are an outcome constraint, not a plan: they say how many hours of care are needed in total, not who is on the unit at three in the morning.
You need both, and you need them separate. The standard sizes the overall establishment; coverage per band builds the week. Confusing the two produces plans that satisfy the monthly average and leave Sunday afternoon uncovered.
What to do when the sum does not work
If the sum says the establishment is short, there are three honest routes and one dishonest one.
The honest ones: recruit, reduce the number of continuous posts by merging where it is clinically safe, or use fixed-term staff in a structured way in the bands where demand is highest.
The dishonest one is to keep calling it overtime. It works for a few months, then people leave and the sum that did not work becomes a much worse sum. The signals that precede resignations are readable in the rota, and we gathered them in burnout signals you can read in a rota.
In Sked Solve coverage is declared per time band and per role, and the pre-generation check flags which bands cannot be covered with the available team before the solver even runs. See how it works at Sked Solve.
