Resignations on a ward rarely surprise the person receiving them. They surprise them at that moment: the decision had been forming for months, and throughout those months the rota already held everything needed to notice.
The problem is that nobody reads it that way. A rota gets consulted to find out who is on tomorrow, not to understand how the team is doing. Yet the data is the same.
Five indicators the rota already contains
Concentration of heavy shifts. Not the ward's total nights, but their distribution. If three people out of twelve work half the nights, the ward is formally compliant and substantially in debt to three people.
Quick returns. The gap between one shift ending and the next starting, per person, looked at as the minimum rather than the average. Anyone accumulating short gaps has been sleeping badly for weeks, and it shows in the rota before it shows in behaviour. The sums are in rest between shifts.
An hours balance that only ever moves one way. A balance that oscillates is normal. A balance climbing for eight consecutive weeks on one person is not an hours problem, it is a signal that this person is covering something structural.
Declined swap requests. Not the accepted ones: the ones that go unanswered. They are the most direct measure of how rigid the system feels from the inside.
Availability narrowing. When someone who used to offer wide availability starts offering very little, they usually have not changed their life. They have started protecting themselves, and it is the earliest signal of all.
Why nobody looks at them
Not through carelessness. Because looking requires reading the rota per person and over time, while a rota is read per day and in the present. They are two different views of the same data, and the second is the only one most tools offer.
There is also a more uncomfortable reason: these numbers, once you look at them, ask you to do something. As long as they stay unexamined, the problem remains manageable one day at a time.
What to do when the signal is there
Name it, do not wait for the person to. "I noticed you have done six more nights than average over the last two months, how are you holding up?" is a different conversation from the one you have when the letter arrives. It is also the only one that can still change anything.
Correct it in the rota, not in intentions. "We will try to lighten your load" produces nothing. Taking two nights off next month, in writing, does.
Check whether the signal is isolated or systemic. If three people out of twelve show the same indicators, the problem is not those three: it is the staffing sum, and it is time to redo it. The method is in round-the-clock cover in care homes.
The cost of finding out late
A resignation on a short-staffed unit does not cost you one person: it costs the months of emergency cover that follow, the overtime of those who stay, and very often a second resignation, because whoever remains absorbs the load of whoever left.
Which is why these five numbers are worth the half hour a month it costs to look at them.
In Sked Solve heavy shifts, rest periods and the hours balance are counted per person and over time, so the distribution is visible without reconstructing it by hand. See how it works at Sked Solve.