Every line in the Ministry of Justice's prison deaths table is a person who went into custody and did not come out. The table stretches back to 1978. It records how each of those people died: self-inflicted, natural causes, homicide, other. And then there is a category that records nothing at all, because nobody has decided yet. It is called 'Awaiting further info'.

For deaths in 2024, that category holds 45 cases. For deaths in 2023, it holds 10. Both figures come from the same publication, covering 1978 to 2024 (Source: Ministry of Justice, Safety in Custody -- Deaths_in_prison_custody_1978_to_2024_accessible -- Table_1_2). As a year-on-year change that is a rise of 350%. Read as a percentage, it sounds like something went badly wrong in 2024.

Here is how we got here, and it is less dramatic and more uncomfortable than the percentage suggests.

A death in custody is not classified on the day it happens. It waits. There is an investigation, then a coroner's process, and until a conclusion lands the statisticians cannot honestly write anything in the cause column. So the file gets published with a placeholder. That means the most recent year in any edition of this table is always the least resolved year. Last year's deaths have had twelve extra months of process behind them. This year's have had none.

Which is why the honest comparison is not 45 against 10. It is 45 against what 2024 will look like once the same twelve months have passed. On the evidence of 2023, most of those 45 lines will eventually acquire a recorded cause. The most likely reading is that this is a queue, not a surge, and that the 350% figure is measuring the passage of time rather than any change in how people are dying in prison.

The number that actually deserves attention is the smaller one. Ten deaths from 2023 were still unclassified when this file was published alongside the 2024 figures. Those are cases that have had a full extra year and still carry no cause. If the backlog cleared cleanly, that figure would be zero.

The strongest alternative reading, and it matters: 10 is a very small base, so any movement against it produces a large percentage. There may also have been a genuine rise in deaths in 2024, or a change in the timing of coroner conclusions, or a change in how the Ministry applies the category. This table alone cannot separate those explanations. And these are the most recent available figures, covering 2024, so the classification position has probably moved since.

My own view, and I will own it: the placeholder is the most honest cell in the spreadsheet. A statistician writing 'Awaiting further info' is refusing to guess about how someone died in the state's care. That is exactly right. But 45 open cases from one year, with 10 still open from the year before it, is a reminder that the system counts prison deaths faster than it explains them.

Forty-five families are somewhere in that gap. The number will fall. The question is how far, and how fast.