Start with the number, because the number is the whole story. In the Ministry of Justice's deaths in prison custody tables, the row labelled 'Over one year' records 22.0 for 2024. (Source: Ministry of Justice, Safety in Custody -- Deaths_in_prison_custody_1978_to_2024_accessible -- Table_1_3) That is the most recent year published. It is not this year's figure, and it is worth saying that plainly before anyone reaches for a conclusion about what is happening in prisons right now.

Twenty-two is a small number on a page and a very large one in a family's life. Divide the year by it and you get a death in this category roughly every 17 days. Call it just under two a month, every month, for twelve consecutive months. That is the arithmetic. There is no drama in it and it does not need any.

What makes this line sit uncomfortably is how little the label tells you. 'Over one year' is a bucket. It sorts deaths by a duration, and the person inside the bucket disappears into it. You cannot tell from the label alone who they were, how they died, or what the prison did or did not do in the days before. The category is doing administrative work, not explanatory work.

My view, and I will own it as opinion rather than dress it up as analysis: a public safety statistic should be legible to the public. A row that reads 'Over one year: 22.0' passes the transparency test on paper and fails it in practice. Anyone can download it. Almost nobody can interpret it without the accompanying notes, and most people never see the notes. Publishing is not the same thing as explaining.

The decimal is its own small puzzle. Deaths are whole people. A figure of 22.0 in a table about deaths either means twenty-two, formatted with an unnecessary decimal place, or it means something other than a straight count. Those are two very different facts wearing the same clothes, and the reader is left to guess which one they are looking at.

Here is the strongest honest argument against making too much of this. A single cell lifted out of a large statistical release is the weakest possible evidence base. Without the surrounding years, the prison population denominator, and the definition attached to 'Over one year', the figure cannot tell you whether things are getting better or worse. A count of 22 in a system holding tens of thousands of people could represent an improvement, a deterioration, or no meaningful change at all. Anyone claiming a trend from one number is guessing, and that includes anyone who tells you it is reassuring.

So take the one thing the data does support and stop there. In 2024, in this category, the Ministry of Justice recorded 22.0. Twenty-two reasons for someone to be waiting on an answer. The figure is public. The explanation, on the face of the table, is not.