Less Crime

Comparing rates of custody and non-custody deaths

The police and crime commissioner (PCC) for Northumbria, former solicitor-general Vera Baird, recently announced that she will “take any necessary action” following the death of two men in a week in Northumbria Police custody. The announcement is, of course, a good thing: part of the role of PCCs is to set the local police budget, so if any changes are needed she will be able to fund them.

Baird is right to reassure her constituents, but is two deaths in a week necessarily cause for them to be concerned? One answer to this question is to say that “even one death in police custody is too many” but (given that a lot of people get arrested) that may not be realistic. A more useful question to ask might be: how many deaths in police custody would we expect if the fact of a person being arrested had no bearing on their death? Another way of thinking of this: how many people who die in police custody were going to die even if they weren’t in custody?

The rest of this post involves back-of-an-envelope calculations. Remembering that this is a blog post and not an academic research paper, the results of these calculations should be considered accurate to within an order of magnitude or so, at best.

In 2011–2012 (the most recent year for which figures are available) there were 15 deaths in police custody in England and Wales. Unfortunately, converting this number to a rate (for comparison with other figures) is difficult because there are no published figures on the number of people who are arrested each year.

How many people get arrested?

The Home Office only counts arrests for ‘notifiable offences’, of which there were 1.4 million last year. This includes arrests for all serious crimes, but omits arrests for some crimes for which a lot of people get arrested — of the 15 people who died in custody last year, only seven had been arrested for notifiable offences.

Two examples of non-notifiable offences are drink driving and being drunk and disorderly. We don’t know how many people get arrested for these offences each year, but we do know that 34,000 offenders were given a fixed-penalty notice for being drunk and disorderly (almost all of these would have been issued in custody) and that 93,000 people failed a breath test (which almost always results in the person being arrested).

As well as arrests for non-notifiable offences, people can be arrested if they have deserted from the armed forces, if they entered the country illegally or if a warrant has been issued for their arrest. Statistics are not available on all arrest warrants, but in the past year 68,000 people have been arrested for failing to appear at court. There are also some people in police cells because they have been detained by police officers under the Mental Health Act, despite more-or-less everyone agreeing that these detainees should be taken to hospital instead.

Overall, the number of people arrested each year may be upwards of two million, a number which corresponds with an estimate by the Independent Police Complaints Commission (IPCC) of just over two million people in police custody in 2006.

Calculating a rate of deaths in custody

In order to compare deaths in custody to deaths elsewhere, we need to be able to estimate the total time spent in custody. This is tricky because it involves multiplying one approximate number by another very large approximate number, so the result will be both very large and very approximate.

The Home Office estimated in 1998 that the average stay in police custody is about seven hours, while the same IPCC report mentioned above found that people detained under the Mental Health Act in 2005/06 spent an average of nine-and-a-half hours in custody and the Department of Health (which monitors prisoner healthcare) recently estimated that the average prisoner is detained for 11 hours. Some categories of prisoners are typically held for longer than others: immigration offenders arrested by UK Border Agency staff are held for an average of 18 hours in police cells.

Taking these various reports together, we might assume the average prisoner is detained at a police station for eight hours. Given two million arrests per year, prisoners spend a very-approximate total of 16 million hours (667,000 days) in custody each year. This is equivalent to 2.25 deaths for every 100,000 person-days in custody, or one death per 44,000 person-days.

How do the rate of custody and non-custody deaths compare?

Of the 63.2 million people living in the UK in 2011, 552,000 of them (0.8%) died during that year, or 2.40 for every 100,000 person days. This is slightly more than the rate of deaths in custody, but given the broad assumptions we’ve already made it’s probably reasonable to think of these numbers as ‘similar’ rather than any more-specific comparison.

This suggests that the rate of deaths of prisoners in police custody is similar to the rate of deaths in the general population. However, it doesn’t mean that all, or even most, of the people who died in custody were going to die anyway, because we don’t know if the people in custody are similar to the population overall. We know that 84% of people who get arrested are men and 70% are aged 21 or over, but that is about all we know.

There are some reasons to think that the population in police custody is at a lower risk of dying than the population at large. A 65-year-old is ten times more likely to die in a year than a 35-year-old is, so the age of prisoners would be an important factor in determining their risk of dying in custody. The age-crime curve suggests that most people in custody will be young (although I haven’t been able to find any detailed statistics on prisoner age), so that may indicate that prisoners are at a lower risk of dying than the general population. Similarly, prisoners are regularly monitored (at least every hour, and in many cases more frequently), whereas people outside custody could be taken ill or be injured and not be discovered for many hours.

On the other hand, there is reason to think that prisoners are at a greater risk of dying than other people. Homeless people, alcoholics and drug addicts are all over-represented in police custody and all have much lower life expectancy than the population as a whole. People who are ill are also over-represented in police custody: in one recent survey of prisoners at one custody suite in London, 40% needed to take prescription drugs while detained. Finally, many prisoners are drunk when they are brought into custody, putting them at risk of alcohol poisoning, injuries from falling, dehydration and so on. Taken together, drugs, alcohol or long-term illness were listed as the primary cause of death for ten out of the 15 people who died in custody last year1, so these certainly seem to be significant contributors to deaths in custody.

These factors may balance one another, giving prisoners a similar risk of death to everyone else, but equally they might not. All we can reliably say is that the rate of deaths in custody is similar to that in the population at large.

1 Of the five other people who died in police custody, one choked on a package of drugs, one died from a head injury sustained before arrest and one was poisoned by anti-freeze he drank before arrest. The cause of death for the final two people had not been determined when the latest IPCC report was published.