Alcohol Concern has today published a report surveying how well London has done in reducing the harms of alcohol. It’s a useful audit of the funding and delivery of services for those who need them.

But the press release, drawn from a single claim made by the report, gives a false impression of the nature of the problem. It is headed “Alcohol-related crime higher in London than rest of England” and it is drawn from a single paragraph in the report, which reads:

“London has the highest rate of alcohol-related crime in England. London’s rate of alcohol-related crimes is significantly higher than any other region in England (12.2 vs 8.1 per 1000). London also has the highest rates of alcohol-related violent crimes and of sexual offences.”

No source is given for these figures, but a bit of searching reveals where they came from: the North West Public Health Observatory’s Alcohol Profiles for England, which give estimates for alcohol-attributed crime, violent crime, and sexual crime. This provides the figure of 12.2 per 1000 for alcohol-attributed crime for London in 2009-10, against an average for England of 8.1. It also shows London is higher for alcohol-attributed violent crime (8.45 per 1000 vs 5.83 for England) and for alcohol-attributed sexual crime (0.17 vs 0.13).

The same report reveals that the City of London is far the most dangerous place in the capital, with a figure for alcohol-attributed sexual crime three times higher than that for London as a whole. Bet that came as a surprise! Few people, even the most anxious, anticipate a sexual assault in the Square Mile, even though plenty of drink is consumed there. It’s a consequence, of course, of a low resident population rather than a high level of crime.

That apart, how are these figures worked out?  The NWPHO takes total recorded crime and applies a multiplier to it:  the proportion of crime committed by an offender under the influence of alcohol. These fractions are worked out from the number of people arrested who test positive for alcohol, data coming from 16 police stations in England & Wales. The figures given by the NWPHO are 0.37 for violence against the person, and 0.13 for sexual offences.

For all crime it explains that the fraction is worked out from summing the attributable fractions for six offences: violence, sexual offences, robbery, burglary, theft of a motor vehicle, and theft from a motor vehicle.

Whatever the multiplier may be, it is applied indiscriminately across the country. So to get local figures for alcohol-attributable violent crime you simply multiply local crime rates by 0.37, and sexual crime rates by 0.13, wherever the offence took place.

Crime overall happens to be higher in London than elsewhere, so applying a simple multiplier means that alcohol-related crime is, by definition, also higher in the capital. Take violence against the person: 23 per 1000 in London, 16 per 1000 in England (source: Crime in England & Wales 2009-10). Multiply  by 0.37 and you get 8.51 in London, vs 5.92 in England – figures close but not identical to those quoted by NWPHO, probably as a result of rounding in the published data.

To sum up, the data tells us nothing about whether Londoners are more disposed than others in the rest of England to commit crimes when intoxicated. The use of a fixed multiplier simply translates its greater crime rate into a greater alcohol-related crime rate. It tells us nothing we didn’t already know: crime in higher in the capital city and (if you believe this methodology) worryingly high in the City of London.

I really shouldn’t have to explain to a charity that specialises in alcohol abuse how to interpret these figures.