Homicide data

I am writing to query the homicide statistics quoted by Dr Salib (2003). The figures he quotes for total annual homicides suggest a fall in homicide between 1979 and 2001. The source for his figures is quoted as the Office for National Statistics (ONS).

Homicide statistics are easily available through the website of the ONS and from various other sources, including Home Office statistical bulletins and the House of Commons Library. For example, Richards (1999) describes homicide trends between 1945 and 1997, demonstrating the dramatic rise in rates of offences initially recorded as homicide seen over that time from around 300 or 400 a year in the 1950s to more than 700 a year in the late 1990s. The recent Home Office Statistical Bulletin (Simmons & Dodd, 2003) shows a continuing rise in this trend with 1048 deaths initially attributed to homicide in 2002/2003, although these figures are based on date of notification and thus can include deaths that actually took place in earlier years.

Dr Salib's paper appears to use data on death registrations from the ONS where there has been a conviction for murder or for manslaughter. However, the ONS assigns a temporary ICD–9 code for cause of death for deaths where death was violent, unnatural or suspicious or pending the outcome of inquests and legal proceedings, which are of course often prolonged. The ONS site itself states that it is difficult to present accurate statistics on number of homicides using death registrations, which is what Dr Salib has seemingly attempted to do.

As psychiatry is faced with a Government currently determined to medicalise as far as possible the growing problem of violence in our society, it is essential that psychiatric journals present statistics on this subject in a meaningful fashion. Dr Salib's paper, although not specifically about trends in homicide over time, presents misleading data on this subject, which are neither helpful nor informative to the wider debate on violence in society.