We read the short report by Bennewith et al (Reference Bennewith, Gunnell and Kapur2005) with interest. The authors attempted to address one of the objectives of the National Suicide Prevention Strategy for England, restricting access to means of suicide (Department of Health, 2002).

The authors found 10 cases (6%) of ‘potentially preventable’ suicide by hanging in controlled environments such as hospitals and prisons, among 162 cases of a randomly selected sample of suicide by hanging from a wide geographical area in England.

We would like to make the following comments. The report made no reference to the proportion of older victims in the randomly selected sample. The mean age given in the report (41 years) is almost the same as that for all people over the age of 16 years in England and Wales who hanged themselves in the same year as the study (2001) and over the past 23 years. These cases include, on average, 16% over the age of 64 years. This means that the study sample of 162 contained at least 25 victims over the age of 64, a sizeable older element that was not referred to in the report.

This is important and ought to have been clarified particularly in relation to the deaths that occurred in hospital when the victims were found seated (4.7%), kneeling (7.4%), lying (8.7%) or partially suspended (3.4%) and to individuals who were found alive (4.3%). However, this does not apply to hanging in prison where victims had an estimated mean age of 28 years (Reference Shaw, Baker and HuntShaw et al, 2004).

We calculate the expected annual rate of ‘potentially preventable suicides by hanging’ within institutions (controlled environment) in England and Wales to be about 110 cases of a total of 1300 expected annual suicides by hanging: 86 in prison (Reference Shaw, Baker and HuntShaw et al, 2004) and 24 (about a third of 71 hanging incidents by psychiatric in-patients) in hospital (Department of Health, 1999). ‘Potentially preventable suicide by hanging’ in controlled environments involving prisoners represents 5% of all suicide by hanging in England and Wales and 2% in the case of psychiatric in-patients. This is remarkably similar to the 6% in the report of Bennewith et al.

It would be of real interest, and certainly of practical value, if future studies specifically investigated suicide by hanging within controlled environments such as hospitals and prisons using an appropriately selected sample over a period of time (e.g. 220 incidents expected over 2 years in England and Wales, based on current figures). This would provide a study with acceptable power and some inferential value compared with the modest 10 cases reported by Bennewith et al.