EDITOR—Carrying out a risk assessment can be risky. Doctors should ask patients if they are carrying weapons only if they can safely cope with the immediate production of a weapon. Doctors must consider their own safety and that of the patient, relatives, and colleagues. Sanders et al recommended inquiry into the full range of aggressive ideation but issued no warning about how dangerous this can be.1 They found that one in 20 patients admitted to a psychiatric hospital in Middlesbrough regularly carried weapons. Some patients will produce their weapon on inquiry, and a few may be prepared to use it.

I have been treated to demonstrations of knives, scissors, a machete, and a (replica) gun. In most cases I had arranged for others to be present before asking about weapons, and the situations were managed safely. Not all my colleagues have been so fortunate. The staff of psychiatric wards usually have training and experience in the management of violence. They can also respond quickly to an emergency involving a weapon on the ward. Doctors and nurses who see patients at home or in clinics rarely have such support available unless they have made specific arrangements in advance. It is therefore advisable to organise support before asking about weapons. This support could be a visit with a colleague, or a safer venue such as the ward in preference to a clinic. With good back up and an understanding of the patient's mental state, a handover of most weapons can be instigated with minimal risk to all concerned.

The method employed by Sanders et al compared a semistructured interview about risks with case note records of aggressive ideation. It is not clear if these were medical, nursing, or multidisciplinary records. A similar study in a psychiatric hospital by Harwood and Yeomans showed that a risk interview found more evidence of risk of violence than either medical or nursing records alone.2 If the medical and nursing records were combined, however, the level of routine risk recording was better and approached the quality of the systematic interview. This paper also found that risk assessments were carried out more frequently when there was a regularly reviewed standard for risk assessment audits in place and an effective audit system to back it up.

EDITOR—Sanders et al state that most violence is carried out by mentally well people.1 They then describe an uncontrolled study, in which they report the result of non-randomised interviews of patients admitted to a general psychiatric unit. They report seemingly high rates of violent thoughts, previous offending, and weapon ownership with no reference to a control group or baseline levels in the unit's catchment area.

In the absence of a control group matched for age, sex, social deprivation, and other important non-psychiatric risk factors with regard to violence, the results are at best meaningless and at worst highly stigmatising of psychiatric patients. Many mentally well people have violent thoughts, most own a weapon, and many will have carried a weapon. It is well recognised that psychiatric variables have little bearing with regard to risk of violence. In fact, a diagnosis of a psychotic illness can be seen as a protective factor with regard to repeat offending.2 This work may seem harmless, but poorly constructed and uncontrolled research such as this only helps to further the inaccurate stigmatisation of mentally ill people as potential killers, despite the lack of any evidence base to support this.

Sanders et al conclude that it is important for doctors to inquire systematically about the full range of aggressive ideation in patients admitted to psychiatric hospitals. Their paper provides no evidence as to why such patients should be asked any more than members of the general public. Making such a statement inaccurately implies a general increase in dangerousness associated with a psychiatric label. Had they used evening attendees at a local accident and emergency unit as controls, the “spin” of their article would have been quite different.