Safety issues in psychiatry are, arguably, still not well addressed. This is a matter for concern. The video should be of value to anyone, clinician or researcher, who visits patients at home.

The tape runs for 12½ minutes and is divided into four sub-sections: preparation, the interview process itself, what to do if something goes wrong and post-incident analysis. There are sequences with research workers describing their own experiences, with comments by Professor Kendrick and the reconstruction of some adverse incidents. It is clear, concise and based on common sense principles. The makers have avoided over-emphasising the issue of potential interpersonal hostility or aggression, but not played down the need to maintain awareness of it as a possibility.

Some areas could have been emphasised further. The need to obtain a comprehensive history, especially of previous violence, might have been given greater importance. It is unlikely, as suggested by one researcher, that confidentiality might prevent it. Access to the patient presumes ethical approval which should have dealt with this. If clinicians agree that patients can be approached to participate in a study but will not share information then the individual concerned should be excluded. Initial contact with the patient by letter or through liaison with a professional already involved, and a joint visit, might also have been suggested. Education of research workers, particularly those with limited experience of patients suffering from mental illness, is also vital.

I would thoroughly recommend this short training film to all those supervising community researchers or undertaking the work themselves. It could also be used to teach medical students, trainees in psychiatry and general practice, student nurses, social workers and other community staff. The principles outlined apply equally well to research and clinical practice. The makers do not shy away from saying that one should cancel a visit if there is cause for concern and that safety is more important than research results. They also add that it may be necessary to leave a situation if it becomes dangerous in keeping with the principle that intervention can, at times, increase, rather than decrease, risk.