Case-control study of GP attendance rates by suicide cases with or without a psychiatric history

Targets for reduction in suicide deaths have been set against a background of an increasing number of people committing suicide. It is often assumed that a reduction can be effected by increasing the detection in primary care of patients at risk. This presupposes that there are indicators that enable suicide risk to be detected reliably.

To compare the characteristics of those who commit suicide with an age- and sex-matched control group in terms of level of general practitioner attendance, diagnosis and pharmacological treatment of mental illness, and to compare those suicides with and without a psychiatric history in terms of general practitioner attendance and history of pharmacological treatment.

From a total of 48 deaths attributed to suicide and undetermined causes in the Forth Valley in 1993, general practice case notes were located for 41. Live controls were matched to index cases by age, sex and practice. Information on consultations, referrals to secondary care, medication and diagnoses in the previous 10 years was extracted from general practice and, for suicides, psychiatric case notes.

Over the 10-year period, suicide patients attended their general practitioner at a higher level than control subjects. However, the number of suicide patients who attended their general practitioner in the month before their death did not differ in comparison with control subjects over a similar period. Suicide cases, in comparison with control subjects, were more likely to have received a psychiatric diagnosis from their general practitioner, been prescribed psychotropic medication and received referral to specialist mental health services. Those suicide patients with a psychiatric history had a significantly higher number of general practitioner consultations than those without a psychiatric history in four out of the five years preceding death. Those suicide patients without a psychiatric history did not differ significantly from control subjects on any of the variables assessed.

For those people committing suicide who do not have a psychiatric history and whose consultation patterns do not differ from the norm, it is difficult to suggest how general practitioners might improve their detection of relevant suicidal risk factors. For those patients with a psychiatric history who commit suicide, until we have more detailed information regarding the specific content of general practitioner's consultations before death and how these differed from other consultations of the deceased, then it is premature to assume that general practitioners are failing to identify indicators of impending suicide.