Little is known about the precursors of suicide risk among primary-care patients. This study aimed to examine suicide risk in relation to patterns of clinical consultation, psychotropic drug prescribing, and psychiatric diagnoses.

Nested case-control study in the Clinical Practice Research Datalink (CPRD), England. Patients aged ⩾16 years who died by suicide during 2002–2011 (N = 2384) were matched on gender, age and practice with up to 20 living control patients (N = 46 899).

Risk was raised among non-consulting patients, and increased sharply with rising number of consultations in the preceding year [⩾12 consultations v. 1: unadjusted odds ratio (OR) 6.0, 95% confidence interval (CI) 4.9–7.3]. Markedly elevated risk was also associated with the prescribing of multiple psychotropic medication types (⩾5 types v. 0: OR 62.6, CI 44.3–88.4) and with having several psychiatric diagnoses (⩾4 diagnoses v. 0: OR 31.1, CI 19.3–50.1). Risk was also raised among patients living in more socially deprived localities. The confounding effect of multiple psychotropic drug types largely accounted for the rising risk gradient observed with increasing consultation frequency.

A greater proportion of patients with several psychiatric diagnoses, those prescribed multiple psychotropic medication types, and those who consult at very high frequency might be considered for referral to mental health services by their general practitioners. Non-consulters are also at increased risk, which suggests that conventional models of primary care may not be effective in meeting the needs of all people in the community experiencing major psychosocial difficulties.

Early environmental events may be relevant to the etiology of schizophrenia. Among such events, interest has focused especially on obstetric complications (OCs).

Aims of the study were to compare the incidence of OCs in patients, siblings and normal controls and to examine the relationship between OCs and later schizophrenia.

One hundred and thirteen patients with schizophrenia were recruited, as were 140 patients’ siblings and 113 controls without schizophrenia. The OCs history of patients, their sibs and controls was obtained through interviews with patients’ and controls’ mothers.

The results highlighted that more patients than sibs had at least one definite OC and a higher mean number of OCs; more patients had premature rupture of membranes, threatened abortion and a labor of more than 36 h.

Our data provide some evidence for a link between OCs and later schizophrenia. Furthermore, this study highlights how OCs, which may cause fetal distress through a hypoxic-ischemic mechanism, could increase the risk of schizophrenia interacting with genetic susceptibility.