Intended for healthcare professionals


Effect of general hospital management on repeat episodes of deliberate self poisoning: cohort study

BMJ 2002; 325 doi: (Published 19 October 2002) Cite this as: BMJ 2002;325:866
  1. Navneet Kapur, lecturer (nav.kapur{at},
  2. Allan House, professorb,
  3. Kath Dodgson, research assistanta,
  4. Chris May, research assistanta,
  5. Francis Creed, professora
  1. aDepartment of Psychiatry and Behavioural Sciences, Manchester Royal Infirmary, Manchester M13 9WL
  2. bAcademic Unit of Psychiatry and Behavioural Sciences, University of Leeds, Leeds LS2 9JT
  1. Correspondence to: N Kapur, 7th Floor Williamson Building, University of Manchester, Manchester M13 9PL
  • Accepted 13 March 2002

Provision of services in the United Kingdom for patients who deliberately poison themselves is variable, and many patients leave hospital without adequate assessments.1 This may reflect the equivocal research evidence on the effectiveness of interventions.2 In this cohort study, we aimed to investigate whether aspects of routine hospital management—such as admission, psychosocial assessment, and referral for follow up—had an impact on the repetition of deliberate self poisoning.

Participants, methods, and results

Over eight weeks, we prospectively identified patients aged over 16 years who attended six general hospitals (three teaching; three district) in north west England for deliberate self poisoning. We examined the notes in accident and emergency departments for all patients (regardless of presenting complaint) to ensure that we did not miss any episodes. We also looked at databases held in wards and emergency departments and copies of specialists' self poisoning assessments, and we retrospectively checked the patient administration system in each hospital. We collected information about patients' characteristics, clinical details, and the management of the current episode, including whether the patient had received a psychosocial assessment (as defined by the Royal College of Psychiatrists).3 We followed participants for 12 weeks after their index episode, because half of those who poison themselves again do so within this period.4 Follow up data were taken from case notes and databases by a researcher blind to the patient's initial management.

In total, 604 people deliberately poisoned themselves during the recruitment period; 88 (15%) of these poisoned themselves again within 12 weeks. Overall, 24/246 (10%) patients who had received a psychosocial assessment and 64/358 (18%) who had not received an assessment poisoned themselves again (table). Patients with certain risk factors for repetition—such as previous self poisoning, psychiatric history, and substance dependence—were more likely to receive a psychosocial assessment. We adjusted for these and other risk factors for repetition—such as sex, age, potential lethality of the episode, and self discharge from hospital—with a binary logistic regression model.5 We included psychosocial assessment in the model to investigate whether such assessment independently contributed to risk of repetition. After adjustment, patients who had not been assessed were still more likely to poison themselves again (adjusted odds ratio 2.3 (95% confidence interval 1.4 to 3.9), P<0.005).

Risk of repetition of self poisoning in 604 patients presenting to six general hospitals in north west England

View this table:


Patients who had received a psychosocial assessment after deliberately poisoning themselves were half as likely to poison themselves again as those who did not, even though the patients who were assessed might be regarded as being at higher risk. The association held when we adjusted for risk factors for repetition, although we did not adjust for all possible confounders. The better outcome in those who were assessed compared with those who were not assessed may have reflected the type of aftercare, but this study provides no evidence that referral to a specialist reduced repetition. Our results suggest that 12 patients need to receive a psychosocial assessment to prevent one patient poisoning himself or herself again. If we assume that 50% of patients are assessed currently, we might prevent 7000 repeat episodes of self poisoning by complying with existing guidelines and ensuring that all patients are properly assessed.3

Our results should be interpreted cautiously because they are restricted to one region of the United Kingdom and include a comparatively short follow up period. The hospitals were representative of the services generally available for self poisoning,1 however, and most patients who poison themselves repeat the act soon after the index episode.4 We did not identify repeat episodes in patients who did not present to hospital, and this is a potential source of bias.

Psychosocial assessments have a number of elements.5 Further research should identify and refine the components of psychosocial assessments that reduce the risk of repetition of self poisoning.


We thank Kevin Mackway Jones, Paul Strickland, Mike Brownlee, Bill Williams, Robin Ellis, and Jayne Cooper (director of Manchester and Salford Self Harm Project) for help with data collection and staff from the information departments at all centres for providing admission data. We also thank Iain W McGowan for his comments on the paper.

Contributors: NK and AH devised the study and wrote the initial draft of the paper. FHC advised on aspects of study design. KD and CM collected data and helped with aspects of study design. All authors contributed to the analysis and presentation of data and commented on drafts of the paper. NK is the guarantor for this study.


  • Funding CM and KD were supported in part by an educational grant from SB Pharmaceuticals during the course of this study.

  • Competing interests None declared.


View Abstract