Poisons admissions in Edinburgh 1981-2001: agent trends and predictors of hospital readmissions.

Self-poisoning is a major public health problem. This study describes patterns of admissions and readmissions from self-poisoning to the Royal Infirmary of Edinburgh from 1981 to 2001. A database on hospital discharges with a diagnosis (ICD-9/10) of poisoning between 1981 and 2001 was used. Annual admissions were described for seven main drug categories, and proportions of patients readmitted within 1-5 years from first admission, were computed for each category. Cox proportional hazards regression was used to evaluate prognostic factors for readmission risk over 1981-2001. For both sexes, admissions increased from the early to mid 1990s, and declined thereafter. The proportion readmitted varied with the drug taken at first admission, from 11.9% (95% CI: 10.8-13%) for non-opiate analgesics, to 17.6% (16.5-18.7%) for benzodiazepines. Deprivation was positively related to readmission risk after first admissions with paracetamol (P < 0.001) and benzodiazepines (P < 0.001). Timing of first admissions involving paracetamol (P < 0.01), benzodiazepines (P < 0.001), antidepressants (P < 0.001), non-opiate analgesics (P < 0.001), and opiates (P < 0.05), was inversely associated with readmission risk. In patients admitted for drug overdose, readmission risk is influenced by type of drug taken at first admission. Information on drug type used in self-poisoning may assist in identifying patients at risk for future events, and in reducing hospital readmissions.