Patterns and predictors of re-admission to hospital with self-poisoning in Scotland

Objectives: To identify factors influencing hospital re-admission with self-poisoning. Study design: Retrospective cohort follow-up study using national linked hospital discharge data. Methods: All Scottish adult hospital episodes with self-poisoning admissions were captured using NHS Scotland Information Services Division data, and first-time 'index' admissions between 1996 and 2002 were identified. Re-admission rate was defined as the proportion of index admissions who went on to have one or more further self-poisoning admissions within 2 years. The effects of various potential risk factors for re-admission were examined using logistic regression. Results: In total, 50,891 index admissions were identified; of these, 8278 patients were re-admitted. The 1-year re-admission rate was 12.2%. Older patients (>65 years) were least likely to be re-admitted [odds ratio (OR) 0.40, P < 0.01, compared with patients aged 15-24 years]. No differences were found between males and females. Previous psychiatric hospital admission was associated with an increased readmission rate (OR 2.85, P < 0.01), with a diagnosis of personality disorder associated with the highest rate of re-admission (OR 4.59, P < 0.01). Other factors predicting re-admission were: increased deprivation (quintile 3: OR 1.16, P < 0.01; quintile 5: OR 1.15, P < 0.01, compared with quintile 1); taking medicines for chronic disease, drug dependency (OR 1.6 and 1.19, P <= 0.02) or antidepressants (OR 1.11, P = 0.01) (compared with paracetamol); and co-ingestion of three or more agents (OR 1.37, P < 0.01). Conclusion: Younger age, higher deprivation, ingestion of certain drug groups or multiple drug types, and prior psychiatric hospital admission are all risk factors for re-admission with self-poisoning. Personality disorder carried the greatest risk of re-admission. These findings may provide a basis to develop policies to reduce re-admission rates in the future. (C) 2008 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.