A secondary objective of the study was to ascertain whether two common conditions in inpatient populations (contact abuse experiences and deliberate self-poisoning) occurred together more often than predicted on the basis of chance alone. This was done after controlling for gender, age, ethnicity, marital status, social class, and geographical locality, which are all factors known to be associated statistically with a history of abuse. The design was a matched (1:1) case-control study set in a district general hospital in England. The subjects were 36 women admitted for treatment for deliberate self-poisoning. These women were matched with the next non-overdose admissions to the same hospital on the following six variables: sex, age, ethnicity, social class, marital status, and geographical locality. The main outcome measures used were modified versions of standardized self-report questionnaires of sexual, physical, and psychological abuse, together with measures of parenting style and general psychopathology. The study found that women who had taken an overdose were more likely to have been sexually abused and somewhat more likely to have been psychologically but not physically abused. They also had higher measures of psychopathology and were more likely to have been abused at a younger age, exposed to the "affectionless control" style of parenting by their mothers, and to have harmed themselves in other ways. The authors advise that women treated in hospitals for self-poisoning should be assessed for abuse experiences, and appropriate treatment should be provided for those with significant experiences and symptoms of abuse. 3 tables and 30 references