Childhood abuse and adult violence in homeless people

A recent theory proposed to explain violence towards others invokes the roles of childhood abuse, neuropsychiatric disorder and concomitant substance misuse; all impacting upon the individual perpetrator. If accurate, such a theory has potential implications for treatment strategies and harm reduction within the psychiatric context. With this in mind, we conducted a retrospective analysis of 148 case notes from consecutive referrals to a psychiatric service for the homeless. The homeless are at particular risk of cognitive impairment, and are likely to report multiple aversive experiences in childhood. Hence, we examined the relationship between reports of childhood abuse and violence in adult life. Our sample comprised 123 male and 25 female subjects, of mean age 37.2 ± 12.3 years (range, 18–77 years). Of the primary psychiatric diagnoses recorded, affective disorders were most common (48.6%), followed by schizophrenia (24.3%) and personality disorders (15.5%). Alcohol and substance misuse were ubiquitous (38.5 and 23.6%, respectively). A total of 27.7% of subjects had reported abuse in childhood (12.8% physical, 6.1% sexual, 8.8% both), 33.8% had significant history of violence towards others (recorded under ‘forensic history’), while 40.5% had self-harmed. We performed χ tests, in order to examine whether the frequency of reported violence differed between subjects with and without a history of childhood abuse (all categories, because the samples are small). We found that there was no difference in violence towards others according to earlier abuse. However, abused and non-abused groups differed in terms of their reported self-harm, a finding that approached statistical significance ( = 2.2, P = 0.07, one-tailed). In other words, there was a trend for those who were abused as children to self-harm in adulthood. There are weaknesses in retrospective analyses of data acquired from clinical case notes, where the former were recorded for other purposes. Also, the χ 1 148 2 , ( ) nature and veracity of ‘abuse’ and ‘violence’ are likely to be heterogeneous across subjects. Nevertheless, our sample is one in which aversive experiences might be expected to be particularly frequent, hence providing an enriched sample in which to test the hypothesis proposed. It may be that those patients who gave rise to Pincus’ hypothesis are radically different from those encountered by our service at the Sheffield Care Trust. The former may represent an atypical group, both in terms of the extremity of their violence and the severity of the sadism to which they were exposed as children. For instance, in contrast to Pincus’ cases, none of our sample had killed anyone. Nevertheless, these potential differences are important to acknowledge when attempting to predict violent conduct in the British context, in non-forensic, community samples. Our data suggest that in this setting, even among those who are socially excluded, people who have been abused are of greater risk to themselves than others.