294 Suicide and Life-Threatening Behavior 39(3) June 2009  2009 The American Association of Suicidology The Epidemiology of Homicide followed by Suicide: A Systematic and Quantitative Review Matthew Large, BSc (Med), MB, BS, FRANZCP, Glen Smith, BSc (Hons) MB, BS, and Olav Nielssen, MB, BS, M Crim, FRANZCP This systematic review of population based studies of homicide followed by suicide was conducted to examine the associations between rates of homicide- suicide, rates of other homicides and rates of suicide. The review analysed 64 samples, including the case of an outlier (Greenland) that were reported in 49 studies. There was a significant association between the rates of homicide-suicide and those of other homicides in studies from the U.S.A. Outside the U.S.A. there was no clear association between homicide-suicide and other homicide but there was modest but significant association between rates of suicide and homicide- suicide. Homicide-suicide appears to be closer in epidemiological terms to homi- cide than suicide in regions with high rates of homicide and measures to reduce homicide in these regions may also reduce homicide-suicide. A proportion of people who commit a homi- (Stack, 1997). A small proportion of cases in- volve older people with physical illness or an cide also commit suicide shortly afterwards, in a phenomenon known as homicide-suicide offender with a psychotic illness (Felthous & Hempel, 1995). (H-S). Most H-S offenders are men who kill their partners and sometimes also their chil- H-S has elements of both homicide and suicide, although it is widely believed dren, and a close relationship with the homi- cide victim predicts subsequent suicide risk that homicide-suicide has more in common with suicide than homicide (Coid, 1983; Ro- senbaum, 1990; Saleva, Putkonen, Kiviruusu, Matthew Large is with Mental Health & Lo ¨ nnqvist, 2007). After reviewing 17 avail- Services, The Prince of Wales Hospital, Randwick, able studies Coid (1983) asserted that H-S Sydney, NSW, Australia and the School of Psychi- atry, University of New South Wales, Sydney, rates were unrelated to rates of total homi- NSW, Australia; Glen Smith is with the Northern cide. Coid proposed two epidemiological Sydney Central Coast Area Health Service, Mac- laws regarding H-S: (i) that the higher the quarie Hospital, North Ryde, NSW, Australia; rate of homicide in a population, the lower Olav Nielssen is with the Clinical Research Unit the proportion of homicide offenders will for Anxiety Disorders, School of Psychiatry, UNSW at St Vincent’s Hospital, Sydney, NSW, subsequently suicide, and (ii) that the rate of Australia, and the Discipline of Psychological Med- H-S is approximately the same in different icine, University of Sydney, NSW, Australia. countries despite differences in the overall The authors would like to thank Professor rates of homicide. Given the limited data E. Fuller Torrey of the Stanley Medical Research available at the time, the failure to include Institute Maryland, USA for his comments on the manuscript and Professor Earl Hunt of the Uni- studies such as that of Boudouris (1974) and versity of Washington for his comments on the the lack of a statistical analysis to support the distribution of homicide-suicide rates. conclusions, Coid’s use of the term “epidemi- Address correspondence to Dr Large, The ological laws” may have been premature. Euroa Centre, Mental Health Services, The Prince Nevertheless the existence of these of Wales Hospital, Barker St, Radwick, NSW, Australia, 2031; E-mail: mmbl@bigpond.com laws was supported by the authors of two