Commentary: Homicide-Suicide in Older Adults—Cultural and Contextual Perspectives

Rates of Homicide-Suicide in Different Countries

In a review of the literature from 10 nations, Coid¹² found the incidence of murder-suicide to be 0.2 to 0.3 per 100,000. Similarly, Marzuk et al.¹³ estimated the homicide-suicide rates in the United States to be 0.2 to 0.3 per 100,000 per year. In a comprehensive review of the current literature, Eliason¹⁴ examined the data on murder-suicide of Marzuk and colleagues,¹³ and found an incidence of murder-suicide under 0.001 percent. Specific to country, Jena et al.⁴ found an annual rate of homicide-suicide of 1.0 per 100,000 in Pretoria, South Africa, between 1997 and 2001, while Roberts et al.⁵ reported an annual homicide-suicide rate of 0.89 per 100,000 in Durban, South Africa, between 2000 and 2001. Hansen⁶ reported a rate of homicide-suicide of 1.33 per 100,000 in Greenland in the mid 1970s at a time when the total homicide rate was 15.6 per 100,000, and the suicide rate was an extraordinary 45 per 100,000 per year. Large et al.² reported a rate of 0.313 per 100,000 within the United States. While these rates show that this incident happens around the world, they also demonstrate the rarity of homicide-suicide and the complications involved in prevention. Thus, more research should be conducted to gain a better understanding of the phenomenon.

Similar Characteristics of Homicide-Suicide

Thorough investigation shows commonalities in the details of the spousal/couple homicide-suicide. These similarities (or risk factors) usually include the relationship between homicide victims and homicide-suicide perpetrators (usually intimate), the sex of the homicide victim (usually female), the sex of the homicide-suicide perpetrator (usually male), the age of the victim (usually younger than the perpetrator), the use of a weapon (usually a gun), and a history of mental illness in the homicide-suicide perpetrator (most often a male who is depressed). The resemblances are universal and cut across national lines. In Scandinavia,³ South Africa,^4,5 China,⁷ and the United States,¹⁵ for example, most homicide-suicide perpetrators are male, with most of their victims being female intimate partners who are killed violently by them with weapons.

Dissimilar Characteristics of Homicide-Suicide

While these commonalities can be helpful in identifying risk factors, further study of the broader phenomenon of homicide-suicide renders more variation and difference within and between cultures and age groups, complicating attempts at accurate targeted prevention. The sample in the study by Bourget et al.¹ consisted of 93 percent male perpetrators, with all perpetrators older than 65 years. In this group, there were many Caucasians, a low prevalence of domestic violence history, and a high incidence of physical illness in the victims.¹ Other counties, cultures, and age groups display different trends. In the United States, there was a history of domestic violence in most homicide-suicides involving intimate couples; however, most were not elderly couples.¹⁴ In China,⁷ most of the incidents were related to spousal conflict or were completed for altruistic reasons. The main methods of killing were slashing with a weapon and carbon monoxide poisoning by charcoal burning in a closed space. Homicide-suicide trends in the three counties identified earlier (Canada, the United States, and China) show that there are variations in both country and age, which also demonstrates the dynamic complexity of this phenomenon.

Cultural, Racial, and Ethnic Differences in Homicide-Suicide

Furthering these complications are ethnic differences related to homicide-suicide in South Africa and in the United States. Regarding patterns of homicide-suicide relationships, Jena et al.⁴ found two profiles of typical perpetrators in Pretoria: a younger, single, black male shooting his girlfriend and himself at home and an older, married, Caucasian male shooting his wife and himself at home. In Durban, Roberts et al.⁵ reported that 91 percent of the perpetrators and 87 percent of the victims were black, which proportionately represents the population.

Gender Differences in Homicide-Suicide

Differences based on the sex of the perpetrator have also been studied in the United States. Infant homicide followed by a suicide (filicide-suicide, which is largely preventable by pre-empting postpartum depression¹⁸) is one of two major homicide-suicide patterns (the second being intimate partner homicide-suicide¹⁹). Logan et al.²⁰ found that more than half of filicide-suicides (51.5%) in their sample were committed by females.

Motivational Differences Between Canada and the United States

In almost every country and culture, males perpetrate homicide-suicide. The acts are most often committed by men, the means are usually violent (involving weapons), and the victims are most frequently women. In a review, Scott³⁴ lists 13 suggested motivations for homicide-suicide, the first five of which are impending divorce, previous divorce, departure or perceived loss of a nonmarital partner, jealousy, and retaliation. This finding suggests pervasive and pernicious patriarchy, sometimes called male entitlement dysfunction,³⁵ which should be urgently addressed. Considering that domestic violence is a major motive for homicide-suicide in the United States,¹⁴ it is surprising that a history of domestic violence among the older couples was not found to be more prevalent by Bourget et al.¹ This discrepancy could be attributable to the differences in national culture. The pattern of homicides in the United States is much different from that in Canada, with rates in the United States being about three times higher.³⁶

Complications in Studying Homicide-Suicide

Ostensibly, there is a broad range of research that has been conducted in this area; however, this endeavor is fraught with complications and potential inaccuracies. Bourget et al.¹ identified two limitations of their study: missing data and a small sample size. There are further limitations to the study of this phenomenon. The Institute of Medicine Suicide Prevention report²¹ noted that coroners’ data in the United States are unreliable because of regional differences in requirements for the position of coroner, the definition and classification of suicide, training and background, investigation of cases, and the quality of data management. IOM gave various examples and stated: In the United States, the qualifications range from simply having an interest in the job (e.g., Indiana) to specialized training in forensic pathology (e.g., Oklahoma). Medico-legal officials may be elected, appointed or serve ex-officio (e.g., elected county sheriffs). Investigations may be centralized within a state (e.g., Rhode Island) or organized by each county (e.g., Utah). Each of these factors affects the nature, extent, and quality of the investigation and the classification of deaths as suicide [Ref. 21, pp 380–1].

These nuances can lead to inadvertently erroneous statements and confusion among researchers.

Religion and stigma leading to reporting bias can confound the study of both suicide alone and homicide-suicide. For instance, predominantly Catholic countries, in which the religion-related stigma on suicide is severe, are less likely to make explicit attribution in the public record of deaths due to suicide.^37,38 The reported suicide rates in countries that are predominantly Muslim may also be influenced by this bias.³⁹ It has been postulated that most, if not all, of the cases of death of undetermined cause are actually suicides.²¹

The rarity of suicide, as recognized by Bourget et al.,¹ further complicates precise inquiry. The IOM committee also noted that suicide is a rare event. Despite these higher rates, the reality is that suicide and homicide have low base rates: 0.00011 and 0.000062, respectively. Thus, if they are to yield significant results, studies must include large population samples. IOM asserted this necessity: Large sample sizes are required to provide statistical power for studies of events with a low base rate. For example, to determine the overall incidence rate of suicide in a general population within plus or minus five per 100,000 with 90 percent confidence, would require about 100,000 participants. Many statistical approaches exist that can be effectively applied to the field of suicide [Ref. 21, p 411].

As the homicide-suicide rates are even lower than individual suicide rates, we believe that attempting to base prevention efforts on the findings of these limited samples is problematic and typical of the current state of suicide prevention in Western countries.

Prevention of Suicide in the Elderly

In 1990, the United States had a suicide rate of 24.9 per 100,000 among men aged 75 to 84. In 1998, the rate had risen to 42.0 per 100,000. Although older individuals comprise approximately 10 percent of the U.S. population, they account for 20 percent of completed suicides.^40,41 Because the rates of suicide and homicide in the elderly are so low, it is difficult to develop a profile of the perpetrators. Given the apparent protective factors that are in operation, profiling with the current limited information would be likely to yield frequent false positives.²¹ However, Bourget et al.¹ provide information to build a foundation for effective prevention. These Canadian authors suggest that terminal illness in a spouse (the usual victim) and depression in the other spouse (the usual perpetrator) may be risk factors for homicide-suicide in an older couple, proposing that it would be prudent to screen for such disorders to prevent suicide and homicide-suicide. IOM stated: Untreated or undertreated pain, anticipatory anxiety regarding the progression of medical illness, fear of dependence, and fear of burdening the family are the major contributing factors in the suicidality of elderly with medical illness [Ref. 21, p 44].

Primary care physicians saw 70 percent or more of elderly suicide victims within 1 month of their deaths.^42–44 It makes sense that if in Canada the victims of homicide were gravely ill,¹ individuals at risk could be identified in primary care offices. Accordingly, the problem of prevention of older-partner domestic homicide-suicide¹ could be approached in the same way as prevention of elderly suicide. Foley⁴⁵ and Hendin⁴⁶ found that elderly and cancer patients who once wanted to die found a stronger will to live when their pain or depression was adequately treated and managed. Specifically, the Prevention of Suicide in Primary Care Elderly (PROSPECT): Collaborative Trial⁴⁷ proposed a way of identifying at-risk perpetrators of suicide and providing them with protective factors that keep their risk factors from becoming predictive of bad outcomes. PROSPECT⁴⁸ found that patients in the intervention group had seven percent lower suicidal ideation than did those in conventional treatment. Further, 12 percent fewer patients expressed feelings of hopelessness than those in conventional treatment. Bruce et al.⁴⁷ found that patients in PROSPECT experienced a faster rate of decline of suicidal ideation than did those in regular treatment. Alexopoulos et al.⁴⁹ similarly found that those in PROSPECT were more likely to receive psychotherapy or antidepressants and to have a remission of depression than were those in the treatment-as-usual group. The effectiveness of this approach suggests that the same strategy could be used to prevent homicide-suicide in elderly couples.

Furthermore, although some homicide-suicide patterns hold across cultures, there are culturally specific patterns as well (e.g., China's homicide-suicide pattern of chopping or charcoal burning). Prevention research and intervention development must further ascertain and study these differences. As in most cases in mental and physical health prevention and intervention,¹⁸ treatment of homicide-suicide must seek to be culturally responsive to be effective. The clinician who disregards the patient's cultural origin may lose the patient. If prevention and intervention strategies are not sensitive to the understanding and worldview of patients, then these strategies may not be as effective or may even fail completely.