Comorbid Depression and Substance Abuse in Domestic Homicide: Missed Opportunities in the Assessment and Management of Mental Illness in Perpetrators

Abstract

Almost 40% of murdered women are victims of domestic homicide across the world. However, research has yet to examine comorbid depression and substance abuse in domestic homicide perpetrators, despite comorbid mental health conditions being associated with homicide in the general population. A retrospective case analysis approach was performed using domestic homicide cases that had been reviewed by the Domestic Violence Death Review Committee in Ontario, Canada. Group comparisons were made by compiling cases into groups based on perpetrators with depression and/or substance abuse: a no history of depression and/or substance use group, depression only group, substance abuse only group, and comorbid depression and substance abuse group. Statistical analyses compared groups on number and types of risk factors and service provider contacts. A major finding of the current study was the resounding lack of mental health care and batterer intervention program engagement across groups. Results also indicated that perpetrators with comorbid depression and substance abuse have an elevated number of risk factors for domestic homicide and elevated number of service provider contacts. Furthermore, results indicated that perpetrators with comorbid depression and substance abuse had an increased likelihood of having engaged in hostage-taking behavior and increased likelihood of having contact with mental health and health care providers. The study demonstrates the necessity for future research into the barriers associated with help-seeking by perpetrators, family and friends, as well as the barriers to agency referral and to mental health agencies providing service to perpetrators. It also highlights the need for service providers to take multiple mental health conditions into account when working with perpetrators of domestic homicide. Overall, this study underscores the importance of mental health and domestic violence training for service providers in different sectors. Moreover, it emphasizes the necessity of collaboration among service providers to address both violence-specific and mental health-specific concerns in perpetrators of domestic violence.

Keywords

domestic homicide perpetrator mental health risk factor service provider

Literature Review

Domestic violence, also referred to as intimate partner violence, is defined by the World Health Organization (WHO; 2016) as “behaviour by an intimate partner or ex-partner that causes physical, sexual or psychological harm, including physical aggression, sexual coercion, psychological abuse, and controlling behaviours.” Statistics indicate that domestic violence is perpetrated by more men than women and affects upward of 35% of women worldwide (WHO, 2016). In the extreme, domestic violence can escalate into domestic homicide, whereby the current or former intimate partner kills the victim (Jaffe, Dawson, & Campbell, 2013). Almost four in 10 murdered women are victims of domestic homicide (Boyce & Cotter, 2013). In Canada, domestic homicides account for 20% of all solved homicides (Boyce & Cotter, 2013).

Male Depression, Substance Abuse, and Domestic Violence

At the forefront of the domestic violence literature is the examination of victim mental health due to the acute and chronic effects of domestic violence on well-being (Ferrari et al., 2016; Goodman, Fauci, Sullivan, DiGiovanni, & Wilson, 2016; Knight & Hester, 2016). However, a study by Sesar, Šimić, and Dodaj (2015) highlights the importance of examining perpetrator mental health in addition to victim mental health. After conducting their review of the literature, the authors concluded that research findings concerning perpetrator mental health are “insufficient” (Sesar et al., 2015). This conclusion emphasizes the need for investigations of perpetrator mental health because psychological disorders have been implicated as both predictors and outcomes of domestic violence in the literature (Jones, Hughes, & Unterstaller, 2001; Mason & O’Rinn, 2014; Mitchell & Anglin, 2009). Despite the scarcity of research, mental health issues including anxiety, depression, and substance use disorders have been found in perpetrators of domestic violence (Dinwiddie, 1992; Okuda et al., 2015; Shorey, Febres, Brasfield, & Stuart, 2012).

The present study focused on depression and substance abuse because they are among the most commonly documented perpetrator factors in the domestic homicide literature. Depression has been identified as a significant risk factor in research on domestic homicide (Hirose, 1979; Rosenbaum, 1990; Rosenbaum & Bennett, 1986; Schlesinger, 2000). Researchers have found that more than half of domestic homicide perpetrators had a substance abuse problem (Dobash, Dobash, Cavanagh, & Lewis, 2004). Perpetrators who abuse alcohol were 8 times more likely to abuse their partners and twice as likely to murder their partner (Sharps, Campbell, Campbell, Gary, & Webster, 2001).

Comorbidity was first used in medicine by Feinstein (1970) to describe any “additional co-existing ailment.” The term is still used in today’s mental health field, as illustrated by its usage throughout the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association, 2013). However, within this field it is more narrowly defined as the simultaneous existence of two or more mental health concerns within an individual. One aspect of the mental health literature centralizes on investigations of comorbid mental health conditions and their association to violence (Corrigan & Watson, 2005; Van Dorn, Volavka, & Johnson, 2012). In a U.S. sample of over 5,000 people, researchers found that individuals with comorbid mental health conditions were more likely to have committed violent acts in the past 12 months than individuals without comorbid mental health conditions (Corrigan & Watson, 2005). Importantly, this study also adjusted for demographic factors and found that psychiatric diagnosis was still a significant predictor of violence (Corrigan & Watson, 2005). In another study by Van Dorn et al. (2012), within their sample of over 34,000 adults, researchers found a strong relationship between individuals with co-occurring substance abuse and another mental illness, and perpetration of violence. However, both of these studies used self-report measures to examine violence perpetration which may not have fully captured actual perpetration. Despite this limitation, the findings of these studies suggest that comorbid mental health conditions are essential to consider in domestic violence, due to the association of comorbid mental disorders to violence.

The association between comorbid mental health conditions and homicide has also been investigated in the literature. Fazel, Gulati, Linsell, Geddes, and Grann (2009) reported that individuals with comorbid substance abuse and other mental illnesses were at higher risk of committing homicide than individuals without comorbid mental health concerns. A study in the United Kingdom, which examined homicides that were committed by individuals in contact with mental health services between 1999 and 2003, found that 52% of their sample had at least one secondary diagnosis (Swinson & Shaw, 2007). Both of these studies provide further support for examining comorbid mental health conditions in domestic homicide, because comorbid mental health disorders have a relationship to homicidal behavior.

Risk Factors and Service Providers

In a major multisite study of domestic homicide cases, there was an indication of consistent risk factors which suggest the predictability of these murders (J. C. Campbell et al., 2003). The authors identified risk factors such as access to weapons, separation from partner, and prior domestic violence in the relationship which were associated with increased risk of domestic homicide (J. C. Campbell et al., 2003). Because risk factors are present before a murder occurs, these authors stressed that the identification of such factors could be used to prevent deaths (J. C. Campbell et al., 2003). A recent development in the examination of risk factors has been the formation of domestic violence death review (DVDR) teams globally (Dawson, 2017). DVDR teams have been examining cases of domestic homicide to deduce trends and risk factors that can inform domestic violence and homicide prevention techniques (Dawson, Jaffe, Campbell, Lucas, & Kerr, 2017). Risk factors are also important to research in domestic violence and homicide because they help to inform risk assessment strategies. Risk assessments are used to assess the likelihood of domestic violence occurring again (recidivism) and assess the severity of future domestic violence (lethality) (Messing & Thaller, 2013).

Numerous service providers can come into contact with perpetrators of domestic violence and each has their own mandate on how to effectively assess and deal with violence through the usage of structured or unstructured risk assessment, risk management, and safety planning techniques. Research into the mental health correlates of domestic homicide aims to create comprehensive and effective assessment and management protocols for all service providers working with perpetrators of domestic violence. In turn, these protocols can assist in the reduction of harm and of deathly outcomes in domestic violence. Research into the mental health correlates of domestic homicide can also clarify what interventions may be needed in cases of comorbidity, as individuals with multiple mental health concerns may have more risk factors and contact with more service providers due to the complexity of their mental health status.

The current study was a retrospective examination of domestic homicide cases and sought to elucidate risk factors and service provider contacts in cases involving perpetrators with comorbid depression and substance abuse. The study separated perpetrators from the cases into four groups based on mental health status: a comorbid depression and substance abuse group, depression only group, substance abuse only group and with no history of depression and/or substance use group. It was hypothesized that the comorbid group of perpetrators would have an increased presence of risk factors and service provider contacts due to their struggle with multiple mental health concerns. This research aims to aid in the development of risk assessment, risk management and safety planning protocols, and specifically for cases of comorbid depression and substance abuse, so that perpetrators of severe domestic violence can be managed more effectively and comprehensively.

Method

Data Collection

The current study used data from domestic homicide cases that have been reviewed by the Domestic Violence Death Review Committee (DVDRC) within the Office of the Chief Coroner in Ontario, Canada. The review is based on information gathered by the coroner and the investigating police officer. The information varies by case but most often includes interviews with friends and family as well as records from community professionals who have been involved within social service, health, and justice agencies. This information is gathered together and reviewed by a multidisciplinary committee (Dawson et al., 2017). This study focused on 267 domestic homicide cases between 2003 and 2015 (DVDRC, 2016). The DVDRC acts to compile information on perpetrators, including personal, familial, and service provider involvement, to understand and prevent the occurrence of domestic homicide in similar circumstances in the future (DVDRC, 2016). Out of the 267 cases reviewed by the DVDRC (2016), the researchers in the current study had access to the data from 219 of these cases. The dataset came from two preexisting coding forms used by the DVDRC to organize data from all cases.

DVDRC risk factor coding form

The first coding form, the DVDRC risk factor coding form, was created by the DVDRC to code information pertaining to each of the DVDRC’s 40 risk factors, including whether the risk factor was present (P), absent (A), or unknown (Unk) based on all compiled case reports. The risk factors were chosen based on existing literature on domestic violence as variables associated with repeat violence or lethal violence (see DVDRC, 2016, for risk factor descriptions). The focus of this study was depression and substance abuse.

Depression was inferred by the DVDRC when there were either documented cases where the perpetrators were diagnosed with depression, or if they had family, friends, or coworkers report that they displayed depressive symptoms. In a similar vein, substance abuse was inferred by the DVDRC according to formal diagnoses or inferred from friends, family or coworkers. The working definition for this factor with the DVDRC is

Within the past year, and regardless of whether or not the perpetrator received treatment, substance abuse that appeared to be characteristic of the perpetrator’s dependence on, and/or addiction to, the substance. An increase in the pattern of use and/or change of character or behaviour that is directly related to the alcohol and/or drug use can indicate excessive use by the perpetrator. For example, people described the perpetrator as constantly drunk or claim that they never saw him without a beer in his hand. This dependence on a particular substance may have impaired the perpetrator’s health or social functioning (e.g., overdose, job loss, arrest, etc.). Family, friend, and acquaintances have made comments that are indicative of annoyance or concern with a drinking or drug problem and attempts to convince the perpetrator to terminate his substance use. (DVDRC, 2016)

Views of friends and family are important in this process because many perpetrators did not seek help and did not have a diagnosis from a health care professional.

Each of the 40 risk factors has a similarly detailed definition found on publicly available annual reports on the website of the Chief Coroner. Consensus is required by the DVDRC member to include a risk factor for each case. The most common risk factors across all cases are history of domestic violence, actual or pending separation, obsessive behavior/stalking, depressed perpetrator, prior threats or attempts to commit suicide, escalation of violence, prior threats to kill the victim, prior attempts to isolate the victim, victims who had an intuitive sense of fear, a perpetrator who was unemployed, and substance abuse (full report at https://www.mcscs.jus.gov.on.ca/english/DeathInvestigations/office_coroner/PublicationsandReports/coroners_pubs.html).

DVDRC data summary form

The second coding form, the DVDRC data summary form, is a 15-page summary based on all case information, including perpetrator-specific information. This form was used to deduce sociodemographic information and service provider involvement. Service provider involvement was noted from the agencies/institutions section of the coding form, which asked about the involvement of 34 different service providers including criminal justice, child protection, and mental health agencies.

Procedure

The study was a retrospective case analysis and used quantitative data. Only cases that contained a male perpetrator and an adult female victim were examined. This inclusion criterion was due to male perpetrators and female victims being more prevalent in cases of domestic violence (WHO, 2016). Furthermore, because this study was examining perpetrator depression and substance abuse, cases in which perpetrators had other mental illnesses were excluded. This exclusion criterion was necessary to eliminate the compounding effects of other mental health conditions and thus make interpretations from the dataset more evident.

All cases which met the above inclusion criteria were examined. Cases were separated into four groups based on perpetrators with depression and/or substance abuse. The first group, the “with no history of depression and/or substance use” group, contained perpetrators who had no documented mental health diagnoses or symptoms. The second group contained perpetrators who only had diagnoses or symptoms of depression and the third group contained perpetrators who were only abusing substances. The final group, the “comorbid” group, included perpetrators who had comorbid depression and substance abuse.

Statistical Analyses

Chi-square tests of independence were used to compare the four mental health status groups on categorical dependent variables. Comparisons were made on types of risk factors and service provider contacts. Any cases where a variable being analyzed was unknown were excluded from that analysis. Fisher’s exact test was employed instead of a chi-square test for dependant variables where expected counts less than five made up more than 25% of the cells.

A one-way multivariate analysis of variance (MANOVA) was conducted to reduce the experiment-wise error rate that would have been incurred if separate analyses of variance (ANOVAs) were utilized for continuous dependent variables. In the MANOVA, number of risk factors and number of service providers were the dependant variables and mental health status was the independent variable. The independent variable had four levels, as indicated above, comprised of the comorbid group, depression only group, substance abuse only group, and with no history of depression and/or substance use group.

Results

Descriptive Statistics

In total, 133 cases were excluded from the analyses due to not meeting the requirements for inclusion. Out of these cases, 65 did not contain enough information to infer perpetrator depression and substance abuse, 53 involved perpetrators with other psychiatric concerns, eight involved child homicides, and seven involved same-sex couples or female perpetrators and male victims. Thus, 86 cases were kept in the sample for analysis after meeting requirements for inclusion. The “no history of depression and/or substance use” group contained 30 perpetrators, the depression only group contained 28 perpetrators, the substance abuse only group contained 15 perpetrators, and the comorbid depression and substance abuse group contained 13 perpetrators. All perpetrators in the substance abuse only group abused alcohol and seven (47%) abused street drugs (ex. cocaine, ecstasy, marijuana etc.) in combination with the alcohol. In the comorbid group, 12 (93%) perpetrators abused alcohol and four (31%) abused street drugs in combination with the alcohol. One (8%) perpetrator in the comorbid group abused both alcohol and prescription medication and another perpetrator solely abused streets drugs.

Due to the aforementioned exclusion criteria, all perpetrators in the sample were male. Sociodemographic variables including perpetrator age and citizenship were categorized and subjected to Fisher’s exact test (due to low expected cell counts) to determine whether there was an association between these sociodemographic variables and mental health status. No associations were found between these sociodemographic variables and mental health status (p > .05). In the overall sample, 77% (n = 66) of perpetrators were listed as Canadian, 20% (n = 17) were listed as immigrants or refugees, and 3% (n = 3) had unknown citizenship. To avoid any misuse of data based on race, additional background information about country of origin or cultural background is not kept. Almost 50% of perpetrators were in the 30- to 49-year-old age range (n = 41), whereas 29 (34%) were above 50 years old and 16 (19%) were below 29 years old.

Chi-Square and Fischer’s Exact Tests

Risk factors

All 40 risks factors previously identified by the DVDRC (2016) except for the four factors specific to mental health were subjected to chi-square tests for perpetrators with depression and/or substance abuse groups. Utilizing chi-square analyses, historical violence, separation, new partner, unemployment, threats to kill victim and victim intuitive fear were not found to be significant. However, obsessive behavior, prior threats of suicide, and escalation of violence were found to be statistically significant for perpetrators with depression and/or substance abuse groups, χ²(3, N = 82) = 7.95, p < .05; χ²(3, N = 74) = 16.08, p = .001; χ²(3, N = 80) = 9.10, p < .05, respectively. These associations were all moderate with Cramer’s V (.31, .47, and .34, respectively; Cohen, 1988). Post hoc analyses were conducted using adjusted standardized residuals and an alpha level of .006 with the Bonferroni correction. No significant differences were found for obsessive behavior or escalation of violence (p > .01). For prior threats of suicide, results for the “no history of depression and/or substance use” and depression only groups were significant (p < .01 for both). Perpetrators with no history of depression and/or substance use were less likely to have threatened suicide (19%; n = 5) and perpetrators with depression only were more likely to have threatened suicide (71%; n = 17).

Twenty seven risk factors did not meet the chi-square assumption of less than 25% of cells with an expected count less than five and were instead subjected to Fisher’s exact test. Witnessing/experiencing abuse, common-law, failure to comply with authority, destruction of property, assault with weapon, hostage–taking, and minimization of assaults all obtained significant results with Fisher’s exact test. Post hoc analyses, executed exactly as outlined above for the chi-square analyses, were employed for each of these risk factors to determine which cells were statistically significant. A Bonferroni correction yielding a value of .006 was also used. Witnessing/experiencing abuse, failure to comply with authority, destruction of property, and minimization were not significant in the post hoc analyses. Living in common-law relationships received a statistically significant post hoc result (p < .001), with perpetrators who were in the substance abuse only group having a reduced likelihood of having married the victim (47%; n = 7). Assault with a weapon was also found to be statistically significant in the post hoc analysis (p < .001), with perpetrators in the substance abuse only group having a higher likelihood of having a prior assault with a weapon (50%; n = 6). Finally, hostage-taking was statistically significant (p = .001), with perpetrators with comorbid depression and substance abuse having a higher likelihood of having engaged in hostage-taking behavior (39%; n = 5).

Service providers

The 34 service providers were combined into 11 groups to increase the statistical power of the chi-square and Fisher’s exact tests. Five of these 11 service provider categories contained multiple service providers within them. The courts and counsel category contained court judge, crown attorney, defense lawyer, criminal court, family court, and family lawyer. The corrections category contained corrections, probation, and parole. The mental health category contained mental health providers, mental health programs, anger management, marriage counseling, and substance abuse treatment. The health care category contained health care providers, local hospital, and ambulance. Finally, the immigrant and cultural category contained both immigrant and cultural service providers. Providers that were not subjected to analyses included court-based legal advocacy, victim witness assistance program, domestic violence shelter, sexual assault program, other domestic violence victim service, community-based legal advocacy, batterer intervention program, supervised visitation, and homeless shelter contacts because no perpetrators in any mental health status group had made these points of contact.

Police contact was subjected to a chi-square test for perpetrators with depression and/or substance abuse and obtained a statistically significant result, χ²(3, N = 85) = 18.15, p < .001. This association was moderate with Cramer’s V equal to .46 (Cohen, 1988). A post hoc analysis was conducted, again utilizing adjusted standardized residuals and a Bonferroni correction of .006, and it was found that the perpetrators with substance abuse only group significantly differed (p < .001). Perpetrators with substance abuse only were more likely to have had contact with police (87%; n = 13). Courts and counsel was also subjected to a chi-square test for mental health status and did not receive a statistically significant result; however, it was approaching significance (p = .07).

Due to having 25% or more of their cells with an expected count of less than 5 and violating a major assumption of chi-square, Fisher’s exact test was employed for nine service providers. Fisher’s exact test yielded statistically significant results for corrections, mental health care, and health care. Post hoc analyses, as outlined above, were utilized to deduce which cells were statistically significant for each of these variables. Perpetrators with depression only had a significantly (p = .002) reduced likelihood of having had contact with corrections (4%; n = 1). Similar to the previous police contact chi-square results, perpetrators with substance abuse only had a significantly (p = .005) increased likelihood of having had contact with corrections (53%; n = 8). Perpetrators with no history of depression and/or substance use were significantly (p = .006; p = .003) less likely to have had contact with mental health care (10%; n = 3) and health care (7%; n = 2). Finally, for contact with both mental health care and health care, perpetrators with comorbid depression and substance abuse had a significantly (p = .002 and p < .001, respectively) increased likelihood of having had contact (67%; n = 8 for both). See Table 1 for a breakdown of mental health care connections for the three mental health status groups afflicted by mental illness.

Table 1.

Descriptive Statistics for Mental Health Status Groups With Mental Illness by Type of Mental Health Care.

	n	Depression Only(n = 28)	Substance Abuse Only(n = 15)	Comorbid(n = 13)
	n	n (%)	n (%)	n (%)
Mental health care	51	9 (36)	3 (21)	8 (67)*
Mental health providers	51	8 (32)	0 (0)	3 (25)
Mental health programs	53	1 (4)	0 (0)	0 (0)
Anger management	55	0 (0)	1 (7)	1 (8)
Marriage counseling	55	2 (7)	0 (0)	3 (23)
Substance abuse treatment	54	0 (0)	3 (21)	5 (38)

p < .05.

MANOVA

A MANOVA was used to determine whether the mental health status groups differed in number of risk factors or number of service provider contacts. For risk factors, perpetrators could have up to 36% (excluding the four mental health risk factors) and for service providers perpetrators could have up to 25 contacts.

Assumption testing

Prior to conducting the MANOVA, assumption testing was performed to ensure that the MANOVA would produce valid results for comparisons between the dependant variables (risk factors and service providers) and the independent variable (mental health status). Assumptions of multivariate normality, homogeneity of variance, and absence of univariate outliers were all violated. In an effort to make the service provider data normally distributed, reduce the effects of the univariate outliers, and correct for unequal variances, a log transformation was utilized on the data. However, because MANOVA is relatively robust to deviations of normality, the risk factor data were not transformed. Despite having transformed the service provider data, the “no history of depression and/or substance use,” depression only and comorbid groups still failed to meet the assumption of normal distribution as per Shapiro–Wilk (p = .001, p = .014, and p = .01, respectively). However, upon examining skewness and kurtosis z-scores and using a statistical significance level of .01, the data were considered to be normally distributed because all scores were within ±2.58. Thus, despite the transformed service provider data and existing risk factor data failing Shapiro–Wilk tests, the MANOVA was run due to the calculated z-scores suggesting a normal distribution and due to the robustness of MANOVA from deviations of normality.

Outcome

The multivariate result for risk factors and service providers in mental health status was significant, F(6, 164) = 3.82, p = .016; Pillai’s Trace = .180, $η_{p}^{2}$ = .09. The test of between-subjects effects was also significant for mental health status groups in both risk factors, F(3, 82) = 2.79, p < .05; $η_{p}^{2}$ = .093, and service providers, F(3, 82) = 3.97, p < .05, $η_{p}^{2}$ = .127. Because the F ratio was significant for the between-subjects effect in risk factors and service providers, it indicated that at least one set of the means between the “no history of depression and/or substance use,” depression only, substance abuse only, and comorbid groups were significantly different for both dependant variables. Scheffe post hoc tests were chosen due to unequal group sizes and were performed to deduce which means were significantly different. For risk factors, there were no significant differences (p > .05) between groups; however, the comparison between the “no history of depression and/or substance use” group and comorbid group was reaching significance (p = .06). Overall, means showed a trend of “no history of depression and/or substance use” < depression only < substance abuse only < comorbid in number of risk factors (see Table 2; see Figure 1 for a visualization). There were also no significant differences (p > .05) between groups for service providers; however, the comparison between the depression only group and the substance abuse only group was reaching significance (p = .06). Through examining the original data, means showed a trend of depression < “no history of depression and/or substance use” < substance abuse < comorbid in number of service provider contacts (see Table 2; see Figure 2 for a visualization).

Table 2.

Descriptive Statistics for Risk Factors and Service Providers.

Mental Health Status Groups	n	Risk Factors		Service Providers(Not Transformed)		Service Providers(Log Transformed)
Mental Health Status Groups	n	M	SD	M	SD	M	SD
“No history of depression and/or substance use”	30	6.87	4.62	2.63	3.16	.40	.37
Depression only	28	8.35	5.57	1.89	2.44	.35	.30
Substance abuse only	15	9.80	7.07	4.13	2.62	.64	.28
Comorbid	13	12.15	6.84	4.38	3.73	.63	.31

Note. A log transformation was used on service provider data for the multivariate analysis of variance.

Figure 1.

Number of risk factors across mental health status groups.

Figure 2.

Number of service provider contacts across mental health status groups.

Discussion

Victim mental health correlates of domestic homicide have been a topic of much research to-date; however, researchers had yet to expand investigations to comorbid mental health conditions in perpetrators of domestic homicide. This study sought to elucidate whether differences exist between perpetrators of domestic homicide with comorbid depression and substance abuse, and perpetrators with only depression, only substance abuse, or “no history of depression and/or substance use.” An interesting finding from the current study was the presence of alcohol abuse in 96% of the substance abuse only and comorbid depression and substance abuse cases combined. This finding was not hypothesized; however, it is in line with research showing that alcohol is the most commonly used drug in Canada (Canadian Centre on Substance Use and Addiction, 2017). The perpetrators in our sample seem to be typical in their drug of choice, as alcohol is likely an affordable, convenient, and legal option over other drugs within our Canadian context. However, the types of substances that are abused by perpetrators of domestic homicide should be further investigated, as past research has suggested that professionals may only attend to one problem at a time and the abuse of different substances may not have been fully captured in our dataset of police reports and third-party interviews (Riger, Bennett, & Sigurvinsdottir, 2014).

Another major finding in this study was the lack of perpetrator engagement in mental health care across the three groups afflicted by depression and the lack of engagement of all four groups with batterer intervention programs. When evaluating Table 1, it is clear that mental health care engagement was drastically low for both the depression only and substance abuse only groups, and engagement was lower than ideal for the comorbid group as five perpetrators had not been connected with mental health services. In terms of batterer intervention programs, no perpetrators in any of the four groups were connected despite each group containing perpetrators who had been arrested in the past for domestic violence. These findings demonstrate the necessity of research into the barriers associated with help-seeking by perpetrators, friends and family, as well as the barriers for agency referral and barriers to mental health agencies providing service. A study by M. Campbell, Neil, Jaffe, and Kelly (2010) suggests that perpetrator help-seeking is likely to be easier if help-seeking is viewed more positively by society, as well as if trust in others and trust in professional training are increased. Furthermore, help-seeking by family and friends could be increased if community knowledge about domestic violence is fostered (M. Campbell et al., 2010). Removing barriers to help seeking is a complex challenge because the majority of men in batterer intervention programs are referred there by the courts, and that men who self-refer to batterer programs are more likely to drop out of the program and are more likely to reoffend (Gondolf, 2012). The critical role of the criminal justice system must be acknowledged in ensuring treatment and compliance with treatment (Gondolf, 2012).

Perpetrators With Comorbid Depression and Substance Abuse

In comparing the four mental health status groups, the number of risk factors and service provider contacts were not found to be significantly different from each other. This result was not anticipated, as it was hypothesized that there would be increased risk factors and service provider contacts for perpetrators with comorbid depression and substance abuse based on the complexity of their mental health concerns. Though the lack of significant findings may be due to small sample sizes, it may also illustrate a gap between the needs of perpetrators and the acquirement of help from service providers. In theory, perpetrators with comorbid mental health concerns should have contact with more service providers so that they receive comprehensive care; however, in practice the findings from the current study indicate that this may not occur. The reasons for this gap between perpetrator need and connection with services are speculative; however, it does suggest that perpetrators, community members, and service providers have some work to do in decreasing the barriers to help-seeking as noted above.

Furthermore, it should be highlighted that the presence of risk factors may predict whether perpetrators connect with service providers, and because perpetrators with comorbid depression and substance abuse were not found to have a significantly increased number of risks factors compared to the other mental health status groups, this may partially explain why no significant findings were found for service provider contacts. Though the number of risk factors and service provider contacts did not receive statistically significant results, both were approaching significance, with the comorbid depression and substance abuse group having the largest mean for both risk factors and service providers. A larger sample size is warranted in future studies investigating comorbid mental health conditions in perpetrators of domestic homicide.

Analyses also revealed that the comorbid depression and substance abuse group had a higher likelihood of having had a prior hostage-taking. Literature on hostage-taking in domestic violence contexts is scarce; however, Van Hasselt et al. (2005) used five case examples to elucidate risk factors involved in domestic violence hostage-taking cases. The authors found that substance abuse was a major risk factor for hostage-taking; however, this finding is slightly contrary to our results which did not find a significant number of hostage-taking cases in our substance abuse only group, but rather in our comorbid depression and substance abuse group (Van Hasselt et al., 2005). Hostage-taking could be a specific act occurring from the interaction between the anger and irritability in depression, and a reduction in cognitive processing in substance abuse, which have both been cited in previous literature (Dutton & Karakanta, 2013; Heinz, Beck, Meyer-Lindenberg, Sterzer, & Heinz, 2011). However, future research is warranted to examine this notion.

Perpetrators with comorbid depression and substance abuse also had an increased likelihood of contact with mental health and health care. The increased likelihood of having had contact with mental health care may suggest that the presence of multiple mental health concerns does translate to more mental health care involvement. However, the increased contact with health care was an interesting finding which may relate to the complex mental health needs of these perpetrators, as well as to the accessibility of health care in Ontario. This result could also be representative of comorbidities that exist between mental and physical conditions, as shown by a 1998 review which found an increased risk of premature death and chronic physical conditions in individuals with mental illness (Harris & Barraclough, 1998). Perpetrators with multiple mental illnesses may be more likely to have physical illnesses in addition to their mental health concerns which may make them more likely to contact health care providers. Future research is needed to test this notion.

Perpetrators With “No History of Depression and/or Substance Use”

Though this study aimed to specifically examine hypotheses in relation to perpetrators with comorbid depression and substance abuse, significant results obtained for the other mental health status groups are also important to discuss. The “no history of depression and/or substance use” group had a decreased likelihood of having had threatened to die by suicide and had a decreased likelihood of having contact with mental health care which relates well to the absence of mental illness in this group. However, the decreased likelihood of having had contact with health care providers was an interesting finding. If this result is related to the absence of mental illness in the group, then it may suggest that mental health conditions are a main reason why perpetrators may connect with health care providers. This idea is logical as many family doctors in Canada act as a primary contact for individuals with mental health challenges. Whatever the cause for this result, it does suggest that perpetrators without mental illness may not be connecting with health care providers as much as perpetrators with mental illness which emphasizes the need for health care providers to be aware of the mental health correlates of domestic violence and homicide.

Perpetrators With Depression Only

The depression only group achieved results in line with previous research reporting the association between depression and suicidal behaviors as they had an increased likelihood of having had threatened to die by suicide (Eliason, 2009). Results also showed that this group had a decreased likelihood of having had contact with corrections which makes intuitive sense given that they are not struggling with substance abuse which might amplify criminal behavior. This result is also important to emphasize since corrections, probation, and parole play a major role in monitoring perpetrators and preventing violence and homicide. Because these results indicate that these service providers will have a decreased likelihood of connecting with perpetrators who are struggling with depression, the onus may be on other service providers, such as health care and mental health providers, in addition to the perpetrators themselves or their friends and family, to seek assistance for this subgroup’s violence and mental health concerns.

Perpetrators With Substance Abuse Only

Results indicated that the substance abuse only group had a higher likelihood of having had a prior assault with a weapon which was not anticipated. However, this result is similar to the results of a Canadian study involving over 10,000 male offenders which found that over 53% of their sample were abusing substances and were involved in crimes involving weapon use (Correctional Service Canada, 2011). Perpetrators with substance abuse would likely struggle with cognitive processing and self-control which may explain why this group had a higher likelihood of having assaulted their victim with a weapon (Heinz et al., 2011; Room, Babor, & Rehm, 2005). The findings also indicated that perpetrators in the substance abuse only group had an increased likelihood of being in a common-law relationship with the victim. This result was another surprising finding which may relate to the reduced conflict resolution skills, financial difficulties, and infidelity that can be present in relationships when one partner is abusing substances (Room et al., 2005). Finally, the substance abuse only group had an increased likelihood of having had prior contact with police and corrections which are similar to findings from other studies which demonstrates increased criminal behavior in individuals abusing substances (Public Safety Canada, 2015). Moreover, this result emphasizes the necessity of substance abuse and domestic violence training for police, corrections, probation, and parole officers so that these concerns can be appropriately assessed and managed.

Implications

The implications of this study relate to service providers in different disciplines about their potential role when connecting with perpetrators of domestic violence.

Mental health agencies

Each perpetrator who was struggling with depression in our sample did not have contact with mental health providers or programs in their communities which indicates that barriers exist to perpetrators accessing mental health treatment. As indicated above, these barriers could include family members and friends not providing assistance, perpetrators having a lack of trust or not wanting to appear weak, or a lack of domestic violence training in agencies (M. Campbell et al., 2010). Though future research should aim to further identify these barriers, it is clear that mental health agencies need to make their services accessible to individuals who commit domestic violence. These agencies should ensure that they are equipped with enough training for their mental health workers to be comfortable and capable of working with perpetrators of domestic violence, including coordination with the justice system.

Mental health professionals should be able to recognize and address domestic violence when perpetrators present with mental health and/or relationship concerns. An initiative through the Oklahoma Domestic Violence Fatality Review Board has resulted in a specialized Mental Health and Domestic Violence committee that has brought awareness to the relationships between mental health, substance abuse, and domestic violence (Oklahoma Domestic Violence Fatality Review Board, 2014). This initiative has also established domestic violence liaison positions within mental health and substance abuse agencies across the state (Oklahoma Domestic Violence Fatality Review Board, 2014). The development of committees such as this one could assist in the unification of professionals from across sectors. Furthermore, these collaborative engagements could help to educate professionals on the assessment and management of co-occurring mental health and violence concerns. We acknowledge that there are multiple challenges to implementing these changes such as putting victims at risk by asking questions about domestic violence. A lack of professional training could lead to asking questions without knowing what to do with the answer. Another problem acknowledged in the field is the lack of any consistent liaison between mental health and anti-violence agencies.

Other service providers

Because the current study indicates that mental health agencies do not always have contact with perpetrators who are struggling with mental health concerns, it is important that other service providers, like health, justice, and social services, are able to recognize mental health concerns and subsequently refer perpetrators of domestic violence to mental health agencies. Therefore, increased training and education could be considered for service providers outside of mental health agencies on both simple and complex mental health presentations, and domestic violence. A police practice review completed by the Mental Health Commission of Canada (2010) revealed that mental health training was extremely variable across different policing units in Canada. This variability resulted in some police officers receiving extensive mental health training, and other officers receiving little to no mental health training. The current study indicates that all professionals having contact with perpetrators of domestic violence should receive extensive mental health training because a substantial amount of their clientele will be suffering from mental illness.

Overall, it is critical that service providers work together so that issues and concerns in the realm of domestic violence can be effectively mitigated, especially in cases of severe violence. One way to eliminate barriers to help-seeking would be through the creation of a comprehensive program which addresses the triad of violence, mental health, and relationship issues. As an example of this innovation, a Domestic Violence Mental Health court has been implemented in Miami-Dade County, Florida, which prescribes specialized risk management techniques to perpetrators of domestic violence who are suffering from mental health concerns (Winick, Wiener, Castro, Emmert, & Georges, 2010). This hybrid judicial model brings together frameworks for both domestic violence and mental health which enables perpetrators to receive the assistance that they need. Organizations that are able to adopt a similar model in their setting could make their services more accessible to perpetrators and also make it easier for service provider referral.

Limitations

Though the sample of DVDRC cases that were used in this study was a rich data source, there were limitations to its usage. First, this study had a limited sample size due to the inclusion of only Ontario domestic homicide cases which reduced the power to detect effects. Though our sample of Ontario domestic homicide cases may not generalize to other Canadian provinces, Ontario comprises 40% of the Canadian population which makes our sample extremely important in conveying an overall picture of Canadian domestic homicides. Also, again due to limited sample size, same-sex couples, female perpetrators, and male victims were excluded from this study because of their underrepresentation in the population. Categorization of perpetrators into mental health status groups was also a limitation of this study, as it was based on information that was obtained after the homicides had occurred, and in some cases it was based on the reports of family, friends, and/or acquaintances. Inaccurate categorizations may have occurred due to the inaccessibility of all possible case and background information. The study was not able to address the diversity of the Ontario population because ethnic and cultural background information were missing and not coded in the data set. We recognize that some populations may be more vulnerable to depression and substance abuse and may lack access to services such as Indigenous people who have suffered from historical abuse and oppression as part of colonization (Truth and Reconciliation Commission of Canada, 2016).

Conclusion

In summary, the study’s findings illustrate that perpetrators of domestic homicide are a heterogeneous group made up of a large proportion of individuals who are struggling with mental health concerns, but also of individuals who do not have any mental health concerns. Due to the heterogeneity of this group, it is crucial to consider the complexities of multiple mental health conditions and how that may relate to risk factors for domestic homicide and contact with service providers. Specifically, when comorbid depression and substance abuse is involved, perpetrators may have higher incidences of victim hostage-taking and connection with mental health and health care in the community. Furthermore, a major finding of the current study was the lack of mental health care and batterer intervention program engagement by perpetrators in all four groups. The findings of the current study demonstrate the need for service providers to have policies and protocols surrounding recognizing and addressing risk for domestic violence/homicide when risk factors are present along with mental health concerns. It is imperative that service providers work collaboratively so that perpetrator concerns are effectively managed and victims are fully protected.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Casey L. Oliver, MA, is a graduate student in the Centre for Research and Education on Violence Against Women and Children at Western University. Under the supervision of Dr. Peter Jaffe, her MA research focused on the mental health correlates of domestic homicide.

Peter G. Jaffe, PhD, is a psychologist and professor in the Faculty of Education at Western University. He is also the Academic Director of the Center for Research and Education on Violence Against Women & Children. He has coauthored 10 books, 24 chapters, and over 75 articles related to violence and abuse involving children, adults, families, and the justice system. In 2009, he was named an Officer in the Order of Canada by the Governor General for his work preventing domestic violence in the community.

References

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: American Psychiatric Publishing.

Boyce

Cotter

(2013). Homicide in Canada, 2012. Juristat: Canadian Centre for Justice Statistics, 1.

Campbell

J. C.

Webster

Koziol-Mclain

Block

Campbell

Curry

M. A.

. . . Laughon

(2003). Risk factors for femicide in abusive relationships: Results from a multisite case control study. American Journal of Public Health, 93, 1089-1097.

Campbell

Neil

J. A.

Jaffe

P. G.

Kelly

(2010). Engaging abusive men in seeking community intervention: A critical research & practice priority. Journal of Family Violence, 25, 413-422. doi:10.1007/s10896-010-9302-z

Canadian Centre on Substance Use and Addiction. (2017). Alcohol. Retrieved from http://www.ccdus.ca/Resource%20Library/CCSA-Canadian-Drug-Summary-Alcohol-2017-en.pdf

Cohen

(1988). Statistical power analysis for the behavioral sciences (2nd ed.). New York, NY: Psychology Press.

Correctional Service Canada. (2011). Linking type of substance use and type of crime in male offenders. Retrieved from http://www.csc-scc.gc.ca/research/005008-rs11-06-eng.shtml#_ftn5

Corrigan

Watson

(2005). Findings from the National Comorbidity Survey on the frequency of violent behavior in individuals with psychiatric disorders. Psychiatry Research, 136, 153-162.

Dawson

(2017). Domestic homicides and death reviews: An international perspective. London, England: Palgrave McMillan.

10.

Dawson

Jaffe

Campbell

Lucas

Kerr

(2017). Canada. In Dawson

(Ed.), Domestic homicides and death reviews: An international perspective (pp. 59-90). London, England: Palgrave McMillan.

11.

Dinwiddie

(1992). Psychiatric disorders among wife batterers. Comprehensive Psychiatry, 33, 411-416.

12.

Dobash

R. E.

Dobash

R. P.

Cavanagh

Lewis

(2004). Not an ordinary killer—Just an ordinary guy: When men murder an intimate woman partner. Violence Against Women, 10, 577-605.

13.

Domestic Violence Death Review Committee. (2016). 2015 annual report. Toronto, Ontario, Canada: Office of the Chief Coroner. Retrieved from https://www.mcscs.jus.gov.on.ca/sites/default/files/content/mcscs/docs/DVDRC%202015%20Annual%20Report%20English%20Accessible.pdf

14.

Dutton

D. G.

Karakanta

(2013). Depression as a risk marker for aggression: A critical review. Aggression and Violent Behavior, 18, 310-319. doi:10.1016/j.avb.2012.12.002

15.

Eliason

(2009). Murder-suicide: a review of the recent literature. Journal of the American Academy of Psychiatry and the Law Online, 37, 371-376.

16.

Fazel

Gulati

Linsell

Geddes

Grann

(2009). Schizophrenia and violence: Systematic review and meta-analysis. PLoS Medicine, 6(8), e1000120.

17.

Feinstein

A. R.

(1970). The pre-therapeutic classification of co-morbidity in chronic disease. Journal of Chronic Diseases, 23, 455-468.

18.

Ferrari

Agnew-Davies

Bailey

Howard

Howarth

Peters

T. J.

. . . Feder

G. S.

(2016). Domestic violence and mental health: A cross-sectional survey of women seeking help from domestic violence support services. Global Health Action, 9(1), Article 25519.

19.

Gondolf

(2012). The future of batterer intervention programs: Reassessing evidence-based practice. Boston, MA: Northeastern University Press.

20.

Goodman

L. A.

Fauci

J. E.

Sullivan

C. M.

DiGiovanni

C. D.

Wilson

J. M.

(2016). Domestic violence survivors’ empowerment and mental health: Exploring the role of the alliance with advocates. American Journal of Orthopsychiatry, 86, 286-296.

21.

Harris

Barraclough

(1998). Excess mortality of mental disorder. The British Journal of Psychiatry, 173, 11-53.

22.

Heinz

A. J.

Beck

Meyer-Lindenberg

Sterzer

Heinz

(2011). Cognitive and neurobiological mechanisms of alcohol-related aggression. Nature Reviews Neuroscience, 12, 400-413.

23.

Hirose

(1979). Depression and homicide. Acta Psychiatrica Scandinavica, 59, 211-217.

24.

Jaffe

P. G.

Dawson

Campbell

(2013). Developing a national collaborative approach to prevent domestic homicides: Domestic homicide review committees. Canadian Journal of Criminology and Criminal Justice, 55, 137-155.

25.

Jones

Hughes

Unterstaller

(2001). Post-traumatic stress disorder (PTSD) in victims of domestic violence. Trauma, Violence, & Abuse, 2, 99-119.

26.

Knight

Hester

(2016). Domestic violence and mental health in older adults. International Review of Psychiatry, 28, 464-474.

27.

Mason

O’Rinn

S. E.

(2014). Co-occurring intimate partner violence, mental health, and substance use problems: A scoping review. Global Health Action, 7(1), Article 24815.

28.

Mental Health Commission of Canada. (2010). Police interactions with persons with a mental illness: Police learning in the environment of contemporary policing. Retrieved from https://www.mentalhealthcommission.ca/sites/default/files/Law_Police_Interactions_Mental_Illness_Report_ENG_0_1.pdf

29.

Messing

Thaller

(2013). The average predictive validity of intimate partner violence risk assessment instruments. Journal of Interpersonal Violence, 28, 1537-1558.

30.

Mitchell

Anglin

(2009). Intimate partner violence: A health-based perspective. New York, NY: Oxford University Press.

31.

Oklahoma Domestic Violence Fatality Review Board. (2014). Domestic violence homicide in Oklahoma. Retrieved from http://www.okdhs.org/OKDHS%20PDF%20Library/DomesticViolenceHomicideinOklahomaDVFRBReport_cws_01192016.pdf

32.

Okuda

Olfson

Wang

Rubio

J. M.

Blanco

(2015). Correlates of intimate partner violence perpetration: Results from a National Epidemiologic Survey. Journal of Traumatic Stress, 28, 49-56.

33.

Public Safety Canada. (2015). Substance abuse. Retrieved from https://www.publicsafety.gc.ca/cnt/cntrng-crm/crrctns/sbstnc-bs-en.aspx

34.

Riger

Bennett

L. W.

Sigurvinsdottir

(2014). Barriers to addressing substance abuse in domestic violence court. American Journal of Community Psychology, 53, 208-217.

35.

Room

Babor

Rehm

(2005). Alcohol and public health. The Lancet, 365, 519-530.

36.

Rosenbaum

(1990). Role of depression in murder-suicide in couples involved and homicide. American Journal of Psychiatry, 147, 1036-1039.

37.

Rosenbaum

Bennett

(1986). Homicide and depression. American Journal of Psychiatry, 143, 367-370.

38.

Schlesinger

L. B.

(2000). Familicide, depression and catathymic process. Journal of Forensic Sciences, 45, 200-203.

39.

Sesar

Šimić

Dodaj

(2015). Differences in symptoms of depression, anxiety and stress between victims and perpetrators of intimate partner violence. Journal of Sociology & Social Work, 3, 63-72. doi:10.15640/jssw.v3n2a7

40.

Sharps

P. W.

Campbell

Gary

Webster

(2001). The role of alcohol use in intimate partner femicide. The American Journal on Addictions, 10, 122-135.

41.

Shorey

R. C.

Febres

Brasfield

Stuart

G. L.

(2012). The prevalence of mental health problems in men arrested for domestic violence. Journal of Family Violence, 27, 741-748.

42.

Swinson

Shaw

(2007). Homicides and mental disorders: The national confidential inquiry. Psychiatry, 6, 452-454. doi:10.1016/j.mppsy.2007.09.001

43.

Truth and Reconciliation Commission of Canada. (2016). Canada’s residential schools: The history, Part 2, 1939 to 2000: The final report of the truth and reconciliation commission of Canada (Vol. 1). Montreal, QC: McGill-Queen’s University Press.

44.

Van Dorn

Volavka

Johnson

. (2012). Mental disorder and violence: Is there a relationship beyond substance use? Social Psychiatry and Psychiatric Epidemiology, 47, 487-503.

45.

Van Hasselt

V. B.

Flood

J. J.

Romano

S. J.

Vecchi

G. M.

Fabrique

N. D.

Dalfonzo

V. A

. (2005). Hostage-taking in the context of domestic violence: Some case examples. Journal of Family Violence, 20, 21-27. doi:10.1007/s10896-005-1506-2

46.

Winick

B. J.

Wiener

Castro

Emmert

Georges

L. S.

(2010). Dealing with mentally ill domestic violence perpetrators: A therapeutic jurisprudence judicial model. International Journal of Law and Psychiatry, 33, 428-439.

47.

World Health Organization. (2016). Factsheet N° 239: Intimate partner and sexual violence against women. Retrieved from http://www.who.int/mediacentre/factsheets/fs239/en/