Community violence exposure and substance use: cross-cultural and gender perspectives

Although the rates of community violence exposure have declined since the peak in the 90s, it is still regarded as a “major public health problem” [1]. WHO has declared violence” a leading worldwide health problem” [2], and yet, most research regarding community violence exposure has focused on high-risk youth in the United States [3‐5] giving a rather narrow and perhaps, even skewed picture of the consequences of violence exposure.

The negative effects of community violence exposure on child and adolescent mental health are well documented and exposure to community violence has been linked both to a number of internalizing symptoms, such as anxiety, depression and somatization, as well as externalizing symptoms, like aggression and delinquency (for a review see e.g. [4, 6, 7]). Some studies have also explored the role of substance abuse, as a possible outcome of violence exposure [8‐11] linking the exposure to both magnitude and frequency of substance abuse. However the effect of violence exposure on substance use is far less examined [12].

It has been suggested that the effects of community violence can be long-lasting and can impact psycho-social functioning of youth in years after the exposure took place, which has been demonstrated by a number of longitudinal studies [6, 13]. Several factors have been identified that can potentially influence the extent and the consequences of traumatic exposure, including some aspects of trauma itself, age, and gender of the victim. The degree of exposure and proximity to trauma both influence psychopathology, developing after traumatic exposure, with a stronger association found in relation to victimization, as compared to witnessing [4]. When examining the consequences of violence exposure, it is also important to consider the fact that some youth may report higher levels of violence exposure by virtue of themselves being involved in violence as perpetrators and traumatic response among perpetrators might differ from that of innocent bystanders [13]. Antisocial youth may differ from others both in terms of their reaction to trauma, as well as in terms of higher levels of other externalizing problems, such as substance abuse.

There is still a certain controversy regarding gender specific effects of traumatic experience. While some studies suggest that girls are less exposed to community violence, as compared to boys [14], others found that females are more likely to experience trauma than males [15]. It has also been suggested that boys tend to react with more externalizing behaviors, whereas the girls to a higher degree respond with internalizing symptoms following exposure to community violence [13]. Age is another important factor both concerning the prevalence of exposure as well as the effects of it, where older children tend to report a higher prevalence [5] and seem to react with more externalizing symptoms compared to younger ones [4].

However, questions remain about the generalizability of the above findings to different cultures. There is some evidence that socio-cultural factors can influence both the prevalence and nature of symptoms. Some argue that the adverse effects of trauma exposure may differ by culture/ethnicity and that the expression of emotional distress is culture bound [16, 17], whereas others found that response to trauma is universal [18]. There are also differences in how symptoms are described and the explanatory models tend to vary across cultures, with attribution of symptoms to medical illness, family, work, environmental stress or culture-specific phenomena [19].

In summary, the mental health outcomes of community violence exposure are by now well established, however it’s relationship to substance use as well as how this relationship varies by gender and culture is poorly examined. The present study builds on one of the few studies on community violence with a cross-cultural perspective [10] however using a larger, culturally different sample and adding a gender perspective. The present study hypothesized that increased severity of exposure (from no exposure to witnessing to victimization) is associated with an increase in substance abuse, even when controlling for own perpetration of violence and for emotional problems. We further hypothesized that these relationships may be gender-specific, with higher levels of substance use in relation to community violence exposure among boys.

Table 1 shows the prevalence of violence exposure by country and gender. Similar to the previous studies, students from all three countries reported relatively high prevalence of exposure to community violence. Within each study group, more boys were exposed to episodes of violence than girls. The proportion of boys and girls in Russia and Czech Republic who witnessed at least one episode of violence was similar within country, but in the US a higher proportion of girls than boys reported witnessing violent events. In all three countries, more boys reported episodes of victimization by violence than girls (Table 1).

Table 2 presents the descriptive statistics (M, SD) for MANCOVA regarding differences in substance use according to the degree of severity of violence exposure for boys and girls in three countries. Table 3 presents effect sizes for each dependent variable (alcohol use and substance use scales), as well as the summary statistics. The main effect for the degree of exposure for the total group was significant (Wilks’ lambda = 0.966; F(10, 21,098) = 36.33, p < 0.000, η ² = 0.017), with increasing substance use scores for increasing exposure to community violence. The main effect for Gender was significant (Wilks’ lambda = 0.994; F(5, 10,549) = 12.71, p < 0.000, η ² = 0.006), demonstrating a difference in substance use between boys and girls. The main effect for Country was significant (Wilks’ lambda = 0.788; F(10, 21,098) = 266.48, p < 0.000, η ² = 0.112), suggesting differences in baseline levels of substance use in these three samples. Also, the interaction effect for Country × gender was significant (Wilks’ lambda = 0.988; F(10, 21,098) = 12.38, p < 0.000, η ² = 0.006), suggesting that gender differences in substance use followed different patterns in different cultures. The interaction effect for degree of exposure × Country was also significant (Wilks’ lambda = 0.972; F(20, 34,988) = 15.13, p < 0.000, η ² = 0.007), suggesting that the substance use levels differ depending on the level of violence exposure and country. However, neither the interaction effect for degree of exposure × gender (Wilks’ lambda = 0.999; F(10, 21,098) = 1.49, ns, η ² = 0.001), nor the interaction effects for degree of exposure × Country × gender (Wilks’ lambda = 0.997; F(20, 34,988) = 1.46, ns, η ² = 0.001) were significant, suggesting that patterns of substance use in response to varying degree of violence exposure were not gender-specific, and that despite substantial differences in the levels of substance use by country, by gender, and by varying degree of exposure to violence the gender-specific patterns of response to community violence exposure were similar across the three samples. In addition, the main effects for age (Wilks’ lambda = 0.858; F(5, 10,549) = 349.49, p < 0.000, η ² = 0.142), for own violent behavior (Wilks’ lambda = 0.853; F(5, 10,549) = 363.11, p < 0.000, η ² = 0.147), for emotional problems (Wilks’ lambda = 0.997; F(5, 10,549) = 5.93, p < 0.000, η ² = 0.003), and for SES proxy (Wilks’ lambda = 0.998; F(5, 10,549) = 4.80, p < 0.000, η ² = 0.002), were all significant and suggesting increasing use of substances with age, with increased involvement in violent behavior and with increased levels of emotional problems, as well as with lower SES. As shown in Table 3, in the three samples, substance use scores increased with severity of violence exposure and, as demonstrated by the follow-up univariate effects for degree of exposure, those who were victimized by violence reported the highest levels of substance use.

Considering that the differences by outcome, country and gender could have been masked by use of the MANCOVA analysis (i.e. by simultaneously assessing all three outcomes in one model), we have also attempted to examine each outcome separately in order to determine whether the patterns that are reported from the MANCOVA hold up with each outcome individually. The results obtained have been largely similar.

This cross-sectional, cross-cultural study found a significant relation between exposure to community violence on the one side, and alcohol and substance use on the other side, and this association seems to be robust across culture and gender. Across all three countries boys were more often victimized by violence than girls. After controlling for the subjects own level of violent behavior and for emotional problems, levels of alcohol and substance use increased along with increased level of violence exposure. Although, this association seems in some ways to be gender-specific (i.e. as concerns binge drinking), the pattern of these associations seems to hold in all three cultures.

The relation between violence and emotional problems on the one side and violence exposure and substance use on the other side is intricate and seems to be suggestive rather than conclusive, as they all represent risk factors for one another and may often co-occur [30, 31]. Some theories suggest that engagement in high-risk behaviors, like substance use, may represent a way of coping with negative effects following violence exposure for adolescents [32]. From this perspective, substance use has been considered as a “self-medication” in order to cope with the aftermath of violence exposure [9, 31]. This hypothesis is further supported by a prospective study of young adults that concluded that individuals with PTSD had four times higher risk of developing substance use disorder than those without the diagnosis [33]. At the same time, youth who use drugs tend to spend more time in risky environments and hence, are at greater risk for victimization by community violence [28]. Furthermore, substance use is connected to violent behavior either directly though disinhibition, or indirectly through association with aggressive peers, reduced coping skills, and increased life stressors [34]. In addition, those youth who perpetrate violence themselves are also more likely to witness or become victims of violence [35]. Hence, controlling for own violent behavior and emotional problems is essential when examining the relationship between violence exposure and substance abuse.

The present study found that in all three countries substance use increases along with the degree of violence exposure from no exposure, to witnessing to victimization. Previous studies have found a strong association between witnessing community violence and substance use [9] and one longitudinal study even found that this association was stronger with witnessing compared to victimization [36]. Assuming that the patterns of psychopathology in perpetrators and innocent bystanders tend to differ, one explanation for this discrepancy might be that the later study did not control for the subjects own violent acts.

Although, the research on gender differences with regard to the prevalence and consequences of community violence exposure has been rather inconclusive, several American studies have reported higher levels of exposure to community violence among boys than girls [5, 37] with more pronounced differences for victimization rates, as compared to witnessing. Our findings of boys reporting more episodes of victimization, as compared to girls, support the previous research and seem to hold across different countries, except for the high number of girls in the US sample reporting witnessing violence, as compared to boys.

There have been only a handful of studies regarding gender differences in the association between violence exposure and substance use and the findings tend to be inconclusive. One longitudinal study, Pinchevsky et al. [36], found that substance use in girls (but not in boys) was more strongly associated with witnessing, as opposed to direct victimization and that effect of victimization on binge drinking was also stronger in girls than in boys. However, Kilpatrick et al. [31], found no significant gender differences in the effect of witnessing violence on alcohol and drug use in youth, and neither did Kaufman [38], when examining the associations between victimization by violence and regular drinking. In the present study, the interaction effect of degree of exposure × gender was significant only for binge drinking, suggesting gender-specificity of this variable, possibly because of heavier drinking in response to violence exposure among boys. The interaction effects for all other substance use variables were non-significant, suggesting that although the rates of victimization and substance use varied by gender, the general patterns of increase in other types of substance use along with the levels of exposure were not gender-specific.

The role of culture in the development of adverse effects as a result of community violence exposure also remains unclear. Some argue that not only the prevalence of community violence exposure differ between cultures [5, 39], but that adverse effects of traumatic exposure may differ by culture/ethnicity and that culture plays a significant role in the expression of emotional distress [16, 17]. Others argue that trauma response knows no cultural boundaries [18]. Although the present study found significant differences between the three countries both with regard to the prevalence of violence exposure, and to the levels of substance and alcohol use, we refrain from making any direct comparisons between the study groups, as the groups differ substantially with regard to the background factors and the findings in the present study groups may not necessarily be generalizable to the respective countries as a whole. The important finding, however, was that in line with some previous studies on community violence exposure and substance and alcohol use e.g. [10], in all three countries an increase in the degree of violence exposure was similarly related to increase in substance and alcohol use. Hence, despite substantial between-county differences in levels of substance use and violence exposure, the pattern of response was similar cross-culturally.

The use of a cross-sectional design prevents us from determining causality and prospective studies are needed. The reliance of self-report on sensitive issues like violence exposure and substance use can produce recall-bias, however prior studies have indicated that self-reports on sensitive behavior by adolescents tend to be valid and reliable [40]. In fact when it comes to reporting exposure to violence adolescents themselves are the best informants, as parents often tend to underestimate the extent of violence exposure [24]. A strength of the study was the possibility to control for the participants’ own violent behavior in the multivariate analyses. Finally, the data collected in the study are over one decade old. The prevalence of the exposure to violence and of the substance use during this time could have changed, thus potentially changing the picture. However, the same pattern of associations observed in all three countries, despite the different levels of problems reported at each site, suggests that the associations would probably hold in spite the time has passed.

It is evident that community violence exposure, violence and substance abuse are interlinked. This study underline that an increase in violence exposure is associated with an increase in alcohol and substance use and that this pattern of association is neither gender nor culture bound. Hence, globally clinicians working with youth exposed to violence should assess substance use, and prevention programs aimed to reduce alcohol and substance use should address the issues of violence exposure. Finally practitioners working with delinquent youths need to take into account the effects of alcohol and substance abuse when intervening with youth at risk of violence exposure.