Changes in and Challenges for Intentional Injury in Vietnam

Family Violence and Related Factors

Youth in the SAVY2 sample lived primarily with their biological parents (about three fourths of respondents). Among 10 044 respondents, male participants accounted for 51% (vs 49% female), and the age groups 14 to 17, 18 to 21, and 22 to 25 years accounted for 48%, 29%, and 23% of the sample, respectively. The majority of the sample was unmarried (83%), and reflecting the rural dominance of the population in Vietnam, it was found that 75% of the sample came from rural settings.

Although there were no specific questions in SAVY for violence severity, the survey included questions about youth injury as a result of being beaten by a family member(s). In SAVY2, the overall proportion for intentional injury/violence caused by family members was 3.0% (an increase over the SAVY1 result of 2.2%). Among those who reported having ever been beaten, nearly one third said that they were injured in the 12 months prior to the SAVY2 survey. The proportion tended to be higher when the adolescent was younger; for male participants, the prevalence of having been beaten by a family member by age group was 4.4%, 3.2%, and 1.7% for 14- to 17-, 18- to 21-, and 22- to 25-year-olds, respectively. For female participants, the proportions in these 3 age groups were 3.7%, 2.2%, and 0.8%, respectively. These proportions were higher than that in SAVY1 (for male participants aged 14-17, 18-21, and 22-25 years, the proportions were 3.7%, 3%, and 1.2%, and for female participants, they were 1.6%, 1%, and 2%, respectively). Generally, the rate of being hit among male participants tends to decrease with age.

Overall, male youth are more vulnerable to violence. Figure 1 shows the proportion of injury caused by violence by age/gender and marital status. This percentage was significantly lower in the married compared with the unmarried group, suggesting the potential of violence being caused by parents and other family members to unmarried youth, whereas married people (with a certain proportion of them having a more independent life) were at less risk of being injured by family members other than their spouse.

OPEN IN VIEWER

Figure 1.

Percentage ever been injured by family members disaggregated by marital status, age group, and gender: SAVY 2009.

Bivariate analysis shows that urban youth are 0.5 times more likely to be injured by family members compared with their rural counterparts: 3% and 2%, respectively. This problem, however, did not significantly differ between ethnic groups or between economic status levels.

With regard to married youth, SAVY2 also explored the percentage who were ever hit by their spouse (Figure 2). The proportion of 4.1% was higher than the proportion of those who had ever been injured by other family members but lower than that in SAVY1 (5.2%). Women were nearly 6 times more likely than their male counterparts to report violence within marriage (1.0% vs 5.8%, P < .001). For women, this overall proportion in SAVY2 was slightly lower than that in SAVY1 (Figure 2). However, in younger age groups (14-17 and 18-21 years), the proportion in SAVY2 was higher compared with SAVY1. These data suggest that although the overall rate of exposure to violence within marriage was low, women disproportionately bear the brunt of it.

OPEN IN VIEWER

Figure 2.

Percentage of married youth ever hit by spouse: SAVY 2009.

Regarding urban/rural differences, in terms of spousal violence, there was a trend for urban females to report more violence than their rural counterparts (8.4% vs 5.3%), but this difference was not statistically significant. With regard to the question about having been hit, married men were less likely to report it than their single peers. Similarly, the proportion of married men who reported perpetrating spousal violence was lower than the proportion of married women who reported victimization. Of the married sample, among married women, 1.4% reported that the wife beat the husband, and 0.9% reported that the wife was beaten by the husband.

Violence and Its Correlates Outside the Home

SAVY gathered information on youth victimization as well as risk behaviors and the proportion who were intentionally injured by someone outside their family. Findings indicated that 7.6% of the sample had ever been injured as a result of nonfamily violence—no real change from SAVY1. This proportion (both sexes) ranged from 7.1% to 8.9% across age groups but was highest among 18- to 21-year-olds, and as with SAVY1, it was significantly higher among male than female participants: 12.8% versus 2.3% (proportions found in SAVY1 were 13.6% in male vs 2.4% in female participants).

In addition to such questions on victimization, SAVY2 also collected the information on other violent behaviors, including gang membership, participation in group riots, motorcycle racing, carrying weapons, and whether the young person ever hurt anyone badly enough to require medical attention. Overall, incidence of these behaviors, although more commonly reported than in SAVY1, remain quite low. Of the sample, only 2.6% had participated in motorbike racing, and 4.7% participated in group riots (the proportions in SAVY1 were 1.2% and 2.5%, respectively). The proportion of youth who ever carried a weapon was 2.8% (vs 2.3% in SAVY1). With regard to group riots, this was much more likely to be a male and urban than rural (10.7% vs 6.8%, P < .05) and female (2.6% vs 1.2%, P < .05) phenomenon. Also, as income rises, so too does the reporting of participating in group riots: 9.4% in male participants from the wealthiest households versus 8.4% in midlevel and only 4.9% in households with the lowest income level; a similar pattern was seen among female participants: 2.2% versus 1.5% and 0.9%, respectively.

Bivariate analysis showed an association between violent behaviors and other risk behaviors among youth. Youth who had ever been drunk were 8 times more likely to hurt someone seriously enough for them to require treatment, compared with peers who had never been drunk (3.5% vs 0.4%, P < .01). Also, young people who had ever been injured by family members were more likely to cause injury to others outside the home. Multivariate analyses identified the strongest predictors of violence toward others. The following factors were explored using binary logistic regression:

Taking into account the fact that the young people who were no longer in school might be exposed to different factors compared with those who were still in school, 2 regression models were developed. Model 1 (Table 1) was for the whole sample, and model 2 (Table 2) was for only those who still had a connection to school (this restriction of being in school certainly resulted in a smaller sample size compared with model 1).

OPEN IN VIEWER

Table 1.

Logistic Regression Models to Predict the Likelihood That Youth Had Ever Caused Intentional Injury to Others (Model 1 for the Whole Sample). ^a

Predictors (Independent Variables)	B	Standard Error	P Value	Adjusted OR
Gender
Male	0.940	0.333	.005	2.559
Female b	NA	NA	NA	1
Area
Urban	0.653	0.205	.001	1.922
Rural b	NA	NA	NA	1
Family connectedness scale
Weakly connected	0.561	0.207	.007	1.753
Strongly connected b	NA	NA	NA	1
Had ever been drunk
Yes	1.204	0.271	.000	3.332
No b	NA	NA	NA	1
Had ever been injured as a result of violence outside the home
Yes	1.170	0.216	.000	3.223
No b	NA	NA	NA	1
Had ever taken part in a group riot
Yes	2.118	0.240	.000	8.315
No b	NA	NA	NA	1
Had ever carried a weapon
Yes	0.873	0.259	.001	2.394
No b	NA	NA	NA	1
Had ever been injured as a result of violence of family members
Yes	1.023	0.327	.002	2.781
No b	NA	NA	NA	1

Abbreviation: OR, odds ratio; NA, not applicable.

a

Number of observations N = 8901. Goodness-of-fit Hosmer and Lemeshow test χ² = 10.136; df = 8; P = .256.

b

Reference group.

OPEN IN VIEWER

Table 2.

Logistic Regression Models to Predict the Likelihood That Youth Had Ever Caused Intentional Injury to Others (Model 2 for Those Still in School). ^a

Predictors (Independent Variables)	B	Standard Error	P Value	Adjusted OR
Gender
Female	−1.018	0.563	.071	0.361
Male b	NA	NA	NA	1
Area
Urban	0.752	0.364	.039	2.121
Rural b	NA	NA	NA	1
Family connectedness scale
Weak connected	0.263	0.365	.471	1.300
Strong connected b	NA	NA	NA	1
Had ever been drunk
Yes	1.584	0.433	.000	4.876
No b	NA	NA	NA	1
Self-esteem scale
Low	1.025	0.462	.026	2.787
High b	NA	NA	NA	1
Had ever taken part in a group riot
Yes	1.983	0.416	.000	7.262
No b	NA	NA	NA	1
Had ever carried weapons
Yes	1.637	0.419	.000	5.140
No b	NA	NA	NA	1
Sadness scale
Sad	0.702	0.410	.087	2.018
Not sad b	NA	NA	NA	1

Abbreviation: OR, odds ratio; NA, not applicable.

a

Number of observations N = 4297. Goodness-of-fit Hosmer and Lemeshow test χ² = 3.487; df = 8; P = .900.

b

Reference group.

As presented in Table 1, the final regression model for the whole sample affirmed several associated factors previously found in SAVY1 as risks of violent behavior:

In addition, SAVY2 data found 3 other important correlates: (1) urban respondents were at a higher risk (OR = 1.9); (2) those who had been injured by family members were 2.8 times more likely to behave violently; and (3) those who had strong family connectedness had nearly an 80% lower risk. Clearly, family connectedness was strongly protective. Age, ethnicity, geographic regions, socioeconomic status, sadness status, and self-esteem did not prove to be significant factors.

It is interesting to note in Table 2, with the regression model for youth who were still in school, that connectedness to school was protective; however, for school-going youth parental connection was no longer protective. This may reflect the very high prevalence of respondents reporting parent connectedness. Urban residence, participation in group riots, carrying weapons, and ever having been drunk are consistent risk factors for violence among school youth. Differently from the first model for the whole sample, for those still in school, self-esteem played a more important role. These data again confirm the importance of reducing alcohol abuse, participation in group riots, and weapon carrying, and at the same time, the data indicate that family connection and self-esteem are protective factors against violence toward others among those still going to school.

Self-inflicted Injury and Suicide Attempt

Similar to interpersonal violence, self-directed violence, including suicide thoughts and attempts are also addressed by SAVY2. The 3 main questions were about intentional self-injury, suicidal thoughts, and suicide attempts.

The findings show that 7.5% of youth in the SAVY2 sample reported having ever tried to injure themselves (vs only 2.8% in SAVY1), and 4.1% reported having ever thought of suicide (vs 3.4% in SAVY1). Of this SAVY2 sample, 102 (1%) young respondents ever attempted suicide (vs 0.55% in SAVY1). Although the rate of self-inflicted injury was relatively low, the lethality of attempts and the increasing trend since SAVY1 suggests that this is an issue that warrants close attention. There were also significant differences between male and female participants as well as youth in rural and urban settings. The highest proportion was found in the male, 14- to 17-year-old age group in rural areas (10.9%), followed by urban boys aged 14 to 17 years old (10.6%) and rural male participants aged 18 to 21 years old (9.1%). In both rural and urban settings, male participants were usually more likely than female participants to intentionally inflict self-injury; however, in contrast to SAVY1, for the whole sample aged 14 to 25years old, rural youth were overrepresented when compared with urban peers. Prevalence of suicidal thoughts was lower in male compared with female participants in both SAVY surveys (Figure 3). In both sexes, suicidal thought was related to alcohol consumption. Those who had ever been drunk were 2 to 3 times more likely to have suicidal thoughts (3.2% vs 1.4% in male and 12.6% vs 4.8% in female participants).

OPEN IN VIEWER

Figure 3.

Percentage ever thought of suicide by gender, age group, and location.

Intentional Injury Caused by Others

Even though the pattern of intentional injury found in the SAVY2 survey is not different from SAVY1 data and although the frequency of violent behaviors is still not very common, the magnitude of this problem seems to be more serious than previously reported. Regarding family, 3% of youth report having been injured by a family member. The rates by age group and sex are all higher than the rates found in SAVY1, and overall, male youth are more involved in violence. Also, urban youth are 50% more likely to be injured by family members compared with their rural counterparts. SAVY data revealed a rather low prevalence of violent behavior in the family compared with neighboring countries and some Latin American countries.^6,13 Only 3% of youth reported that they were injured by a family member, although the prevalence was higher in male than in female participants.

Regarding married youth, the general lifetime prevalence of ever having been hit by a spouse was 4.1% across both sexes—lower than in SAVY1 (5.2%). Family conflicts may result in spousal conflicts and fights, including yelling, swearing, prohibiting each other from doing certain things, or hitting each other. Women are nearly 3 times more likely than men to be beaten (3.9% vs 1.4%). Although married women are hit more frequently, married men experience more of the other abusive behaviors (35% of men vs almost 29% of women, P < .05). This pattern was very similar to findings from SAVY1 where women were nearly 6 times more likely than their male counterparts to report violence within marriage, underscoring the fact that although not common in Vietnamese families, women are now exposed to domestic violence. This finding is comparable to other studies of domestic violence in Vietnam. Krantz and Vung ¹⁴ found that 9.2% of 883 rural married women studied had been exposed to physical or sexual violence during the preceding 12 months. How this compares with SAVY2 is difficult to say because SAVY collected only lifetime prevalence data, not data relating to abuse or violence over the preceding 12 months.

At 7.6%,violence among Vietnamese youth is lower than in neighboring countries in Asia. ¹⁵ The pattern of interpersonal violence outside the family setting in SAVY2 was quite similar to the pattern found in SAVY1 data. As reported in other countries, the proportion of being intentionally injured by others, however, was significantly higher among male than female participants. ² Only 1.4% of the respondents reported that they had hurt other people seriously enough to require medical attention. Even though the proportion carrying weapons, taking part in a group riot, or taking part in motor racing was not high (2.8%, 4.7%, and 2.5%, respectively), these rates reflected an increase over the 5 years since SAVY1. Gang membership, participation in a group riot, motorcycle racing, and carrying a weapon have already been reported worldwide as risk factors for interpersonal violence by youth. ⁷ In the Vietnamese context, such behaviors have been seen as serious social problems. Overall, only a small proportion of youth had participated in motorbike racing or group riots; the proportion of youth reporting ever carrying a weapon was 2.8%, lower than in most other studies worldwide.^6,7 It is important to note that with the accumulated data from these 2 rounds of national youth surveys, the following characteristics are associated with juvenile violence: males in the urban setting, those who had ever been drunk, those who had been intentionally injured by others, those who took part in group riots and carried a weapon, and those who had low self-esteem. Interestingly, youth who reported a strong connection with their family were 80% less likely to cause intentional injury to others. This specific protective factor has not been reported elsewhere in the Vietnamese literature, though it has been studied elsewhere.^3,13

Some limitations of this study should be noted. The cross-sectional nature of both SAVY1 and SAVY2 does not allow us to deduce causality among those variables. Thus, it is not possible to clearly understand the complexity of the associations. For instance, a young girl in a highly violent family may be hit and beaten resulting in a hopeless feeling. Conceivably, she may then try to use alcohol, get drunk, and subsequently attempt suicide. This series of events cannot be documented in SAVY; however, such a pathway is described in the international literature on suicide.^16-18 The long recall period is also a potential source of bias. The sampling design of these 2 rounds of SAVY did not allow us to interview the same individuals or panel of youth over time. Therefore, there is no way to measure the change in behavior of the same cohort of people over 5 years.

Self-inflicted Injury

The findings showed that Vietnam has one of the lowest rates of youth suicide attempts in the world. ³ Higher rates have been found in Malaysia, China, and Hong Kong.^9,10,19 In contrast to the finding for violence against others, female respondents were found to be more likely to have suicidal thoughts and attempt suicide more than their male counterparts. The prevalence of attempted suicide was higher in married than unmarried youth, particularly among female respondents. This may reflect the fact that the difficulties and problems of married life have a more negative influence on the young wife’s emotional and psychological life than that of her husband. However, this finding also could be a result of the fact that married respondents were usually older.

Self-harm was more common in urban than in rural areas. Alcohol use was found to be a strong predictor of both violence against others and self-inflicted injury. Domestic violence and unhappy emotional life put youth, particularly young girls, at a higher risk of attempted suicide. These risk factors have been observed in similar studies in other countries. ⁸ Although the prevalence remains quite low in Vietnam compared with neighboring countries, there is a disturbing upward trend. This is an issue that deserves more attention because it may be an indication of social stress and future social problems. A better understanding of the factors that contribute to suicidal thoughts and suicide attempts among young people of Vietnam will be important if we are to avoid the same patterns here as are seen in much of Southeast Asia.