Purpose:  Youth living in the urban slums in Uganda have many unmet needs. The purpose of the current study is to briefly examine the associations between Adverse Childhood Experiences (ACEs) and alcohol use patterns, an understudied area among urban vulnerable youth, in order to provide guidance and insight for service provision for youth.
Methods: The current data analysis is based on a cross-sectional survey conducted in June 2011 which consisted of a convenience sample of youth living in the slums (ages 14-24) attending a drop-in center, Uganda Youth Development Link (UYDEL). Descriptive statistics were computed. Bivariate and multivariable logistic regression analyses were used to determine psychosocial correlates with being an orphan.
Results: Factors associated with alcohol use varied by sex (p<.05) (alcohol use was more common among girls), age (p<.000), (alcohol use more common among youth 18 years of age and older) and alcohol use were also more common among youth who reported the following ACEs; parents hitting each other (p<.000) , parental use of alcohol (p<.000), being hungry (p<.000), having ever lived on the street (p<.000) and having been raped (p<.000).
Conclusions: The prevalence of ACEs were high in this study population which is a great concern. Moreover ACEs were associated with earlier age of alcohol use initiation, frequent and heavy drinking underscoring the need for additional support for past childhood trauma as well as alcohol prevention efforts for these vulnerable youth.

Globally, alcohol is the third-leading risk factor for premature death and disability [1] and it accounts for 5.9% of all deaths and 5.1% of the disease burden [2]. The World Health Organization (WHO) has highlighted the detrimental use of alcohol globally including causing disease, social and economic burden in societies [3]. Disease burden has been found to be closely linked to volume of alcohol consumption, with disproportionate effects on poor people and those marginalized in society [4].

Alcohol use has been found to be associated with alcohol dependence [5,6], other substance use [6], criminal activity [6], unintentional injuries [5,7], suicidal ideation and attempts [8], and is an important risk factor for chronic disease and injury [9,10]. Recent findings from research also suggest associations between harmful alcohol use and infectious diseases such as tuberculosis and HIV/AIDS [11]. Findings have also shown differences and historical trends in alcohol consumption and related harm which can be explained by environmental factors such as economic development, culture, availability of alcohol and the level and effectiveness of alcohol policies [12].

Correlates of alcohol use include: genetic factors [12,13], age [12], gender [12,14,15], family characteristics [12,16], socio economic status [12,17], culture [12] and childhood maltreatment [18,19]. Findings on alcohol consumption particularly in Africa have shown that alcohol use and risky sexual behaviors are linked to drinking venues and alcohol serving establishments, sexual coercion, and poverty [21,22]. These results point to the pressing public health issue of alcohol use in sub-Saharan Africa which is linked to other health-risk behaviors and adverse outcomes.

Alcohol use as a result of ACEs is an important public health concern because of the several detrimental health and social outcomes associated with it [23].

Despite the vast literature on alcohol use and ACEs, there is limited research on alcohol use and ACEs among youth in Kampala, particularly those youth who do not attend school. This is concerning because these youth are vulnerable and have a high prospect to drink alcohol at young ages, drinking frequently and heavily [24]. As such, this study seeks to examine alcohol drinking patterns and ACEs, that is; early alcohol use initiation, frequency and intensity of drinking with the goal of informing program planning and intervention strategies.

A key area that has yet to be explored is the role of early Adverse Childhood Experiences (ACEs) in early alcohol use initiation and consumption patterns. Research in the U.S. on ACEs shows that; an association between ACEs and alcohol use at an early age which transcends time [20], experiencing 2 or more ACEs increases the risk of alcohol dependence by a similar margin, psychological distress increases the  risk of self-reported alcohol problems [6,25] and there are differences in how ACEs affect men and women [6,25] thereby indicating that ACEs likely have a strong association with early alcohol use and alcohol consumption patterns. Alcohol use is frequently used as a coping strategy among youth based on research in North America and Europe. It is very likely that youth in sub-Saharan Africa, and in Kampala specifically, will have similar experiences because research among the youth living in the slums and streets of Kampala has shown that 30.2% report problem drinking and 32.8% report drunkenness [24]. As such, the following study was guided by two research questions: What is the prevalence and socio demographic correlates of ACEs among high -risk youth living in the slums of Kampala? Do high-risk youth living in the slums of Kampala who report ACEs initiate alcohol use earlier and use alcohol frequently and heavily than those who did not experience ACES?  In this study, ACES were broadly defined and used indicators to reflect whether youth did not having living biological parents, did not actively talk to parents, saw parents hitting each other, parents hit the youth, parental use of alcohol, hunger, having ever lived on the street and having been raped.

These indicators were guided by extensive previous research [12,24,26,27,28]. Moreover, alcohol use initiation (<13 years and >13 years) was a key outcome as it is a key metric for efforts to delay and reduce alcohol use. Also, having had more than 5 drinks in the past month and having been drunk on more than 5 days in the past month were used to measure frequent and heavy drinking respectively.

Youth in the slums of Kampala lack so many basic needs [24,26,29]. Youth who live in the slums and on the streets rarely attend school and are particularly vulnerable to a range of health risks including substance abuse [30]. As such, early public health interventions are needed and the current study is conducted to inform prevention strategies in resource-poor settings.

Data from the Kampala Youth Survey conducted in May and June 2011 in Kampala, Uganda was used. The study protocol was approved by the Georgia State University Institutional Review Board and also by the Uganda National Council for Science and Technology. The main objective of this cross-sectional survey was to quantify and describe high-risk behaviors and exposures in a convenience sample of urban youth living on the streets or in the slums, 14–24 years of age, who were participating in a Uganda Youth Development Link (UYDEL) drop-in center for disadvantaged street youth. UYDEL is a non-profit making organization that operates in Kampala, Uganda whose goal is to enhance human capital development among the disadvantaged youth in the areas of child rights protection, HIV prevention, substance abuse, sexual and reproductive health, skills training and social research [31].

Face-to-face surveys, lasting about 30 minutes, were administered by social workers or peer educators employed by UYDEL. The study was implemented across eight drop-in centers across Kampala. Detailed methodology from the cross-sectional study has been reported previously [26,27].

Participant Selection, Inclusion and Exclusion Criteria

The youth who were selected to be a part of the study were those who had been currently receiving services from any of UYDEL’s 8 drop-in centers for street youth (UYDEL, 2013), which at that time, served approximately 650 youth per month.  For the present study, 507 youth were approached to participate in the survey. Among these, 46 declined and 461 agreed to participate, yielding a participation rate of 90.9%. Four surveys were missing substantial numbers of responses and were therefore excluded, yielding 457 completed surveys for the final analytic sample of youth between the ages of 14 and 24 (31.1% boys and 68.5% girls).

Several variables informed by previous research [12] andthe CDC ACE Study questionnaire [28], were used in the analysis to find relationships between adverse experiences and alcohol use status. Independent variables included were: hunger, having parents, talking to parents, ever having lived on the street, parents hitting each other, parents hitting children and parental use of alcohol, These were dichotomized to form two groups of either possessing the characteristic or not. For all selected variables, all missing values were excluded.

A description of the eight measures used as broad indicators of ACEs are described below; having parents - Participants were asked if their parents (fathers and mothers) were alive at the time of the survey (Categories were: yes, no and N/A were categorized as absence of the indicator). Talking to parents - Participants were asked how often they talked to their fathers and mothers (Categories were: ‘Yes’ for daily, weekly, monthly and annually or ‘no’ for never). Parents hitting each other - Participants were asked if they had seen or heard their parents hitting each other (Categories were: yes, no and N/A were categorized as absence of the indicator). Parents hitting you - Participants were asked if they had ever been hit or beaten by their parents (Categories were: yes, no and N/A were categorized as absence of the indicator). Parental use of alcohol - Participants were asked if their parents used alcohol (Categories were: yes, no and N/A were categorized as absence of the indicator). Hunger – Participants were asked if they were ever hungry (Categories were: yes, no and N/A were categorized as absence of the indicator). Having ever lived on the street – Participants were asked if they had ever lived on the streets or were currently living on the streets (Categories were: yes, no and N/A were categorized as absence of the indicator). Having been raped – Participants were asked if they had ever been raped (Categories were: yes, no and N/A were categorized as absence of the indicator).

The three outcome measures of interest were measured by; age of alcohol initiation (<13 years or >13 years), frequency of drinking (< 5 drinks or >5 drinks) and intensity of drinking (< 5 days or >5 days). All three categorical measures were coded to have three levels - presence or absence of the measure under study and the third level was non-drinkers who served as the reference group. Drinking status was assessed by participants’ response to the question “In the past month, on how many days did you drink alcohol? (Current drinking)”; Age of alcohol use initiation was assessed by participants’ response to   the question “How old were you when you had your first drink of alcohol? Heavy drinking was indicated for respondents who dank so much that they were really drunk on 5 or more days.

Independent Variables

Indicators to measure ACEs were informed by the World Health Organization [12] and the CDC ACE Study questionnaire [28] and included; hunger, having parents, talking to parents, ever having lived on the street, having ever been raped, parents hitting each other, parents hitting the youth and parental use of alcohol.

Three statistical techniques were used; Chi-Square tests of associations, Bivariate and Multivariate multinomial logistic regression analyses were computed to determine statistical association between ACEs and each of the 3 alcohol use measures among youth. Analyses were conducted using the SPSS statistical software package version 21. Initial bivariate analyses were performed. Measures with significant associations with the outcome variables were also examined in multivariate analyses.  All independent variables were dichotomized to indicate the presence or absence of the particular risk factor while all three dependent variables had 3 levels that is; age of initiation (age 13 or under, over age 13 and non-drinkers), frequent drinking (5 drinks or less, more than 5 drinks and non-drinkers) and heavy drinking (5 days or less, more than 5 days and non-drinkers).

The description of indicators used in this study among Kampala youth are presented in Table 1. The prevalence of ACEs and their distribution by demographic characteristics are presented in Table 2. In this study, 36.5% said they did not talk to their mother, 58.4% did not talk to their father, 32.4% had lost their mother, 51.6% had lost their father, 62.6% reported being hit by their parents, 30.9% had witnessed their parents hitting each other, 38.9% reported parental alcohol use, 60.6% reported ever going hungry, 22.1% reported having ever lived on the street while 24.1% reported having been raped.

Table 2. Demographic characteristics and ACEs measures across gender among Kampala youth (N=457)

The characteristics and indicators’ association with alcohol use are presented in Table 3. Factors associated with alcohol use varied by sex (p<.05) (alcohol use was more common among girls), age (p<.000), (alcohol use more common among youth 18 years of age and older) and alcohol use were also more common among youth who reported the following ACEs; parents hitting each other (p<.000) , parental use of alcohol (p<.000), being hungry (p<.000), having ever lived on the street (p<.000) and having been raped (p<.000).

Table 3. Characteristics and risk factors’ association with alcohol useamong Kampala youth (N=457)

Bivariate associations between alcohol use initiation, frequency of alcohol use, intensity of alcohol use and Adverse Childhood Experiences among youth living in the slums of Kampala are presented in Table 4. Bivariate multilogistic regression analyses were performed to; examine associations between ACES and early alcohol use initiation. Early alcohol use initiation was identified as youth who started drinking under age 13 and those who started drinking after they were 13 years old. The reference category was nondrinkers. Findings show that youth who started drinking before age 13 and those who started drinking after age 13 on the measures of; parents hitting each other (p<.001), OR <13=2.30 (95% confidence interval (CI):1.22-4.35), OR>13=2.89 (95% CI:1.85-4.54); parents hitting the youth (p<.05), OR<13=1.31 (95% CI:0.69-2.47), OR>13=1.85 (95% CI:1.16-2.93); parental use of alcohol (p<.001), OR<13=7.21 (95% CI:3.66-14.2), OR>13=4.50 (95% CI:2.86-7.10); being hungry (p<.001), OR<13=4.55 (95% CI:2.06-10.1), OR>13=2.09 (95% CI:1.34-3.29); having ever lived on the street (p<.001), OR<13=4.84 (95% CI:2.41-9.72), OR>13=4.95 (95% CI:2.94-8.32) and having been raped (p<.001), OR<13=3.22 (95% CI:1.63-6.35), OR>13=4.73 (95% CI:2.89-7.72); to identify group differences between youth who had drank alcohol on 5 days or less or more than 5 days in the past month and ACEs. Results indicated that significant differences existed between frequency of drinking and ACEs on the following measures: parents hitting each other (p<.001), OR<5=2.26 (95% CI:1.353.79), OR>5=3.62 (95% CI:2.01-6.52); parental use of alcohol (p<.001), OR<5=3.79 (95% CI:2.24-6.45), OR>5=6.01 (95% CI:3.16-11.4); being hungry (p<.001), OR<5=2.31 (95% CI:1.33-4.03), OR>5=3.06 (95% CI:1.56-5.99); having ever lived on the street (p<.001), OR<5=4.45 (95% CI:2.50-7.92), OR>5=8.62 (95% CI:4.63-16.04) and having been raped (p<.001), OR<5=3.54 (95% CI:2.05-6.12), OR>5=5.75 (95% CI:3.16-10.46); to identify group differences between youth who had been drunk on 5 days or less during the past month or were drunk on more than 5 days in the past month. Results indicated that there were significant differences between heavy drinkers and non-heavy drinkers on responses to: parents hitting each other (p<.001), OR<5=2.81 (95% CI:1.69-4.66), OR>5=1.77 (95% CI:.79-3.92); parental use of alcohol (p<.001), OR<5=4.45 (95% CI:2.61-7.60), OR>5=3.49 (95% CI:1.55-7.88); being hungry (p<.001), OR<5=2.99 (95% CI:1.66-5.39), OR>5=2.09 (95% CI:.91-4.83); having ever lived on the street (p<.001), OR<5=6.09 (95% CI:3.51-10.57), OR>5=10.52(95% CI:4.69-23.6) and having been raped (p<.001), OR<5=5.23 (95% CI:3.09-8.83), 2.81 (95% CI:1.27-6.21).

Table 4.Bivariate associations between age of alcohol use initiation, frequent, heavy drinking and ACEs among youth living in the slums of Kampala (N = 457)

The multivariate associations between alcohol use patterns (age of alcohol use initiation, frequency and heavy drinking) and Adverse Childhood Experiences among youth living in the slums of Kampala are presented in Table 5.Multivariate analysis was performed on the measures that came up significant in bivariate analyses. Results indicated that parental use of alcohol, having ever lived on the street and having been raped were all significantly associated with age of alcohol initiation, frequent drinking as well as heavy drinking.

The purpose of this study was to examine the association between ACEs and alcohol consumption patterns among youth living in slums in Kampala. Our findings show that several ACEs vary by sex and age and also are highly prevalent among these youth (Parental alcohol use, being hungry, and ever having lived on the street were common among while having been raped was more common among boys).

These findings of relatively high prevalence of ACEs (> 20% for each indicator) were expected because of the unique circumstances that these youth [26].

Moreover, several ACES are also strongly associated with all three outcome measures examined that is; parents hitting each other (p<.001), parental use of alcohol (p<.001), hunger (p<.001), ever having lived on the street (p<.001) and having been raped (p<.001).

Bivariate analyses showed gender differences for; parental alcohol use (p<.05), being hungry (p<.001), ever having lived on the street (p<.001) and having been raped (p<.001).

Findings from the bivariate analyses also indicated that parents hitting each other (p<.001), parents hitting the youth (p<.05), parental use of alcohol (p<.001), being hungry (p<.001), having ever lived on the street (p<.001), and having been raped (p<.001) were significantly associated with the youth’s age of alcohol initiation, frequent drinking and heavy drinking. These findings were consistent with previous research that has found similar results [30,32,33] as well as latter studies [34,35].

It was surprising that several parental involvement characteristics were not significantly associated with alcohol consumption patterns among these youth but this may be because of the unique circumstances of these youth that make them resilient or that parental involvement is viewed differently in this culture thus its impact on alcohol use would be different.

The indicators that were significant in bivariate analyses were then included in a multivariate model and significant associations were found between parental use of alcohol (p<.001), having ever lived on the street (p<.001) and having been raped (p<.001) for all three dependent variables of; age of alcohol initiation, frequent drinking and heavy drinking. These findings reflect previous research that has found similar results [36].

An interesting finding was the consistency of findings from the multivariate analysis, that is: parental alcohol use, ever having lived on the street and having been raped which were all statistically significant for age of alcohol initiation, frequency and intensity of drinking.

This study, a first of its kind, examined the associations between ACEs and several alcohol measures in an under studied, vulnerable, urban population in Kampala, Uganda and this gives an opportunity for knowledge about this particular population to be shared beyond the borders of the slums in which these high risk youth live thus providing an opportunity for more exploratory research and possibly interventions to improve public health in this and such populations. The study also looked at multiple outcome measures of alcohol consumption patterns, that is; age of initiation, frequency and intensity of drinking to assess their association with ACEs and this was able to give a better understanding of alcohol use behavior and not just current drinking patterns.

The results stated here should be viewed in light of several limitations; the study participants were not randomly selected, but were youth who self-selected to attend the drop-in centers and to take part of the study therefore, the findings may not be representative of street and slum youth in Kampala and may not be generalizable to populations elsewhere. The definition of youth who live in the slums was broad and included a range of circumstances and family contexts (both street youth who were homeless and youth who lived in the slums but may have had a stable living arrangement). Due to limited literacy rates, participants were read the questionnaire which could have led to bias. Most of the questions regarding alcohol use and high-risk behaviors were selected from previously established self-administered surveys, specifically the Youth Risk Behavior Survey conducted in the U.S. and the Global School-based Student Health Survey conducted primarily in Africa, Asia, and Latin America where the wording of the questions used was simplified in most cases and the response options narrowed to facilitate the administration of the survey by the interviewers. As such, the reliability and validity of some of the measures may have been altered. With the cross-sectional nature of the survey, the temporal ordering of ACEs and alcohol use cannot be determined, nor can causation be inferred.

For public health practice, it is important to recognize that evidence-based strategies used elsewhere may not be relevant for a low-income country such as Uganda or for vulnerable youth with particular needs and circumstances. With the limited healthcare resources available for treatment and prevention of infectious and other diseases in this and similar settings, it is clear that public health issues such as alcohol use and ACEs have been given less priority. However, the findings from this study can be used to advocate for the urgency of providing more resources and services to these vulnerable youth, most of who have lost one or more parents and have faced significant ACEs thereby compounding existing problems including alcohol abuse and its rippling effects.  As such, resources need to be focused on awareness and emphasize primary prevention of harmful alcohol use and ACEs among youth. The priority of providing more services to vulnerable youth needs to be considered complimentary to, and not competing with, other prevention efforts underway to address other critical health problems including disease, hunger and poverty.

Lastly, it is important to create a national child protection plan with local community support so that future interventions, policies, and resource allocation can produce the impact needed and to improve the conditions and health outcomes of children and vulnerable youth later in life.

Including ACEs into alcohol abuse prevention efforts could prevent a number of negative consequences and have significant impact on a range of serious health problems [37]. Specifically, prevention efforts cab be supported by: collecting ACE data to inform policy makers, increasing awareness of ACEs nationwide and at the community-level, emphasizing the relevance of ACEs to multiple behavioral health disciplines, and using ACEs research and collected data to identify groups of people who may be at higher risk for substance abuse and related behavioral health problems and tailoring interventions to best assist them [38].

The WHO highlights the need for a multi-sectoral approach in preventing child maltreatment and to maximize the effects of prevention and care, and recommends that interventions are delivered as part of a four-step public health approach: defining the problem; identifying causes and risk factors; designing and testing interventions aimed at minimizing the risk factors; disseminating information about the effectiveness of interventions and increasing the scale of proven effective interventions [39].

Additional research is needed to replicate these findings since they are primarily based on a convenience sample of youth who have sought out services offered by a drop-in center. As such, their experiences may be different than that of other youth and should be documented in greater detail and also preferably in a longitudinal cohort to better determine the initiation of risk behaviors and the identification of modifiable factors that may be suitable for prevention.

Future research is also needed to better determine strategies for providing additional services and treatment to these youths who may be difficult to reach. This is particularly troubling given the acute shortage of psychiatrists, psychologists, nurses, and social workers in Africa and Uganda specifically.

Despite the large health, social and economic burden associated with harmful use of alcohol, it has remained a relatively low priority in public policy, including in public health policy particularly in Africa [12]. More culturally appropriate research is needed in Uganda and sub-Saharan Africa at large in the area of youth alcohol consumption and ACEs and their association with social and behavioral problems that these individuals face later in life.

This study sought to assess the prevalence of ACEs among high-risk youth living in the slums of Kampala and whether the prevalence varies by sex and to examine if ACEs are associated with alcohol use initiation, frequent alcohol use and level of intensity of alcohol use. ACEs have serious long lasting implications on the life of the individual including susceptibility to alcohol use and this is especially true for high risk youth who have limited resources and support to improve their lives.

Alcohol has serious long lasting implications on the life of the individual especially for those who experience adversity early on in life. This is more pronounced among high risk youth who lack resources and support to improve their lives. The destructive use of alcohol should be made a priority for researchers and policy makers particularly in understudied populations like the one studied here but also largely in sub-Saharan Africa.

Research reported in this manuscript was conducted as part of an MPH thesis by the lead author Nina K. Babihuga under the supervision of the co-authors Drs. Swahn and Dube.