Prevalence and predictors of alcohol use among adult males in Ethiopia: multilevel analysis of Ethiopian Demographic and Health Survey 2016

Background Alcohol is a psychoactive substance that is widely consumed in the world. Alcohol use is one of the world’s leading risk factors for disease and disability. It affects individuals’ physical, mental, economic, and social issues. To our knowledge, there is limited study on alcohol consumption and associated factors. Therefore, this study aimed to determine the prevalence and predictors of alcohol use in Ethiopia by using the 2016 Ethiopian Demographic and Health Survey. Methods This study was based on the most recent Ethiopian Demographic and Health Survey 2016. A total of 12,594 men at the age of 15 to 59 were included in this study. Considering the hierarchical nature of EDHS data, a multilevel logistic regression model was applied. The ICC, MOR, and the LR test were done to assess the presence of a significant clustering effect. Besides, deviance was used for model comparison since the models were nested models. Variables with a p value ≤ 0.2 in the bivariable analysis were considered for the multivariable analysis. In the multilevel logistic regression, the adjusted odds ratio (AOR) with 95% confidence interval (CI) was reported to declare the strength and significance of the association between the dependent variable and independent variables. Results The prevalence of alcohol drinking in this study was 46.64% with a 95% CI of 45.00 to 47.00%. Age groups 30–44 (AOR = 1.30, 95% CI 1.08, 1.56) and 45–59 (AOR = 1.38, 95% CI 1.10, 1.74), Orthodox religion follower (AOR = 0.36, 95% CI 0.24, 0.55), media exposure (AOR = 1.67, 95% CI 1.41, 2.20), khat chewing (AOR = 3.08, 95% CI 2.54, 3.74), smoking (AOR = 2.18, 95% CI 1.71, 2.79), having no occupation (AOR = 0.34, 95% CI 0.22, 0.51), and region were the predictors of alcohol use in Ethiopia. Conclusions Nearly half of the Ethiopian population reported alcohol use at least once in their lifetime. Old age, Orthodox religion followers, media exposure, khat chewing, smoking, and having no occupation were predictors of alcohol use in Ethiopia. Therefore, health education about the risk of alcohol used is highly recommended. In addition, khat chewing and smoking control mechanisms should be designed and given special attention. Advertising alcohol through media is better to be controlled. Job opportunities should also be created for those who have no occupation to mitigate alcohol use in Ethiopia.


Introduction
Alcohol is a liquid that contains ethanol and the most predominant beverage worldwide [1]. Alcohol use is one of the world's leading health risks that results in 2.5 million death each year [2]. It is also a causal factor in many diseases and a precursor to injury, violence, and cardiovascular diseases [3]. Worldwide, alcohol use is associated with maternal and child health problems, risky sexual behavior, unintended pregnancy, injury, and poisoning [1]. Alcohol use triggers a host of public health harms, from injury and death accompanied by excessive drinking to increased violence, crime, poverty, and other forms of social destabilization, financial, disease, and death [4][5][6].
Alcohol is one of the most used and misused substances in different societies [7]. Due to alcohol drinking, 3 million Canadians are at risk of acute illness [8]. Alcohol is widely consumed by more than half of the population in the Americas, Europe, and the Western Pacific [1]. In Europe, alcohol leads to ischemic cardiovascular disease and injury, and death [9]. In England, about 10 million people are drinking at a level of which increases their risk of health problems. In age 15 to 49, alcohol is the leading cause of ill health, early mortality, and disability and the fifth leading cause for ill health in all age groups [10].
In low-and middle-income countries, alcohol use disproportionately affects premature mortality and disability [11]. In African countries, alcohol consumption has a large impact on the burden of disease and mortality, and alcohol exposure is expected to increase in the next years [12]. In the African continent, alcohol industry involvement and investment are rising following a general strategy to increase demand, availability, and access to alcoholic beverages [13].
Alcohol advertising and marketing are misleading the public in order to entice them to consume alcohol [21,29]. As a result, alcohol consumption is common and widely acceptable across all categories of people [30].
In Gondar Ethiopia, the prevalence of ever alcohol use was 48.23% [22]; in Ethiopia, systematic review and meta-analysis, pooled current and lifetime alcohol use were 23.86% and 44.16% respectively [31]. In another community study conducted in Ambo, Ethiopia, the prevalence of alcohol use disorder was 27% [32]. Alcohol consumption in Ethiopia is a risk factor for infectious diseases (tuberculosis, lower respiratory infections, viral hepatitis, sexually transmitted diseases including HIV), non-communicable diseases (heart diseases, noninfectious liver diseases, cancer), and mental disorders (alcohol use disorders including depression) [2]. As far as we know, in Ethiopia, studies showed that alcohol is the risk for different diseases and injuries rather than showed cause and effect relationship. Even though the risk of alcohol use is known in Ethiopia, less emphasis is given to the prevention strategies and the management of hazardous alcohol drinkers and addicted individuals [7]. Today, in Ethiopia, alcohol advertisement is prohibited with proclamation No. 759/2012. Advertised liquor with more than 12% alcoholic content may not be disseminated through mass media. Any liquor outdoor advertisement may not be placed within 100-m radius of children care center, school, medical or historical institution, cinema, theatre, and stadiums [33]. However, the fact is that this is not implemented. To our knowledge, there is a limited study on alcohol consumption and associated risk factors. Therefore, this study aimed to determine the prevalence and predictors of alcohol use in Ethiopia using the 2016 Ethiopian Demographic and Health Survey. The finding is crucial for policymakers and health professionals for effective intervention.

Study area
Ethiopia is found in East Africa of WHO region. It is located in the horn of Africa. Ethiopia had nine regions and two city administrations as shown in Fig. 1.

Data source
We used data from the most recent Ethiopian Demographic and Health Survey 2016 conducted in January 18, 2016, to June 27, 2016. It is conducted every 5-year interval. The survey had different datasets (individual records, kids record, household record, men records, birth cohort records etc.). For this study, men's record (MR) dataset was used. The data was freely accessible, and permission was obtained after projects are designed and submitted. The detail is found elsewhere [34].

Sampling procedure
A two-stage stratified sampling procedure was adopted in selecting study participants. The detail of sampling procedure is found elsewhere [35]. All men aged 15-59 who had been interviewed about ever alcohol drinking were included in the study. However, respondents with missing data for the outcome variable were excluded. A total of 14,795 eligible male respondents were selected,  and 12,594 were successfully interviewed, and the response rate was 85.12%.

Variables of the study
Outcome variable Respondents' ever alcohol drinking status, the outcome variable in this study, was defined as a person who ever drinks alcohol in his lifetime.

Independent variables
The independent variables were grossly classified into the individual-level and communitylevel variables: Individual-level variable includes ever chewed khat, age, religion, marital status, educational status, sex of household head, working status, occupation, source of information (reading newspaper, reading magazines, and watching television), and wealth index (poor, middle and rich). The community-level variables include residence and region.

Operational definition
Ever alcohol drinking was defined as a respondent who drinks alcohol during his lifetime. Ever chat chewer was defined as a respondent who chewed chat during his lifetime.

Data analysis procedure
To identify the predictors of alcohol use, the STATA 14 software was used. Sampling weight was done before any statistical analysis to adjust for the non-proportional allocation of the sample to different countries and the possible differences in response rates. Since the DHS data has hierarchical nature, measures of community variation/random-effects (intraclass correlation coefficient, median odds ratio [36], and proportional change in variance [37]) were estimated. The values of these measures were significant, indicating the use of a multilevel logistic regression model than ordinary logistic regression.
Model comparison was done using deviance between the null-model (a model with no independent variable), model I (a model with only individual-level factors), model II (a model with community-level factors), and model III (a model that contain both individual and communitylevel independent variables). A model with the lowest deviance (model III) was the best-fitted model.
Both bivariable and multivariable multilevel logistic regression were performed to identify the determinant factors of zinc utilization in Ethiopia. All variables with a p value < 0.25 at bivariable multilevel logistic model analysis were entered into the multivariable multilevel logistic regression model. p value ≤ 0.05 was used to declare statistically significant variables in the final model.

Sociodemographic characteristics
Total weighted samples of 12,594 participants were included in the analysis. The median age of the respondent was 29 with an interquartile range (IQR) of 21-39. Almost half of the participants, 6426 (51.03%) were between the age of 15 and 29 years. The majority, 10,098 (80.18%), of the men were rural. The majority, 8154 (64.74 %), of them had media exposure. 5876 (46.66%) men were in the primary education class. Around two-third of 7705 (61.17 %) participants were married (Table 1).

Random effect analysis
This study fits a model that considers the nature of the dataset. As known, the EDHS dataset had hierarchical nature. Therefore, fitting models that consider nature of the data is important. We fitted a generalized linear mixed effects model that had two component random effect and mixed effect. The fixed effect measures using odds ratio with the selected independent variables to qualify the effect size of low intake of food rich in vitamin A and independent variables. The random effect   (Table 2).

Discussion
This study showed the prevalence and associated factors of alcohol use in Ethiopia using the Ethiopian Demographic and Health Survey of 2016. The prevalence of lifetime alcohol use was 46.64% (95% CI 45-47%). The current finding was significantly lower than other findings in New Zealand [5], in India [14], in Nigeria [15], in Uganda [16], and in Nigeria [17]. The discrepancy might be due to that in New Zealand the age was 16-64 years, and the survey was a cohort study. In India, the study was conducted on only rural area dwellers, who are a highly vulnerable population for alcohol use, which is supported by other findings [24]. In Nigeria, researches were conducted in semirural dwellers and militaries who were more prevalent in the alcohol use population. In Sri Lanka, the study participants were mentally ill individuals.
Muslim, Protestant, and other religion followers decrease the odds of alcohol drinking 99.4%, 95%, and 64% as compared to Orthodox religion followers. The finding is consistent with other findings (51.6%) [15,20]. The reason might be Orthodox followers are culturally accepted in Ethiopia to drink alcohol.
In this study, the odds of alcohol use was about 1.69 [AOR = 1.69; 95% CI (1.14, 2.20)] times higher than those who were following media as compared to individuals who were not following media. There is evidence that alcohol consumption increase in Ethiopia from time to time [22]. Currently in Ethiopia, alcohol advertising though media is stopped by the Ethiopian government [31].
The study participants who smoked tobacco and chewed khat were highly significant with alcohol use when compared with their counterparts. This finding is supported by other findings [20,23,24]. The reason could be the action of alcohol is sedation [38]; to break the sedation, alcohol user also uses stimulants like khat and cigarette to get the feeling of being energized and hyperalert [39].
The odds of alcohol drinking among men who had occupations decreased by 66% as compared to men who had no occupation. This finding is supported by studies elsewhere [40]. Unemployment is an important factor for alcohol use, and problematic alcohol use crimp the likelihood of unemployment and decreases the chance of finding and holding down a job [41][42][43]. In Ethiopia, the government has no control over the production of locally brewed alcoholic drinks. Therefore, alcoholic beverage is found everywhere and everybody can access at low cost [44]. Unemployed men have more time available during the day in which to drink alcohol.

Conclusion
Nearly half of the Ethiopian population reported alcohol use at least once in their lifetime. Old age, being Orthodox religion followers, media exposure, khat chewing, smoking, and having no occupation were predictors of alcohol use in Ethiopia. Therefore, health education about the risk of alcohol used is highly recommended. In addition, khat chewing and smoking control mechanisms should be designed and given spatial attention. Job opportunities should be also created for the majority of society to alleviate alcohol use in Ethiopia.

Limitation of the study
Since the study is cross-sectional, it is not possible to establish a causal relationship between the independent and dependent variables. The study did not look at the pattern (frequency, dose) of alcohol use, harmful drinking, clinical aspects and consequences of dependence, and the consequence. The outcome measure for this study was by asking questions not by blood chemistry confirmation, and this may affect this study result.
Abbreviations AOR: Adjusted odds ratio; CI: Confidence; LLR: Likelihood ratio; RR: Relative risk; SNNPR: South Nation Nationalities of People Regions; EDHS: Ethiopian Demographic and Health survey