Impact statement
This review fills a critical gap in the understanding of alcohol consumption patterns among university students in the Association of Southeast Asian Nations (ASEAN), a population with unique sociocultural dynamics and limited prior investigations. By systematically analyzing sociodemographic factors such as gender, age and parental alcohol consumption, this study aims to provide actionable insights for policymakers, educators and public health practitioners seeking to design targeted interventions.
The findings reveal that male students are significantly more likely to consume alcohol than their female counterparts within ASEAN countries. Positive associations are found between consuming alcohol and both older age (within the university cohort) and parental consumption of alcohol. The review suggests the importance of culturally tailored, family-inclusive educational campaigns and policy measures aimed at reducing alcohol consumption among students. Furthermore, the identification of gaps in geographic representation and study designs offers a roadmap for future research to ensure more comprehensive regional coverage and robust methodologies.
The wider impact of this research lies in its potential to inform evidence-based policies and interventions that can improve student well-being and reduce the public health burden associated with alcohol misuse. At the regional level, it supports the United Nations’ agenda on youth health and safety. Internationally, the study contributes to the global understanding of alcohol consumption behaviors in diverse cultural contexts.
Introduction
Health-risk behaviors are actions that can lead to an increased risk of diseases and injuries (Surís et al., Reference Surís, Michaud, Akre and Sawyer2008; Peltzer and Pengpid, Reference Peltzer and Pengpid2016), and vary across different age groups, environments and cultures (Duell et al., Reference Duell, Steinberg, Icenogle, Chein, Chaudhary, Di Giunta, Dodge, Fanti, Lansford, Oburu and Pastorelli2018; Rattay et al., Reference Rattay, von der Lippe, Mauz, Richter, Holling and Lange2018; Wattanapisit et al., Reference Wattanapisit, Jiraporncharoen, Pinyopornpanish, Jiraniramai, Thaikla and Angkurawaranon2020). Among these, alcohol consumption is a significant concern, particularly for university students (Mekonen et al., Reference Mekonen, Fekadu, Chane and Bitew2017). In developed and emerging economies alike, university students, especially first-year students who are in early adulthood and transitioning from high school to university, have long been observed to engage in excessive alcohol use (Gill, Reference Gill2002; Hebden et al., Reference Hebden, Lyons, Goodwin and McCreanor2015). Habits formed during these years may also have long-lasting effects on health, potentially increasing the likelihood of noncommunicable diseases (NCDs) and chronic ailments in later life, such as diabetes, heart disease, stroke and certain cancers ( Shield et al., Reference Shield, Parry and Rehm2014; Budreviciute et al., Reference Budreviciute, Damiati, Sabir, Onder, Schuller-Goetzburg, Plakys, Katileviciute, Khoja and Kodzius2020). The burden of NCDs in emerging economies has been increasing in recent decades (Boutayeb and Boutayeb, Reference Boutayeb and Boutayeb2005; WHO, 2018) and reducing alcohol abuse is an explicit goal of the United Nation’s Sustainable Development Goal (SDG) 3.5 (Flor and Gakidou, Reference Flor and Gakidou2020).
More importantly, such behaviors are modifiable and preventable. While more intense forms of alcohol consumption, such as heavy episodic or binge drinking (defined by the World Health Organization as consuming at least 60 g of pure alcohol on a single occasion within the past 30 days), are significantly associated with alcohol-related harm (Moure-Rodríguez et al., Reference Moure-Rodríguez, Caamano-Isorna, Doallo, Juan-Salvadores, Corral, Rodríguez-Holguín and Cadaveira2014), researchers have found it easier to measure and study ordinary alcohol consumption (Yi et al., Reference Yi, Ngin, Peltzer and Pengpid2017; Wattanapisit et al., Reference Wattanapisit, Abdul Rahman, Car, Abdul-Mumin, de la, Chia, Rosenberg, Ho, Chaiyasong, Mahmudiono and Rodjarkpai2022), with which binge drinking is also correlated (Caffrey et al., Reference Caffrey, Caffrey, Puapan and Jariyapayulkert1996).
Focusing specifically on alcohol consumption among university students, Peltzer and Pengpid (Reference Peltzer and Pengpid2016) cite it as a significant public health issue, but also note that most research focuses on North American and European populations (Dantzer et al., Reference Dantzer, Wardle, Fuller, Pampalone and Steptoe2006; Wicki et al., Reference Wicki, Kuntsche and Gmel2010; Perera and Torabi, Reference Perera and Torabi2012; Venegas et al., Reference Venegas, Cooper, Naylor, Hanson and Blow2012). In developing economies, according to Peltzer and Pengpid (Reference Peltzer and Pengpid2016), estimates of drinking alcohol among university students range from 16.7% (China) to 48.8% (Malawi) for males, and from 3% (South Africa) to 24% (Colombia) for females. Other available evidence, while limited, also suggests that the prevalence of hazardous drinking in developing economies may be approaching that of industrialized countries. Taking an international perspective and reviewing published articles from 2005 to 2006 on alcohol use among university students Karam et al. (Reference Karam, Kypri and Salamoun2007) states that “in Europe, Australasia and South America, college student drinking is problematic, to an extent similar to that reported in North America.” More recently, Francis et al. (Reference Francis, Grosskurth, Changalucha, Kapiga and Weiss2014) conduct a systematic review of alcohol use among youth in East African countries, and find a 33% prevalence rate for university students, which was second only to male sex workers (69%).
Such observations highlight the need for more research into alcohol-related health risks across diverse global contexts. One such context is the set of rapidly developing and influential economies in the Association of Southeast Asian Nations (ASEAN), a regional intergovernmental organization founded in 1967 with shared goals and commitments to address challenges in areas such as human rights and public health (Lamy and Phua, Reference Lamy and Phua2012). Its founding members were Indonesia, Malaysia, the Philippines, Singapore and Thailand, and it has since expanded to include 10 member states, including (besides the ones above) Brunei Darussalam, Vietnam, Laos, Myanmar and Cambodia.
The population in ASEAN economies is young, educated and increasingly globalized, leading to a concurrent rise of modern health challenges, such as alcohol addiction among university students. In an influential study of 15,366 university students from seven ASEAN countries, Wattanapisit et al. (Reference Wattanapisit, Abdul Rahman, Car, Abdul-Mumin, de la, Chia, Rosenberg, Ho, Chaiyasong, Mahmudiono and Rodjarkpai2022) identify “alcohol drinker” as one of five clusters of health-risk behaviors and find that the prevalence of alcohol use exceeds 10%. Similarly, Yi et al. (Reference Yi, Ngin, Peltzer and Pengpid2017) survey 8,809 university students from nine ASEAN countries and determine a prevalence ranging between 12.8% (infrequent drinking) and 6.4% (binge drinking). At the country-specific level, prevalence estimates can vary significantly. Wattanapisit et al. (Reference Wattanapisit, Abdul Rahman, Car, Abdul-Mumin, de la, Chia, Rosenberg, Ho, Chaiyasong, Mahmudiono and Rodjarkpai2022) find that the odds of Singaporean university students consuming alcohol are 14 times that of students in Brunei, whereas, in Islamic countries like Indonesia and Malaysia, the odds are significantly lower.
Developing a better understanding of the sociodemographic drivers contributing to alcohol consumption among university students in ASEAN countries is a well-motivated and timely research agenda – one that could be critical for formulating effective and sustained strategies, policies and interventions. In other regions, several studies have shown positive associations between some sociodemographic determinants and alcohol consumption among youth. In the United Kingdom and Ireland, a systematic review found that gender was correlated with alcohol use disorder. However, the authors noted that the gender gap has narrowed recently (Davoren et al., Reference Davoren, Demant, Shiely and Perry2016). Other sociodemographic determinants that have been studied include factors such as whether the student lives on campus, is a member of a sports club, has parents who drink, academic performance and age (Mekonen et al., Reference Mekonen, Fekadu, Chane and Bitew2017; Wattanapisit et al., Reference Wattanapisit, Jiraporncharoen, Pinyopornpanish, Jiraniramai, Thaikla and Angkurawaranon2020). High alcohol consumption among university students is now known to be associated with a range of negative outcomes, including injuries, violence, academic underperformance, long-term health complications and increased healthcare burdens (White and Hingson, Reference White and Hingson2014; Paul et al., Reference Paul, Ganie and Dar2024), highlighting the urgent need for early prevention and policy intervention among university youth.
This study aims to conduct a systematic review of the sociodemographic factors associated with alcohol consumption among university students in ASEAN countries. The specific objectives are to (a) determine specific sociodemographic factors that have been found to be positively or negatively associated with alcohol consumption in university students within ASEAN countries; (b) compare these studies in terms of quality, rigor and key findings and (c) conduct a meta-analysis of sociodemographic factors found to be most commonly studied.
Methods
Design
The Population, Intervention, Comparator, Outcome, Study Design and Timeframe framework was used to define inclusion criteria, search concepts and research questions, as well as to guide decisions on study eligibility, data extraction and analysis (Appendix C of the Supplementary Material).
Inclusion and exclusion criteria
Following the PRISMA guidelines of (Moher et al., Reference Moher, Liberati, Tetzlaff, Altman and Group2009) and the Cochrane Handbook of (Higgins et al., Reference Higgins, Thomas, Chandler, Cumpston, Li and PageMandWelch2019), this review focuses on university students within ASEAN countries (Indonesia, Malaysia, the Philippines, Singapore, Thailand, Brunei, Vietnam, Laos, Myanmar and Cambodia). The primary outcome is alcohol consumption. In most of the included studies, alcohol consumption was treated as a binary outcome (drinker vs. nondrinker), often assessed through self-report or validated tools such as the Alcohol Use Disorders Identification Test (AUDIT). Variations in measurement were documented, and where necessary, outcomes were harmonized into a binary framework to enable meaningful comparison across studies.
Studies reporting sociodemographic exposures (e.g., gender, age, income, religion, peer influence) are prioritized, with students who do not consume alcohol as the comparator group. Both observational and interventional study designs were considered, without publication date restrictions, although recent studies (1990–2024) were prioritized. Only English-language, peer-reviewed articles and relevant reports (e.g., from WHO) were included, acknowledging the limitation of excluding non-English papers from the region. Studies only involving other drugs or substance use addictions were specifically excluded, as were studies not involving university students as the primary (or, at minimum, separately studied) population of interest. Conference abstracts, gray literature, opinion articles, editorials and review articles, including systematic reviews and meta-analyses, were excluded from the analysis.
Data
Nine databases were searched for published peer-reviewed articles with quantitative data, encompassing international (PubMed, Web of Science, Scopus, Medline, Embase, Cochrane Library), regional (Global Health, Global Index Medicus) and country-specific (Garuda Rujukan Digital, focused on Indonesia) sources to ensure broad coverage. The search strategy for each database is detailed in Appendix A of the Supplementary Material, along with a detailed count of documents found (where possible), thereby establishing a comparative baseline between the databases in terms of search coverage. Both thesaurus searching (e.g., using MeSH terms) and free-text searches were used. Although the search strategy mostly utilized international databases, only studies focused on ASEAN populations were included for the review and meta-analysis, as a key objective of the study is to fill ASEAN-specific knowledge gaps on alcohol use disorder among university students. Nevertheless, in the subsequent discussion, we do compare our findings with selected systematic reviews that have performed a similar analysis in other international contexts.
The references from the identified studies were also reviewed to identify other potentially relevant studies. The general strategy was designed to be inclusive, with no time limits applied to the search, and with synonyms, subject headings and Boolean operators like OR and AND judiciously used to obtain an initial set that could be reviewed for further inclusion. Specific inclusion and exclusion criteria are discussed next. The search was carried out between May 20 and 25, 2024. EndNote 21 was used as the reference manager.
The initial search results were first imported, and duplicates were removed based on a judicious combination of identical references found by matching the title and author, followed by a cursory manual review. Titles and abstracts were then screened, with studies excluded if they did not meet the inclusion and exclusion criteria. The full-text articles were then evaluated more closely in accordance with the established criteria. The reference lists of eligible studies were reviewed for any additional relevant articles, and suggested or similar articles found online were then evaluated for eligibility.
Each included full-text article was comprehensively reviewed, and the following study characteristics were extracted and tabulated both for individual study synthesis and statistical analysis and meta-analysis, using a data extraction form: first author and publication year, sample size, percentage of sample size that is female, period of study, country, age range of participants in years (with mean and standard deviation, where available), study design and relevant details on sampling method.
All included studies were ultimately found to be cross-sectional in design, and each study was further analyzed manually to extract the sociodemographic characteristics and exposures in the study as well as specific details on the outcome measured and key findings of interest distilled from the full text. These are tabulated separately, with a brief narrative synthesis provided. For sociodemographic characteristics that are studied in a sufficient number of studies, associations are further subjected to quantitative analysis and meta-analysis.
Statistical analysis and heterogeneity
All statistical analysis was conducted using Stata 17, and meta-analysis on each identified sociodemographic characteristic was conducted for which a sufficient number of studies were available (typically,
$ n>=5 $
, but exceptions could be made for
$ n=4 $
, with the caveat noted in the results that it is a small-scale meta-analysis in the results, and not necessarily generalizable). Considering the outcome measure (alcohol consumption), the specific way in which it is measured could differ (e.g., the form of the question that is asked) across studies, which is a likely source of heterogeneity in the meta-analysis. Therefore, an important goal of the results is to provide sufficient details on the measurement of how alcohol consumption was measured.
Because the studies together span multiple countries and include substantial heterogeneity, the random-effects model was used for the meta-analysis. For completeness, a fixed-effects meta-analysis was also conducted and compared to the random-effects results. Heterogeneity is tested based on the methodology described by Higgins et al. (Reference Higgins, Thomas, Chandler, Cumpston, Li and PageMandWelch2019), with both the
$ {I}^2 $
statistic and p-value from Cochran’s Q test reported. Publication bias is measured using both funnel plots and Egger’s test.
A systematic assessment of each study selected for the final review was conducted for quality and risk of bias, using the Appraisal tool for Cross-Sectional Studies (AXIS) instrument (Downes et al., Reference Downes, Brennan, Williams and Dean2016). A complete tabulation of the instrument for all included studies is reproduced in Appendix B of the Supplementary Material. The questionnaire includes 20 items spanning the major sections of the study (methods, results and so on). In using the tool for each cross-sectional study independently evaluated by it, a study is rated as high quality (low risk of bias), medium quality (moderate risk of bias) and low quality (high risk of bias) for total scores between 15–20, 10–14 and < 10, respectively.
Results
A total of 10,351 studies were gathered from searching the nine selected databases, including an additional 23 records identified through citation-searching (Figure 1). Of these, 2,428 (23.46%) were found to be duplicates and were removed. Following this process, the remaining 7,923 records were screened for suitability based on titles and available abstracts. Of these, 163 (2.06%) publications were judged to be suitable for further analysis, but of these, the full-text of four (2.45%) publications could not be found or retrieved. After screening the remaining 159 publications for eligibility, a total of 15 (9.43%) studies were deemed eligible for final inclusion in this systematic review based on screening conducted using the inclusion/exclusion criteria.
PRISMA flow diagram.
Study descriptions
The 15 studies included in this review encompassed a total sample of 35,527 individuals and collectively included all ASEAN countries, although some countries were more represented than others. All studies were cross-sectional in design and primarily used self-reported surveys as their instruments. The basic characteristics of the included studies are described in Table 1. Two studies spanned multiple countries, with Yi et al. (Reference Yi, Ngin, Peltzer and Pengpid2017) covering nine ASEAN countries (Cambodia, Indonesia, Laos, Malaysia, Myanmar, the Philippines, Singapore, Thailand and Vietnam) and Wattanapisit et al. (Reference Wattanapisit, Abdul Rahman, Car, Abdul-Mumin, de la, Chia, Rosenberg, Ho, Chaiyasong, Mahmudiono and Rodjarkpai2022) covering seven ASEAN countries (Philippines, Singapore, Thailand, Vietnam, Malaysia, Indonesia, Brunei Darussalam). Of the other 13 studies, three (exclusively) covered Myanmar, six covered Thailand, and one study each covered Vietnam, Cambodia, Malaysia and Indonesia. Based on the search criteria, no studies were found matching the criteria that exclusively covered the Philippines, Singapore, or Brunei Darussalam, although there is some literature (Lim et al., Reference Lim, Fong, Chan, Heng, Bhalla and Chew2007; Pagkatipunan, Reference Pagkatipunan2017; Lee et al., Reference Lee, Wang, Abdin, Chang, Shafie, Sambasivam, Tan, Tan, Heng, Vaingankar and Chong2020; Rahman et al., Reference Rahman, Julaini, Zaim, Masri and Abdul-Mumin2023) on the prevalence of alcohol consumption among the university population (but without any data on, or associations with, any sociodemographic variables), or covering other youth populations, such as school-going adolescents (Hong and Peltzer, Reference Hong and Peltzer2019; Pengpid and Peltzer, Reference Pengpid and Peltzer2019). With the exception of four studies, all studies reported the study period. Only one included study reported results before 2000 (Caffrey et al., Reference Caffrey, Caffrey, Puapan and Jariyapayulkert1996), with all other studies having been conducted over the last two decades (2008–2022). As expected, the age range of participants skews younger (typically between 18 and 24 years old), and in most studies, skews positively toward females.
Descriptive characteristics of all cross-sectional studies included in review
Identified sociodemographic characteristics, key associations and meta-analysis
In the context of this review’s first objective, the data compiled in Table 2 shows that a wide variety of sociodemographic characteristics have been explored across the 15 included studies. Gender was the most commonly reported exposure, and most studies consistently found that being male was significantly associated with alcohol consumption compared to being female. In a subsequent meta-analysis, a pooled odds ratio is used to quantify the effect on alcohol consumption across studies. In the studies where binge drinking was reported or explicitly differentiated from alcohol consumption per se, the association still remained strong.
Sociodemographic characteristics identified in each selected study, measurement of outcome and key results/associations distilled per study
a Unless otherwise specified through a footnote, all sociodemographic characteristics and measurement of outcome were obtained through self-administered/self-reported questionnaire.
b Abbreviations: AUDIT: Alcohol Use Disorder Identification Test (Babor et al., Reference Babor, Higgins-Biddle, Saunders and Monteiro2001), AOR: Adjusted Odds Ratio, CI: Confidence Interval, OR: Odds Ratio.
c The authors defined religiosity as a “latent construct” examining “how religious the participants are.” It is computed as a sum of mean scores of five specific items on the questionnaire.
d The same terminology (sex or gender) is used here as specified in the cited study.
e Although the authors purported to measure binge drinking, in logistic regressions, they combined infrequent and frequent binge drinking into a single category that has been referred to here simply as alcohol consumption.
f AUDIT four-category model: Abstainers (0 score), low risk drinkers (1–7 scores), high risk drinkers (8–15 scores) and dependence (16–40 scores).
Other common sociodemographic variables explored in the studies, albeit less commonly so than gender, were age, Grade Point Average (GPA) or academic performance, living situation (e.g., whether participants lived with parents or in an on-campus dormitory), smoking and tobacco use, attitude toward alcohol consumption and parental alcohol consumption. San San et al. (Reference San San, Oo, Yoshida, Harun-Or-Rashid and Sakamoto2010) and Yi et al. (Reference Yi, Ngin, Peltzer and Pengpid2017) highlighted that age and life satisfaction, combined with factors like peer influence and physical activity, affect the likelihood of alcohol consumption. Parental influence, particularly the drinking behavior of parents, was repeatedly cited as a strong predictor of alcohol uses (San San et al., Reference San San, Oo, Yoshida, Harun-Or-Rashid and Sakamoto2010; Phoosuwan, Reference Phoosuwan2019). Less certain is the role of religion, academic performance and living situation, which showed variable influence, with religiosity not always significantly associated with alcohol consumption (Nguyen et al., Reference Nguyen, Paulson and Opatrny Pease2024).
Meta-analysis of recurring characteristics
The results of the random-effects meta-analysis for gender are shown in Figure 2a. An overall OR of 3.15 (95% CI: 2.24–4.06) was observed, which suggests that being male is significantly and strongly associated with alcohol consumption. However, the heterogeneity was substantial and significant (
$ {I}^2 $
=85.2%, p < 0.001), suggesting that the results need to be treated with caution. This high heterogeneity is consistent with two other meta-analyses that have also measured associations between gender and alcohol consumption, including a meta-analysis by Nascimento et al. (Reference Nascimento, JdS and CAFd2022) (who observed an
$ {I}^2 $
statistic of over 95%) and Francis et al. (Reference Francis, Grosskurth, Changalucha, Kapiga and Weiss2014) (who also noted significant heterogeneity in all subgroups using the
$ {I}^2 $
statistic, but unlike Nascimento et al. (Reference Nascimento, JdS and CAFd2022) and this study, deferred from reporting a pooled association measure).
Association between consuming alcohol and (a) gender (male vs. female), (b) age (oldest age group vs. youngest age group) and (c) parental consumption of alcohol (having parents who consume alcohol vs. parents not consuming alcohol) using random-effects pooling of odds ratios (ORs) identified in relevant selected studies.
A sensitivity test was also conducted by excluding the studies by Sok et al. (Reference Sok, Pal, Tuot, Yi, Chhoun and Yi2020) and Supit et al. (Reference Supit, Mamuaja and Pissu2017) and redoing the meta-analysis, as they seem to be outliers compared to the other studies. Although not illustrated herein, the heterogeneity was found to reduce to moderate levels when only the other six studies were considered (
$ {I}^2 $
=57.6%, p = 0.038). The effect of being male remained strong and significant, but was reduced (overall OR = 2.28, 95% CI: 1.78–2.77). These ORs are generally consistent with the association measures found in the literature for other regions: Nascimento et al. (Reference Nascimento, JdS and CAFd2022) find that, among Brazilian medical students, 65% of male students engage in binge drinking compared to 47% for women. Results compiled by Peltzer and Pengpid (Reference Peltzer and Pengpid2016) find similar gender-specific prevalence differences in Colombia (46% male vs. 24% female), China (ranging from 16.7 to 37.4% male vs. 5.4 to 11.6% female), Malawi (48.8% male vs. 5.0% female), South Africa (27% male vs. 3% female), Thailand (32% male vs. 7% female), Uganda (34.1% male vs. 23.4% female) and Venezuela (32% male vs. 15% female). In almost all of these cases, the prevalence of alcohol consumption among males is at least two to three times higher than among females.
In a second sensitivity analysis, where the multicountry surveys by Yi et al. (Reference Yi, Ngin, Peltzer and Pengpid2017) and Wattanapisit et al. (Reference Wattanapisit, Abdul Rahman, Car, Abdul-Mumin, de la, Chia, Rosenberg, Ho, Chaiyasong, Mahmudiono and Rodjarkpai2022) were included, heterogeneity went up substantially (
$ {I}^2 $
=93.4%, p < 0.001), but overall OR remained at commensurate levels compared to earlier analyses (OR = 2.49, 95% CI: 1.88–3.10). Together, these analyses confirm that the odds of alcohol consumption by college students in ASEAN countries remain two to three times higher for male students, compared to females.
Figure 2b uses the random-effects model to show that the overall odds of consuming alcohol are 1.5 times (overall OR = 1.5, 95% CI: 1.28–1.72) for students in the oldest age group in college (who still tend to be young; typically in early-mid 20s) compared to the youngest. However, there are only four studies that could be included, so the results should be cautiously interpreted as arising from a small-scale meta-analysis. Heterogeneity is nonexistent (
$ {I}^2 $
=0%, p = 0.554), in stark contrast with the gender analysis. Fixed-effects meta-analysis, while not reproduced in the figure, showed that the overall OR, 95% confidence intervals and heterogeneity were unchanged.
The only other sociodemographic variable for which enough studies were available for conducting meta-analysis was parental consumption of alcohol. Random-effects results, demonstrated in Figure 2c, show that the overall odds of consuming alcohol were 1.58 times higher (OR = 1.58, 95% CI: 1.31–1.85) for those with parents consuming alcohol, compared to those whose parents did not consume alcohol. Heterogeneity was again low (
$ {I}^2 $
=13.6%, p = 0.327) and not significant. Robustness analysis using the fixed-effects model led to similar conclusions, with heterogeneity unchanged but the overall OR experiencing a slight decline (OR = 1.56, 95% CI: 1.34–1.78).
Quality assessment
Using the AXIS instrument (Appendix B of the Supplementary Material), eight studies were classified as high quality, with scores ranging from 17 to 19, indicating robust study design, appropriate population sampling and well-reported methodologies. Notably, the study by Aung et al. (Reference Aung, Ou, Wan, Nyan and Phyo2019) achieved the highest score of 19, with strengths in sample representativeness and efforts to address nonresponders. Other high-quality studies, such as those by Buakate et al. (Reference Buakate, Thirarattanasunthon and Wongrith2022), Htet et al. (Reference Htet, Saw, Saw, Htun, Lay Mon, Cho, Thike, Khine, Kariya, Yamamoto and Hamajima2020) and Jaichuen et al. (Reference Jaichuen, Chaiyasong, Nasueb and Thamarangsi2018) displayed strong adherence to AXIS guidelines, including clear objectives and reliable outcome measures. However, certain areas for improvement, such as better reporting of nonresponse rates and justifications for sample sizes, were noted even in these higher scoring studies. A common issue across almost all studies was the inadequate addressing potential response bias, and the lack of attempting to characterize nonresponse. A smaller number of studies were categorized as medium quality, scoring between 10 and 14. Weakness in these studies included a lack of justification for their sample size or an inadequate discussion of the study’s limitations (Mathialagan and Teng, Reference Mathialagan and Teng2017; Supit et al., Reference Supit, Mamuaja and Pissu2017; Nguyen et al., Reference Nguyen, Paulson and Opatrny Pease2024). Overall, however, the studies’ internal consistency and appropriate use of statistical methods still support their inclusion in the review. Both Tonkuriman et al. (Reference Tonkuriman, Sethabouppha, Thungjaroenkul and Kittirattanapaiboon2019) and Boonchuaythanasit et al. (Reference Boonchuaythanasit, Ponrachom and Cardinal2021) use a structural equation model, with associations that cannot be reconciled with the odds ratios prevalent in other models; hence, they are excluded from meta-analyses. Studies by Yi et al. (Reference Yi, Ngin, Peltzer and Pengpid2017) and Wattanapisit et al. (Reference Wattanapisit, Abdul Rahman, Car, Abdul-Mumin, de la, Chia, Rosenberg, Ho, Chaiyasong, Mahmudiono and Rodjarkpai2022) are both multicountry, in contrast with all other studies, and do not provide separate associations per country. They are also excluded from initial meta-analyses but are included in appropriate sensitivity analyses.
Publication bias
For evaluating the risk of bias across selected studies, funnel plots were generated for each of the three common sociodemographic characteristics studied earlier in the meta-analysis (gender, age and parental alcohol consumption). As illustrated in Figure 3, for two of the sociodemographic characteristics (age and parental alcohol consumption), studies tended to be located within the triangle region and scattered close to the line of
$ \mathit{\log}(OR)=1 $
. The situation is slightly more complex for gender: several studies are located outside the triangle region; however, despite this scattering, there is symmetry on either side. Using Egger’s regression test for small-study effects to formally evaluate publication bias, we obtained
$ p=0.764 $
(gender),
$ p=0.447 $
(age) and
$ p=0.171 $
(parental alcohol consumption). Because all three p-values are greater than 0.05, there is little evidence against the null hypothesis of symmetry (indicating no publication bias), and hence, publication bias in the previously reported meta-analyses is not present.
Funnel plot for assessing publication bias in studies included in the meta-analysis (Figure 1) for (a) gender, (b) age and (c) parental alcohol consumption.
Discussion
Our key objectives were to determine both the set and strengths of sociodemographic variables found to be associated with alcohol consumption among university students in ASEAN countries. In performing a literature search for the systematic review, we found 15 eligible cross-sectional studies covering all the ASEAN countries, either individually or as part of a comprehensive study sampling students across countries (Figure 1). Each study was varied in its coverage and measurements of sociodemographic characteristics (Table 1). Age, gender and parental alcohol consumption were among the most frequently occurring characteristics identified, although characteristics like religiosity and academic performance were also observed in more than one study. In distilling other key findings from these studies (Table 2), peer influence and smoking behavior were also found to be recurrent factors that significantly predicted alcohol use among students (Htet et al., Reference Htet, Saw, Saw, Htun, Lay Mon, Cho, Thike, Khine, Kariya, Yamamoto and Hamajima2020; Buakate et al., Reference Buakate, Thirarattanasunthon and Wongrith2022).
While no other systematic review and meta-analysis of sociodemographic factors exists, to our knowledge, of alcohol consumption among university students in ASEAN countries, a preliminary survey of related reviews (e.g., in other countries, or covering youth populations other than university students) suggests that some of the main findings distilled from this review are consistent with those by Francis et al. (Reference Francis, Grosskurth, Changalucha, Kapiga and Weiss2014), Karam et al. (Reference Karam, Kypri and Salamoun2007), Martens et al. (Reference Martens, Dams-O’Connor and Beck2006) and Wicki et al. (Reference Wicki, Kuntsche and Gmel2010). For example, Karam et al. (Reference Karam, Kypri and Salamoun2007) take a comparative international perspective, and reviews, relative to North America, “similar risk factors and protective factors” (including male gender and family use of alcohol) in regions as diverse as Australasia, North and South America, Africa and Asia. Although their study did not include a meta-analysis, it confirmed significant associations between alcohol use and sociodemographic factors, including male gender, higher socioeconomic status, greater family educational attainment and excessive alcohol use by family members or peers (Karam et al., Reference Karam, Kypri and Salamoun2007).
Another related review by Francis et al. (Reference Francis, Grosskurth, Changalucha, Kapiga and Weiss2014) conducted both a systematic review and meta-analysis of alcohol use prevalence among youth in eastern Africa. They also found that most studies were cross-sectional in design, and because they considered a broader population group (including school students, as well as university students, street children and sex workers), were able to determine that the prevalence of alcohol use was highest among university students and male sex workers compared to other groups. Specifically considering gender-specific differences, they find that, while male college students had higher odds of consuming alcohol, female alcohol prevalence was higher both in primary school and in street children. These findings caution against wholesale statements about gender differences in alcohol consumption. In yet another systematic review conducted for medical university students in Brazil, Nascimento et al. (Reference Nascimento, JdS and CAFd2022) also find gender-specific differences, although they do not study other sociodemographic correlates of alcohol use. Gender-specific differences are also noted in reviews conducted for university populations in industrialized economies like the UK and Ireland (Martens et al., Reference Martens, Dams-O’Connor and Beck2006; Davoren et al., Reference Davoren, Demant, Shiely and Perry2016), but Davoren et al. (Reference Davoren, Demant, Shiely and Perry2016) argue in their review that the gap may be narrowing.
Our study has some limitations that need to be borne in mind. A consistent limitation is the studies’ cross-sectional design, which only provides a snapshot of behaviors and associated factors at a single point in time, making it impossible to determine whether a factor, such as peer influence or academic performance, precedes or follows alcohol consumption. Hence, there is a need for longitudinal studies to better understand the temporal relationships and potential causal pathways between these factors and alcohol use among college students in ASEAN countries. Even for observational studies, a diverse range of designs (including cohort and case–control) may help provide deeper insights into alcohol use disorder.
Another limitation is the studies’ general reliance on self-reported measures of alcohol consumption and related behaviors, which can introduce biases such as social desirability, recall error, or underreporting. Selection bias may also occur if students consuming alcohol are missing class, and hence had less probability of being selected for the survey. More minor concerns include heterogeneity in how questions on the survey are designed, and the lack of distinction or nuance in most studies between binge drinking, alcohol addiction and alcohol consumption. This review had to rely on the last of these categories, as it was most commonly reported, but future research could better aim to distinguish between the three. Last but not least, the generalizability of the findings should be treated with caution, because several studies included in the review used convenience samples from specific universities or cities, limiting their broader applicability.
The findings of this systematic review have implications for both clinicians and policymakers. Early identification of high-risk groups can enable the development of more tailored, preventive health measures, particularly in student health services. Considering the positive association between being a male university student and consuming alcohol, university decision-makers need to recognize the potential consequences of alcoholism among male students, such as an elevated risk of injuries, violence and academic underperformance Paul et al. (Reference Paul, Ganie and Dar2024). As prevention policies, universities could implement routine alcohol use screening for males during health checkups, incorporate brief alcohol intervention programs into campus health centers and train counselors to assess peer influence and smoking status as risk indicators, especially for males deemed to be at higher risk. In the literature, there is some precedent for each of these interventions in other contexts (Barnett and Read, Reference Barnett and Read2005; Samson and Tanner-Smith, Reference Samson and Tanner-Smith2015). Similarly, given the positive association between older age (within the university cohort) and alcohol consumption, educational campaigns geared toward younger or incoming students on the risks of alcohol consumption and resources specifically targeted toward substance abuse disorders and addictions, may yield dividends for reducing alcoholism (Grossbard et al., Reference Grossbard, Mastroleo, Geisner, Atkins, Ray, Kilmer, Mallett, Larimer and Turrisi2016).
For policymakers, the review suggests the need for region-specific public health initiatives that address the broader social influences on drinking behavior, such as peer pressure. Policymakers may consider implementing stricter regulations on alcohol advertising, mandating campus-based substance use programs and engaging parents and communities in prevention efforts. To be truly effective, such implementations would ideally be culturally tailored (Manuel et al., Reference Manuel, Satre, Tsoh, Moreno-John, Ramos, McCance-Katz and Satterfield2015). An example of one such intervention that was trialed in three Native American communities by McDonell et al. (Reference McDonell, Nepom, Leickly, Suchy-Dicey, Hirchak, Echo-Hawk, Schwartz, Calhoun, Donovan and Roll2016) is culturally adapted contingency management, defined as an “addiction intervention where participants receive reinforcers such as vouchers or prizes for providing objective evidence of drug abstinence” and was cited as an effective intervention for illicit drugs, compared to other psychosocial interventions (Lussier et al., Reference Lussier, Heil, Mongeon, Badger and Higgins2006; Prendergast et al., Reference Prendergast, Podus, Finney, Greenwell and Roll2006). However, it needed to be adapted to the specific cultural setting of the Native American reservation to be feasible. Another example of a culturally tailored intervention is peer-to-peer-based motivational interviewing for alcoholism in the context of American Indian Alaska Native women of childbearing age (Montag et al., Reference Montag, Dusek, Ortega, Camp-Mazzetti, Calac and Chambers2017), the design process of which “included various community focus groups, interviews and a final review.” While a number of such adaptations have been proposed both for Native Americans (Richer and Roddy, Reference Richer and Roddy2022), and also for specific demographic groups like Latino males (Valdez et al., Reference Valdez, Flores, Ruiz, Oren, Carvajal and Garcia2018), little work exists on culturally tailored interventions for tackling alcohol use disorders in the ASEAN countries, despite their promise (Jiang et al., Reference Jiang, Xiang, Hao, Room, Zhang and Wang2018).
An important aspect of this study and the formulated research objectives is that it treated alcohol consumption as an outcome only, thereby excluding the many studies where alcohol consumption was an exposure for another outcome e.g., hypertension or depression (Peltzer et al., Reference Peltzer, Pengpid, Sychareun, Ferrer, Low, Huu, Win, Rochmawati and Turnbull2017; Tuyen et al., Reference Tuyen, Dat and Nhung2019; Vo et al., Reference Vo, Nguyen, Vu, Tran and Nguyen2023; Yeo et al., Reference Yeo2024). While this was by design, both because it is necessary to scope the study and because it is generally difficult to isolate the association between alcohol and the outcome variable in the presence of other controls (if alcohol is itself an exposure), real-life policy decisions would need to consider a more holistic picture when designing interventions. This study, while not making causal inferences, suggests some of the risk factors that could be taken into account by policymakers and those in authority in universities (such as Deans) to reduce the harmful effects of alcohol consumption among their student body.
Conclusion
This systematic review aimed to provide a comprehensive analysis of sociodemographic factors associated with alcohol consumption among college students in ASEAN countries. Through the inclusion of 15 cross-sectional studies, key variables such as gender, age, parental alcohol consumption, peer influence and smoking behavior emerged as significant correlates of alcohol use. The review also identified gaps in the literature, including underexplored sociodemographic variables. As countries across the Global South experience rapid growth in their university populations, region-specific insights into alcohol consumption are essential for designing effective public health policies that reduce future healthcare burdens. By highlighting ASEAN-specific risk factors, the review’s findings contribute to the global understanding of youth alcohol consumption and can inform both local and international intervention strategies, aligning with the United Nation’s SDG 3 on reducing harmful substance use worldwide. Future research should further integrate regional specificity with global perspectives, promoting cross-national comparisons that strengthen evidence-based public health policy.
Open peer review
To view the open peer review materials for this article, please visit http://doi.org/10.1017/gmh.2025.10027.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/gmh.2025.10027.
Data availability statement
Data availability is not applicable to this article as no new data were created or analyzed in this study.
Author contribution
M.K. is responsible for all aspects of this study.
Financial support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Competing interests
The authors declare none.
Ethics statement
As this is a systematic review, it did not involve any direct data collection from human subjects.
Open access
Comments
Dear editors,
I would like to submit a standard review titled Alcohol consumption among university students in ASEAN countries: a systematic review and meta-analysis for consideration in Cambridge Prisms: Global Mental Health. Alcohol consumption among university students poses significant public health challenges, especially in the Association of Southeast Asian Nations (ASEAN) region where limited research exists. According to the World Health Association, the burden of such non-communicable diseases and addictions in emerging economies has been sharply rising in recent decades. More importantly, such behaviors are modifiable and preventable. Focusing specifically on alcohol consumption among college and university students, some authors have cited it as a significant public health issue, but also note that most research focuses on North American and European populations.
In ASEAN economies, where the population is young, educated, and becoming increasingly globalized, the rise of modern health challenges like alcohol addiction among university students is a particularly growing concern. Developing a better understanding of the socio-demographic drivers contributing to alcohol consumption in this group is a well-motivated research agenda, and one that that could be critical for formulating effective and sustained strategies, policies, and interventions. It is also timely. Hence, I aim to conduct a systematic review of the socio-demographic factors associated with alcohol consumption among university students in ASEAN countries.
The specific objectives are to: (a) determine which sociodemographic factors have been positively or negatively associated with alcohol consumption in college students in the ASEAN countries, (b) compare these studies in terms of quality, rigor, and key findings, and (c) conduct a meta-analysis of socio-demographic factors found to be most commonly studied.
The review and meta-analysis is backed by fifteen peer-reviewed research studies that were gathered by following a systematic PRISMA search protocol. I hope that this submission is of interest, and look forward to hearing from you.
Best regards
Mayank Kejriwal
Principal Scientist and Research Assistant Professor
University of Southern California