Does academic streams influence alcohol use in colleges?

BACKGROUND: Alcohol use among college students is a major public health priority owing to its high prevalence and numerous negative outcomes. Most interventions targeting alcohol use among college students consider them as a homogenous entity. There is preliminary evidence from high-income countries that patterns of alcohol use differ across academic streams. This remains unstudied in India. AIMS: To compare the prevalence and correlates of alcohol use among students enrolled in various collegiate educational streams (medical, nursing, engineering, arts and science, and others [law/fisheries]) in the state of Kerala, India. SETTINGS AND
DESIGN: This is a cross-sectional survey conducted among college students.
MATERIALS AND METHODS: 5784 students completed a self-administered questionnaire assessing alcohol use and its correlates in the psycho-social domains. STATISCAL ANALYSIS: Lifetime prevalence and severity of alcohol use was determined across examined academic streams. A multivariable logistic regression analysis was done separately for each course, to identify factors influencing alcohol use.
RESULTS: The lifetime prevalence of alcohol use varied between 10.6% among nursing students to 41.7% among students pursuing "other" stream (law/fisheries). Students pursuing medicine and nursing had a relatively lower proportion of hazardous users. Consistently across all academic streams, being male and using tobacco increased the risk, while those from the Muslim community had a lower risk of alcohol use. Other examined psychosocial correlates showed varying relationship across courses.
CONCLUSION: The prevalence and psychosocial correlates of alcohol use vary among students pursuing various academic streams. This finding has public health importance as the incorporation of course level characteristics in intervention programs will improve effectiveness. Copyright:

INTRODUCTION

Alcohol use among college students is a key public health priority, given its adverse health and academic outcomes. Among young people, it accounts for more than 10% of the overall burden of injury and disease.[1] Students with problematic alcohol use have a higher risk for poor academic achievement, risky sexual behaviour, accidents, mental health issues, suicide, and later addiction.[234]

Studies from high-income countries (HIC) have reported that most college students drink (90%), with a significant proportion reporting “binge drinking” (25%–50%).[456] Most studies from Asia have reported lower rates of heavy drinking of around 10%.[5] Studies from India are a few and have reported widely varying prevalence of 20%–50%.[789]

These studies have examined college students as a homogenous group. However, the environment in colleges can no longer be considered a single culture.[10] There is anecdotal evidence suggesting that alcohol use patterns vary according to the choice of an academic field. Different workloads (having no classes on Friday) or specific gender characteristic (e.g., a higher proportion of females in nursing) is reported to influence alcohol use in campuses.[511] Also, several factors such as duration of courses, institution's predominant ethnic characteristics, having active athletic/sports program, and membership in clubs/fraternities have been inconsistently linked to alcohol use.[51213] These findings from Western campuses have limited generalizability to India. There is a lower prevalence of alcohol use in India (and other Asian countries) compared to European countries or the USA.[58] The sororities, fraternities, or special athletic clubs of the Western university systems are nonexistent in India. In contrast to the West, only a few large universities in India offer a wide variety of courses. Most Indian colleges offer courses restricted to a specialized stream (e.g., arts and science, engineering, and nursing).[14]

Despite these differences, there is preliminary evidence that variation in alcohol use across different courses may be the case in India also, with a small study of 297 male students in 4 different educational streams reporting variation of prevalence across courses.[15] Understanding the unique patterns of alcohol use among students in various academic streams has significant public mental health importance. For most young people, the period of transition to adulthood represents greater vulnerability to yield to environmental norms. Exposure to campus with a culture of drinking can thus lead to both initiation and increased alcohol use.[12] It would also be vital to understand whether the commonly-reported correlates of alcohol use in the sociodemographic and psychosocial domains such as age, gender, religion, socioeconomic status, place of residence, tobacco use, family structure, suicidal risk, and emotional distress among students using alcohol pursuing different academic streams.[5816] Incorporating the protective and risk factors of the unique academic environment of various streams of higher education may ensure improved effectiveness of interventions. The reach and impact of any college intervention is more significant as students in college represent the cream of the country's future.[14]

Hence, this study examined the differences in prevalence and correlates of alcohol use among students in college pursuing the five common streams of higher education: that is, medical, engineering, arts or science, nursing, and others (law or fisheries) in Kerala, India. This study is a part of a bigger research project which examined substance use and other psychological issues among students pursuing a college education.

MATERIALS AND METHODS

This survey was done in 58 of the 123 colleges in Ernakulam district, Kerala. The colleges were first categorized to five academic streams (medical, engineering, arts and science, arts or science, and others [law or fisheries]). The colleges were then selected by cluster random sampling with academic streams forming the clustering unit. In each of the selected colleges, the survey was conducted in two classes, either even or odd years (i.e., second and fourth year, or first and third year). The classes were allocated by the head of the institution.

All students were informed about the objectives of this survey and requested to participate. They were assured that their responses would be anonymized with nonparticipation having no impact on their academic grades. Students who gave informed consent were administered the paper and pencil questionnaire under exam conditions. Students could opt for either the English or Malayalam version of the questionnaire. The survey instrument designed in English was translated to Malayalam and back-translated to ensure it is conceptually congruent. The students were supervised by the faculty of the Psychiatry Department, Government Medical College, Ernakulam.

Information on sociodemographics was collected using a checklist (age/sex/religion/socioeconomic status/residence). In addition, the following structured instruments were used.

Alcohol

Alcohol, Smoking, and Substance Involvement Screening Test (ASSIST) was employed to evaluate alcohol use.[17] In addition to lifetime alcohol use, among users, the ASSIST assesses patterns of use, frequency of urges, problems experienced, concerns raised by friends/family regarding alcohol, and attempts to reduce or stop use. The total score as per ASSIST is the “alcohol involvement score.” An alcohol involvement score of 0–10 indicates low-risk user or abstainer (those using alcohol occasionally or not currently using, with a less likelihood of experiencing harm, currently or the future, with the same pattern of use); score of 11–26 indicates hazardous use (pattern of consumption that increases the likelihood of adverse effects, physical, psychological or social); and score 27+ indicates dependence (has difficulty in quitting despite serious problems in various domains). ASSIST has been validated as a screening tool in developing countries and has good test–retest reliability.[17] The study also evaluated the consumption of tobacco but for this paper, lifetime use of tobacco has been considered.

Psychological distress

Psychological distress in the past 2 weeks was assessed using Kessler's Psychological Distress Scale (K10) scale.[18] The scale has ten questions about anxiety and depressive symptoms scored on a 4-point Likert scale. Higher scores suggest greater distress. K10 is validated for the screening of psychological distress in India.[19]

Suicidality

Lifetime suicidality was evaluated by the following screening questions: “Have you ever thought of committing suicide in your life?” and “Have you made a suicidal attempt in your lifetime?”

Ethics

This study received clearance from the Institute Review Board of Government Medical College, Ernakulam.

Statistical analysis

Statistical analyses were conducted by using R software.[20] Lifetime prevalence and alcohol use severity was calculated among students in the examined academic streams. The association of alcohol users and nonusers with sociodemographic variables in different courses was examined using the Chi-square test for categorical and two-way ANOVA with Bonferroni correction for continuous variables. Multivariable logistic regression analysis (full model) was used to examine the relationship of alcohol use with psychosocial variables for each of the courses separately. Students in nursing, however, had low baseline prevalence of tobacco use (n – 3 [0.5%]) and hence tobacco was excluded in the full model examining nursing. The findings are reported as odds ratio and 95% confidence intervals. The statistical significance was set at P < 0.05.

RESULTS

A total of 5784 college students in five academic streams took part in the cross-sectional survey. The break-up of students surveyed and the number of institutions in each stream are described in Table 1.

Table 1

Colleges and sample surveyed for each specialised course

Courses offered	Number of institutions surveyed		Sample size as a proportion of total N (%)
	Sub-stream	Major stream	Sub-stream	Major stream
Medical		8		986 (17)
Allopathy	2		413 (7.1)
Ayurveda	2		208 (3.6)
Homoeopathy	1		149 (2.6)
Dental	3		216 (3.7)
Nursing		8		620 (10.7)
Arts (including Undergraduate and Postgraduates)		25		2436 (42.1)
Engineering (including degree and diploma students)		14		1310 (22.6)
Degree	11		1018 (17.6)
Diploma	3		292 (5)
Others		3		432 (7.6)
Law	2		346 (6)
Fisheries	1		86 (1.6)
Total	58		5784

Of the total of 5784 students surveyed, 342 questionnaires were discarded as it had substantial missing responses providing 5442 (94.1%) questionnaires for analysis. The mean age of the students was 19.8 ± 1.4 years with a higher proportion of females (n – 3545 [65.1%]). This reflects the enrollment pattern in colleges of Kerala.[21]

The lifetime prevalence of alcohol use varied between 10.6% among nursing students to 41.7% among students pursuing “other” stream (law/fisheries). Male students reported prevalence between 15% in nursing to 55.4% among “others.” Among female students, the lowest prevalence was among students pursuing arts and science (9.3%) and highest among students in “other” courses (32.5%). Male students had a significantly higher prevalence of alcohol use when compared to females in all examined streams (P < 0.001). Male students enrolled in nursing was too low (n – 20 [3.3%]) hence gender association could not be examined [Table 2].

Table 2

Prevalence and severity of alcohol use among students across various courses

Streams	Lifetime prevalence
	Male, n (%)	Female, n (%)	Total, n (%)
Medical (n=960)	50 (27.2)	107 (13.8)	157 (16.4)
Nursing (n=603)*	3 (15)	61 (10.5)	64 (10.6)
Engineering (n=1253)	307 (40.7)	62 (12.4)	369 (29.4)
Arts and science (n=2209)	275 (35.7)	134 (9.3)	409 (18.5)
Others (n=417)	93 (55.4)	81 (32.5)	174 (41.7)

Figures in bold - Male versus female indicate P<0.001. *Gender differences not examined as males enrolled in nursing was low (n=20; 3.3%)

The severity of alcohol use calculated by mean ASSIST Scores after Bonferroni correction showed severity was comparable among those pursuing nursing and medical. The severity of alcohol use among those pursuing arts and science, engineering, and other streams was significantly higher compared to those in nursing (P < 0.05). Hazardous use was lowest among alcohol users in the nursing stream (4.7%) followed by medical (12.6%) and arts and science (17.6%). Students in engineering (20.2%) and “others” (26.4%) had the highest proportion of hazardous users. There was a significant association between student course and ASSIST categories (P < 0.001) [Table 3]. As the mean age of the study participants was low, the proportion of dependent users was low across all academic streams.

Table 3

Severity among life time users (as per alcohol, smoking and substance involvement screening test scores)

Streams	Mean ASSIST score (SD)	Low risk users, n (%)†	Hazardous users, n (%)†	Dependent users, n (%)†
Medical	3.73 (5.52)	136 (87.1)	20 (12.3)	1 (0.6)
Nursing	1.39 (3.83)	61 (95.3)	3 (4.7)	0 (0.0)
Engineering	5.69 (6.24)	293 (79.3)	74 (20.2)	2 (0.5)
Arts and science	5.25 (6.53)	332 (81.1)	72 (17.6)	5 (1.2)
Others	6.79 (6.70)	125 (71.8)	46 (26.4)	3 (1.7)

†As a proportion of total users. SD – Standard deviation; ASSIST – Alcohol, Smoking and Substance Involvement Screening Test

Alcohol users among students pursuing engineering and arts and science streams were significantly older. Across all streams, students from the Muslim community were significantly less likely to use alcohol. Students from urban areas pursuing medical and “other” streams and those from rural background pursuing arts and science were significantly more likely to consume alcohol [Table 4].

Table 4

Socio-demographic correlates of alcohol users in various courses

	Medical, n (%)	Nursing, n (%)	Engineering, n (%)	Arts and science, n (%)	Others, n (%)	Total sample, n (%)
Age years (mean±SD)	20.6±1.5	19.6±1.3	19.4±1.3	20.0±2.1	19.3±1.6	19.8±1.4
Family structure
Living with parents	137 (87.3)	54 (84.4)	336 (90.1)	363 (88.4)	151 (86.8)	4815 (88.1)
Single parent family	9 (5.7)	2 (3.1)	15 (4.0)	22 (4.2)	12 (6.9)	297 (5.7)
Living with relatives/others	11 (7.0)	8 (12.5)	18 (5.1)	24 (7.4)	11 (6.3)	330 (6.2)
Religion
Hinduism	73 (46.5)	9 (14.1)	200 (53.8)	215 (52.5)	122 (71.3)	2502 (46.9)
Christianity	80 (51.0)	55 (85.9)	155 (42.2)	177 (43.4)	40 (22.8)	1879 (33.2)
Islam	4 (2.5)	0 (0)	14 (4.0)	17 (4.2)	12 (6.0)	1061 (19.9)
Socioeconomic status
APL (above poverty line)*	150 (95.5)	52 (81.3)	326 (87.7)	321 (78.6)	164 (94.8)	4510 (83.0)
BPL (below poverty line)*	7 (4.5)	12 (18.7)	43 (12.3)	88 (21.4)	10 (5.2)	932 (17.0)
Residence
Urban	85 (54.1)	19 (29.7)	164 (44.5)	186 (45.3)	144 (82.8)	2307 (42.3)
Rural	72 (45.9)	45 (70.3)	205 (55.5)	223 (54.7)	30 (17.2)	3135 (57.7)

Figures in bold indicate P<0.05 for comparison of alcohol users versus nonusers in examined variable/course. *Socioeconomic indicators of Government of India

In the full model of logistic regression examining the psychosocial correlates of alcohol use, college students using alcohol in the medical stream had higher odds of academic failures, psychological distress, and tobacco consumption; those pursuing nursing reported suicidal thoughts; those in engineering reported academic failures, tobacco use, and suicidal thoughts; those in arts and science reported tobacco use, psychological distress, and suicidal thoughts; and in “other” streams tobacco use and psychological distress [Table 5].

Table 5

Psychosocial correlates of alcohol use

Total sample, n (%)	Medical		Nursing		Engineering		Arts and science		Others
	n (%)	OR (95% CI)	n (%)	OR (95% CI)	n (%)	OR (95% CI)	n (%)	OR (95% CI)	n (%)	OR (95% CI)
Part-time job (n=441; 8)
Nonuser	15 (1.8)	1.0	2 (0.4)	-	66 (7.4)	1.0	191 (10.2)	1.0	10 (4.1)	1.0
User	9 (5.7)	2.03 (0.70-5.76)	0	-	61 (16.3)	1.33 (0.80-2.22)	76 (18.6)	1.02 (0.69-1.51)	11 (6.3)	2.71 (0.92-7.92)
Academic failures (failed in a subject) (n=973; 17.7)
Nonuser	175 (21.6)	1.0	192 (35.7)	1.0	238 (26.8)	1.0	64 (3.4)	1.0	4 (1.6)	1.0
User	59 (37.6)	2.10 (1.36-3.23)	28 (43.8)	0.99 (0.55-1.77)	168 (44.8)	1.64 (1.18-2.29)	28 (6.8)	0.82 (0.41-1.63)	17 (9.8)	2.44 (0.5-11.3)
Lifetime tobacco use (n=485; 8.8)
Nonuser	10 (1.2)	1.0	1 (0.2) *	-	22 (2.5)	1.0	48 (2.6)	1.0	10 (4.1)	1.0
User	31 (19.7)	27.19 (10.51-70.3)	2 (3.1)	-	136 (36.3)	38.09 (17.8-80.91)	140 (34.2)	18.16 (10.6-31.16)	85 (48.9)	35.21 (13.5-91.9)
Psychological distress
Nonuser (mean±SD)	17.85±7.70	1.0	17.98±8.36	1.0	16.98±7.61	1.0	16.20±6.87	1.0	18.48±7.18	1.0
User (mean±SD)	20.68±9.22	1.03 (1.01-1.06)	22.83±8.69	1.02 (0.98-1.06)	18.06±8.74	1.01 (0.99-1.03)	18.76±8.40	1.03 (1.01-1.04)	20.14±7.43	1.04 (1.01-1.08)
Suicidal thoughts (n=1188; 21.5)
Nonuser	168 (20.7)	1.0	125 (23.2)	1.0	149 (16.8)	1.0	334 (17.8)	1.0	57 (23.5)	1.0
User	49 (31.2)	1.20 (0.72-2.01)	32 (50.0)	2.15 (1.11-4.16)	101 (26.9)	2.04 (1.36-3.06)	114 (27.9)	1.47 (1.03-2.10)	59 (33.9)	1.04 (0.54-2.01)
Suicidal attempts (n=227; 4.1)
Nonuser	25 (3.1)	1.0	20 (3.7)	1.0	28 (3.1)	1.0	62 (3.3)	1.0	12 (4.9)	1.0
User	17 (10.8)	2.01 (0.88-4.52)	11 (17.2)	1.91 (0.74-4.90)	9 (2.4)	0.50 (0.19-1.32)	27 (6.6)	1.17 (0.61-2.24)	16 (9.2)	1.91 (0.64-5.68)

Figures in bold indicate significant OR for comparison of alcohol users vs non-users in examined variables. *OR not examined as baseline prevalence of tobacco use in nursing was 0.5%. CI – Confidence interval; OR – Odds ratio; SD – Standard deviation

DISCUSSION

This study is among the largest from India that has compared the prevalence of alcohol use amongst students in various academic streams of higher education. Our finding that the prevalence of alcohol use varies across academic streams even when students are from geographically proximal institutions with comparable sociocultural background suggests that course characteristics influence alcohol use. This finding has been reported prior though no clear patterns have emerged.[22232425] These studies are almost exclusively from Europe and North America which have a different academic and drinking culture.[22232425] Other methodological issues such as differing instruments and varying definitions also complicate the direct comparability of our results. A previous study from India suggested a similar trend but was inconclusive owing to its small size.[15] Our study did not, however, examine the possible reasons for the between course variation in prevalence. Though understudied, many factors including academic load, knowledge and attitudes toward drinking, demographic and ethnic profile of institutions, and engagement in physical activity have been reported to influence alcohol use.[13]

Among our students, relatively lower rates of alcohol use have been found in students pursuing nursing and medical streams and higher rates among those in engineering and “other streams” (fisheries/law). The overall rates of alcohol use across examined academic streams were lower than those reported from colleges in HIC.[456] Students pursuing medical and nursing also showed lower severity of alcohol use signified by both lower mean ASSIST scores and fewer hazardous users. The lower rates in academic streams where students are more likely to be knowledgeable of poor outcomes linked to alcohol use have been reported.[26] Alcohol consumption among students in the medical/nursing stream has added importance as their attitudes may influence their professional decisions about clients seeking help for alcohol-related problems.[23]

Male students in our study across all courses had a significantly higher rate of alcohol use as has been consistently reported.[45] Also, across courses, those from the Muslim community had a low rate of alcohol use. Our finding adds to the robust evidence that religiosity influences use of substances including alcohol.[5] Increased age was a risk factor for consumption of alcohol only among our students pursuing engineering and arts and sciences. Previous studies have also reported that effects of age are often blurred by course, academic year, and workload.[27] Place of residence showed an inconsistent association, with urban residence increasing risk among students pursuing medicine and other courses, and rural residence for students in arts and science. Family structure and socioeconomic category did not influence risk of alcohol use in any of the examined courses. Previous studies which have examined the association of these sociodemographic variables with alcohol use in colleges have often reported different conclusions.[513]

Use of tobacco was robustly linked to alcohol use among students in all academic streams. This finding has been consistently reported prior.[45] Academic failures were associated with alcohol use only among students pursuing engineering and medicine suggesting that academic failures in streams with greater intensity of course effort may influence alcohol use. Among our students, psychological distress and suicidality did not show a consistent relationship across all examined academic streams. Most studies report psychological distress and suicidality to be associated with substance use but this relationship is often mediated by other variables such as negative life-events, perceived social support, parent–child conflict, academic performance, and financial stability.[52829]

The study had its limitations. The findings were based on self-report and no formal interview was conducted. This study was cross-sectional. Therefore, no inference regarding the direction of causality of examined correlates can be made. Factors such as peer drinking levels, alcohol-related attitudes, and social affiliation, which are known to mediate the relationship between alcohol use and college environment, were not examined. Suicidality was assessed by limited questions. However, the strengths include assessment of a large sample across various academic streams making our findings generalizable within Kerala. Structured instruments were used where possible.

Our finding has public health implications. Exposure to a heavy drinking environment has been known to increase drinking. However, we are not suggesting that all students exposed to such norms will experience alcohol use. A multitude of psychological (e.g., distress), social (e.g., religion), and individual factors (e.g., impulsivity) have been shown to moderate the influence of the college environment.[30] Currently, college-based interventions in India consider college students as sharing a single homogenous culture. Our study suggests that this may not be the case, and intervention programs need to take into account the environment, student campus culture, and various individual factors to improve effectiveness. Future research needs to focus on the development of integrated intervention models, which incorporates all these domains.

CONCLUSION

Our study suggests that alcohol use and its psychosocial correlates vary among students pursuing various academic streams. Hence intervention strategies need to be refocus to target academic streams with higher prevalence.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.