Basic demographic characteristics according to use of illicit drugs are shown in Table 1. Ever drug users had several different characteristics compared with non-users: they were more likely males, born in urban areas, living in students’ dormitory or alone (in rented apartment), had higher household monthly income, had own income, had higher maternal education level, were smokers and alcohol users. These students also had more chronic diseases and reported higher BDI. A total of 82 students out of 514 (16.0%) reported use of illicit drugs. Frequency of illicit drug use is given in Table 2. The most frequently used drug was marijuana (9.3%). The least frequently reported drug was methadone (0.2%). Males were more likely to have used marijuana (p = 0.001), hashish (p = 0.009), ecstasy (p = 0.008), amphetamines (p = 0.008), and cocaine (p = 0.001).

Variable	Illicit drug use		P-value
	Never (N = 432)	Ever (N = 82)
Age (years)	20.8 ± 1.6	20.8± 1.9	0.725
Gender			0.052
• male	192 (44.4)	46 (56.1)
• female	240 (55.6)	36 (43.9)
Place of birth			0.003
• urban	263 (60.9)	64 (78.0)
• rural	169 (39.1)	18 (22.0)
Region of origin			0.518
• Serbia	126 (49.0)	27 (56.3)
• North of Kosovo	70 (27.2)	13 (27.1)
• Enclave	61 (23.7)	8 (16.7)
• Missing	175 (40.5)	34 (41.4)
Type of current residence			0.001
• home (with parents)	120 (27.8)	20 (24.4)
• students’ dormitory	162 (37.5)	27 (32.9)
• alone (in rented apartment)	143 (33.1)	27 (32.9)
• other	7 (1.6)	8 (9.8)
Family monthly income (in Euros)	515 ± 357	805 ± 811	0.004
Having own income			0.001
• yes	256 (59.3)	65 (79.3)
• no	176 (40.7)	17 (20.7)
Mother’s education level			0.001
• higher	96 (22.2)	34 (41.5)
• other	336 (77.8)	48 (58.5)
Father’s education level			0.280
• higher	142 (32.9)	32 (39.0)
• other	290 (67.1)	50 (61.0)
Type of Faculty			0.868
• social science and humanities	121 (28.0)	22 (26.8)
• medical sciences	198 (45.8)	40 (48.8)
• natural sciences and mathematics	11 (2.5)	3 (3.7)
• technology and engineering science	102 (23.6)	17 (20.7)
Grade point average	7.8 ± 0.9	7.9 ± 0.7	0.447
Smoking			0.001
• yes	70 (16.2)	46 (56.1)
• no	362 (83.8)	36 (43.9)
Alcohol use			0.001
• yes	281 (65.0)	70 (85.4)
• no	151 (35.0)	12 (14.6)
Physical activity			0.285
• yes	347 (80.3)	70 (85.4)
• no	85 (19.7)	12 (14.6)
Having chronic diseases			0.001
• yes	34 (7.9)	21 (25.6)
• no	398 (92.1)	61 (74.4)
Beck Depression Inventory score	2.8 ± 4.8	8.8 ± 9.5	0.001

Type of illicit drug	Count (%) N = 514	Males (%) N = 238	Females (%) N = 276	P-value
Marijuana	48 (9.3)	33 (13.9)	15 (5.4)	0.001
Bromazepam	39 (7.6)	15 (6.3)	20 (7.2)	0.672
Diazepam	22 (4.3)	7 (2.9)	15 (5.4)	0.164
Hashish	12 (2.3)	10 (4.2)	2 (0.7)	0.009
Cocaine	10 (1.9)	10 (4.2)	0 (0.0)	0.001
Ecstasy	6 (1.2)	6 (2.5)	0 (0.0)	0.008
Amphetamines	6 (1.2)	6 (2.5)	0 (0.0)	0.008
LSD	6 (1.2)	5 (2.1)	1 (1.4)	0.067
Heroin	3 (0.6)	3 (1.3)	0 (0.0)	0.061
Tramadol	2 (0.4)	1 (0.4)	1 (0.4)	0.916
Methadone	1 (0.2)	1 (0.4)	0 (0.0)	0.281

Table 3 summarizes the results of univariate and multiple logistic regression analysis of factors associated with ever illicit drug use among students the University of Pristina temporarily settled in Kosovska Mitrovica. All statistically significant and marginally significant variables entered the multiple model. Factors associated with use of illicit drugs among students were: cigarette smoking and drinking alcohol, having chronic diseases and having higher depression score (Table 3). According to a similar multiple model, predictors of marijuana use were: being male, not having own income, being smoker and alcohol user, having chronic diseases and having higher score on BDI (Table 4).

Variable	Univariate logistic regression analysis			Multiple logistic regression analysis
	OR	95% CI	P-value	OR	95% CI	P-value
Age	1.01	0.87–1.16	0.922
Gender male vs. female	0.63	0.39–1.01	0.054	0.54	0.26–1.14	0.107
Place of birth urban vs. rural	0.44	0.25–0.76	0.004	0.98	0.46–2.08	0.953
Region of origin Serbia vs. North Kosovo and Enclaves	0.80	0.54–1.19	0.262
Current residence with parents vs. other	0.83	0.49–1.45	0.528
Household monthly income	n/a
Having own income yes vs. no	2.63	1.49–4.64	0.001	1.01	0.38–2.64	0.991
Mother’s education level higher vs. other	0.40	0.25–0.66	0.001	0.75	0.34–1.67	0.485
Father’s education level higher vs. other	0.77	0.47–1.24	0.285
Type of faculty medical science vs. other	0.89	0.55–1.42	0.624
Grade point average	1.10	0.79–1.53	0.564
Smoking yes vs. no	6.61	3.99–10.96	0.001	3.66	1.77–7.60	0.001
Alcohol use yes vs. no	3.13	1.65–5.97	0.001	2.88	1.08–7.69	0.035
Physical activity yes vs. no	1.43	0.74–2.76	0.287
Having chronic diseases yes vs. no	4.03	2.20–7.40	0.001	2.96	1.21–7.25	0.017
Beck Depression Inventory score	1.13	1.09–1.17	0.001	1.10	1.04–1.16	0.001

Variable	Univariate logistic regression analysis			Multiple logistic regression analysis
	OR	95% CI	P-value	OR	95% CI	P-value
Age	0.94	0.79–1.11	0.465
Gender male vs. female	2.80	1.48–5.30	0.002	3.58	1.61–7.95	0.002
Place of birth urban vs. rural	1.43	0.74–2.75	0.277
Region of origin Serbia vs. North Kosovo and Enclaves	1.06	0.65–1.71	0.807
Current residence with parents vs. other	1.97	0.90–4.33	0.089	1.45	0.59–3.57	0.420
Household monthly income	n/a
Having own income yes vs. no	0.21	0.09–0.51	0.001	0.25	0.08–0.75	0.013
Mother’s education level higher vs. other	2.31	1.25–4.27	0.007	1.76	0.84–3.69	0.136
Father’s education level higher vs. other	1.59	0.87–2.91	0.131
Type of faculty medical science vs. other	1.12	0.61–2.04	0.710
Grade point average	1.13	0.75–1.70	0.563
Smoking yes vs. no	0.12	0.65–0.23	0.001	0.20	0.10–0.43	0.001
Alcohol use yes vs. no	0.28	0.17–0.67	0.005	0.34	0.12–0.98	0.046
Physical activity yes vs. no	0.47	0.18–1.22	0.124
Having chronic diseases yes vs. no	0.35	0.17–0.74	0.006	0.52	0.18–1.46	0.014
Beck Depression Inventory score	0.92	0.88–0.95	0.001	0.94	0.90–0.99	0.015

The HRQoL scores according to ever illicit use are presented in Table 5. Ever illicit drug users reported all domains of HRQoL as worse. Similarly, these students had worse Physical and Mental Composite Scores as well as the total HRQoL score (Table 5).

Scales of SF-36	Illicit drug use		P-value
	Never (N = 432) Mean ± SD	Ever (N = 82) Mean ± SD
Physical functioning	93.5 ± 17.0	89.1 ± 20.7	0.040
Role physical	88.6 ± 23.1	72.0 ± 35.4	0.001
Pain	87.4 ± 16.1	77.2 ± 23.3	0.001
General Health	75.2 ± 16.4	70.4 ± 19.3	0.019
Vitality	70.6 ± 18.9	59.0 ± 24.3	0.001
Social Functioning	83.1 ± 20.9	71.6 ± 26.2	0.001
Role Emotional	83.2 ± 32.5	58.5 ± 40.8	0.001
Mental Health	76.4 ± 18.3	62.7 ± 23.5	0.001
Physical Composite Score	83.1 ± 12.1	73.5 ± 18.0	0.001
Mental Composite Score	77.7 ± 15.7	64.5 ± 20.9	0.001
Total Score	82.2 ± 13.5	70.1 ± 19.1	0.001

Consistent with previously reported prevalence [8], the results of our survey provide evidence that one in six students in North Kosovo have ever used illicit drugs. In an attempt to quantify illicit drug use various authors reported substantially different prevalence estimates [3–7, 23]. These differences may be attributable to inclusion of various substances, which may not be illicit per se in majority or regions worldwide (e.g. alcohol or tobacco) [5, 6, 23] or are not covered by questionnaire batteries (e.g. steroids, inhalants) [4]. Therefore, a comparison of overall prevalence of illicit drug use may not be adequate. Moreover, Shamsipour et al. [24] suggested that conventional methods of direct questioning by self-report in assessment of illicit drug use may yield quite lower prevalence when compared with a ‘crosswise model’ technique applied several months afterwards. Even though analyzing drug metabolites in biological specimens may be the most accurate method to determine the extent of illicit drug use, their applicability in population-based cross-sectional studies is limited.

In our survey illicit drug users most commonly reported use of marijuana (9.3%). These findings are in accordance with those reported in previous studies [4, 10]. However, prevalence of marijuana use was lower compared with college students in the US for example, where lifetime prevalence of 30% at college entry [3] or 60% among sophomores were observed [4]. In other regions, such as Iran or India, the prevalence of marijuana use of 7.4% and 6.8% respectively, has been estimated [5, 6]. Peer use of marijuana seems to be associated with initiation of marijuana use among non-users [25]. Studies have shown that use of marijuana may predict use of other illicit drugs, such as hallucinogens or ‘hard drugs’ [26], which may have other important public health implications.

Furthermore, misuse of tranquilizers, ranked second among our students, has also been recognized in a college setting [4, 27]. A recent study on prescription drug misuse reported that 25% of college students use these drugs without physician’s prescription [27]. Analysis of motives for prescription drug misuse suggested that most common reasons include self-treatment intentions, recreation or both [28]. At the same time, one half of college students who misuse prescription drugs are screened positive for drug abuse [28]. A small percentage of students in our survey reported the use of ecstasy and amphetamines. Ecstasy and amphetamines are commonly labelled as “club drugs”. Their psychoactive effects include enhanced feelings of closeness with others and sensory perception [10]. Sim et al. [29] observed that those students who recreationally use ecstasy are more likely to initiate illicit drug use at younger age and engage in binge drinking and marijuana use. Amphetamines, on the other hand, could be used with the goal to improve academic performance and concentration [30]. Cocaine use was reported by 1.9% of students in our survey. By contrast, cocaine use was ranked second in terms of both lifetime and past-month use among illicit drug users in a US college [4]. In addition, in the US, 13% of students reported cocaine use by fourth year of college [31]. On the other hand, in a survey among male students in India and among undergraduate students in Ethiopia no use of cocaine, amphetamines, sedatives or heroin was reported [6, 32]. These remarkable variations might be a result of differences in definition of time frame for which illicit drug use was assessed: illicit drug use during the past month [6, 31] versus lifetime prevalence of illicit drug use in our study. Also, these variations could be explained by differences in financial means for purchase, as it has been suggested that the trajectory of cocaine and marijuana use, both in college and in general population, is being influenced by changes in price [33].

We observed that students who smoked cigarettes and drank alcohol were more likely to be illicit drug users. Our results are in line with the existing literature data [3, 5, 6, 33, 34]. In addition, male students were more likely to use marijuana. Male sex, smoking and alcohol consumption have been consistently associated with illicit drug use [3, 5, 6, 33–36]. These factors are likely related to particular behavioral patterns, as studies have shown that male college students are more likely to exert high-risk behaviors [35, 36], which could be a result of gender-role stereotypes and the socialization process [37]. Specifically, Davies et al. suggested that boys are encouraged from early childhood to engage in riskier behaviors that are deemed as display of strength and fearlessness based on stereotyping and traditional notions of masculinity [37], which might explain a higher propensity towards risk-taking among males irrespective of ethnicity, religion or socio-economic background.

In this survey, we also observed that ever illicit drug users were more likely to have chronic diseases and a higher level of depression. It has been estimated that around 30% of University students have chronic diseases and/or special care needs [8]. In management of health conditions, use of illicit substances might complicate the diagnosis and treatment [8]. There are some indices in the available body of literature that high school students with asthma are more likely to use illicit drugs [38], which corresponds to our findings. However, illicit drugs use remains understudied topic in the population of University students with chronic diseases.

We also observed that higher BDI score was associated with illicit drug use in our sample of students. Due to our cross-sectional study design we cannot determine whether being depressed lead to use of illicit drugs or vice versa. Nevertheless, it is important that University students are offered screening for depression as part of the periodical general health status assessment during schooling. In this way, vulnerable individuals could be assessed for illegal drug use also. Subsequently, these students could be offered support and counseling with the goal to reduce, minimize or alter their risky behavior. A study of personal concerns of college students revealed a wide spectrum of negative consequences that college student experience due to drug use [4]. The highest proportions of lifetime negative consequences accounted for having said or done something embarrassing (46%), felt guilty and ashamed (45%), not having done their academic duties or received a lower grade (44%) and feeling bad physically (43%) [4]. Those students who reported ever illicit drug use had worse scores in all HRQoL domains. Similarly, physical and mental composite scores as well as the overall HRQoL were worse in students who have ever used illicit drugs. Studies exploring HRQoL among University students are scarce [39]. Caldeira et al. [39] followed a cohort of 1,253 college students throughout six years, starting from their freshmen year. Over this period, out of six subgroups describing marijuana-use frequency in college students, chronic marijuana users had worse physical and psychological health [39]. Lower HRQoL has also been identified among adolescents who engage in risky behaviors as opposed to those students who do not [40]. To explain this finding, Topolski et al. [40] suggested that factors such as relationship with parents might mediate the relation between illicit drug use and worse HRQoL. This means that adolescents who engaged in risky behaviors were less likely to get along with their parents [40]. Parental monitoring and vigilance concerning their adolescent's activities and friends could be seen as proxy of getting along with parents. Taking this into consideration, a higher level of parental monitoring and supervision during senior year in high school has been associated with having less drug-using peers, which was, in turn, associated with decreased illicit drug exposure in college [25].

Information bias should be taken into account as a limitation of this study, because data on illicit drug use, smoking, alcohol and physical activity were self-reported. Thus, it is possible that data collected in this study were under-reported. The cross-sectional study design precludes us from drawing the direction of inference, and therefore associations could only suggest potential associations and not determine direction of causality.