INTRODUCTION

Smoking-related illnesses have been a significant contributor to preventable diseases and deaths in Malaysia over the past few decades1. This is expected to continue given the current trend, i.e. plateau in adult smoking prevalence, with small reduction in teen smoking prevalence over the past decade2. Studies have shown that smoking is a behavior initiated during adolescence3. Adolescents display evidence of addiction at much lower levels of tobacco consumption compared to adults, and smoking initiated in adolescence results in high level of nicotine addiction and lower quit success rates4,5. Those who do not smoke as adolescents generally do not start smoking as adults5. Therefore, reducing the incidence of smoking among adolescents is one of the long-term measures to reduce the prevalence of smoking in the population6.

Following the introduction of some tobacco products such as shisha (waterpipe or hookah) and e-cigarettes in the market, efforts to reduce smoking among the population in the country have become more difficult. Available in an array of flavors, colors and packaging, they have become a big attraction among youths7,8. As a result, there has been promotion of tobacco products among smokers, non-smoking adolescents’ trial and initiation of smoking, and a potential increase in the number of dual (smokers using of two types of tobacco product) or poly tobacco users (using three or more types of tobacco products). Globally, it is estimated that more than 20% of smokers aged ≥15 years use cigarettes and at least one other tobacco product concurrently9. In the United States, it was estimated that 4.6 million middle and high school students were current users of tobacco products (i.e. have used a tobacco product in the past month) in 2014, 2.2 million of whom used at least two types of products10. Studies showed poly users were more likely to be male11, Caucasian11, had more exposure to and were more receptive to tobacco advertising12, more exposed to secondhand smoke and have lower perception of harm of smoking and SHS13. However, with respect to urban and rural areas there was no significant association between locality with dual/poly tobacco product use14.

Studies have found that smokers who use two or more tobacco products are more vulnerable to nicotine addiction15. Users of two or more types of tobacco products were also found to have a higher risk of relapse 30-days post-cessation compared to single users15, especially among adolescents16, as well as higher risk of using other substances such as alcohol, cannabis, marijuana or other illicit drugs17. Dual/poly tobacco product use is linked to increased exposure to toxicants and carcinogens18. As a result, dual/poly tobacco use may subsequently result in an increased risk of tobacco-related morbidity and mortality.

Although studies on cigarette and other alternative products in the country are abundant, data on dual/ poly tobacco use among adolescents are lacking as previous studies reported the prevalence and factors related to cigarette smoking3 and e-cigarettes19 use individually. This information is crucial to support policy efforts and interventions to curb the use of tobacco products among adolescents in the country. The aim of this study is to determine the prevalence and factor(s) associated with dual/poly tobacco use among school going adolescent in Malaysia.

METHODS

Sample and data

We analyzed data from the Tobacco and E-Cigarette survey among Malaysian Adolescents (TECMA) conducted in 2016, which was a large nationally representative survey that collected data on tobacco smoking, e-cigarette use, and shisha use, among public primary and secondary school-going adolescents aged 11–19 years. Multistage cluster sampling was used to select a sample of students aged 11–19 years. Fifteen states in Malaysia were the first strata, while urban and rural areas within each state in Malaysia were the second strata. Schools were the primary sampling unit selected, which were based on the proportion to the school’s enrollment size for the survey year and the latest sampling frame provided by the Ministry of Education Malaysia. The sample selection was followed by the selection of classes in selected schools using the simple random sampling method. Finally, all the school-going adolescents in the selected class were invited to participate in the study. The sample size was 13980 based on the estimated prevalence of 3% of e-cigarette users, type one error (alpha of 5%), design effect of 3 to cater for clustering effect, a tolerable error of 1.5% and expected non-response rate of 20%.

Data collection

Only students with written consent from parents were included as participants in the study. Data collection was carried out in a designated area identified by school administrators during the school day. No school teachers or staff was present during the data collection to avoid the Hawthorne effect, i.e. the tendency of research subjects to act atypically as a result of their awareness of being studied. To ensure response validity, the questionnaires were self-administered and assurances of respondent anonymity and data confidentiality were given. Trained research team members briefed the respondents in detail about the items in the questionnaire, and team members were at hand to provide clarification on any of the items when needed. Respondents were also told that their participation was voluntary and could skip any of the items in the survey. In addition to obtaining guardian/ parental consent, selected respondents signed an assent form before participating in the study.

Questionnaire

The questionnaire used in the TECMA study was adopted from the Global Youth Tobacco Survey (GYTS) 2003, GYTS 2009 and Global School Health Survey (GSHS) 2012 in Malaysia, and had been validated. To establish the face-validity of the questionnaire, the questionnaire was pretested on thirty students from primary and secondary schools in Kuala Lumpur to ensure that the items suited the local sociocultural context. Minor modifications were made to the questionnaire based on the feedback from the pre-test.

Dependent variables

Non-use, mono, and dual/poly use of tobacco products

Classification of non-use, mono use and poly tobacco use was based on the frequency of using three types of tobacco products (cigarettes, shisha and e-cigarettes) in the past one month. Those who used any of the products for at least one day in the past month were categorized as current users of that product. Respondents who answered ‘yes’ to items using tobacco products were assigned a value of 1, while those who answered ‘no’ were assigned a value of 0. We totaled the number of tobacco products they were currently using (range: 0–3). Students were classified as: 1) non-users if they did not use any tobacco products (score=0); 2) mono users if they only used one type of tobacco product (score=1); and 3) dual/poly users if they used two or more types of tobacco products (score=2–3).

Independent variables

The independent variables in the study were sociodemographic variables, namely gender, age (i.e. aged ≤12, 13–15, and 16–19 years), ethnicity (i.e. Malay, Chinese, Indian, Bumiputra Sabah, Bumiputra Sarawak, and other), locality of schooling (i.e. urban or rural), ever been taught at school about the dangers of tobacco product, exposure to secondhand smoke (SHS) at school/home/public areas - other than home, perceived the hazardous health effects of tobacco, and ever exposed to tobacco promotion at point of sale.

Statistical analysis

Data management and analysis

Data were keyed in twice by two data entry teams (double-entry) as a quality control measure. Discrepancies between the datasets were resolved by referring to the original questionnaire. The data were cleaned, and design weights based on the study design and sampling method and post-stratification weights to adjust for non-response were calculated for each case in the dataset and applied during analysis. Descriptive statistics were used to describe the demographic characteristics of the respondents. The chi-squared test was used to determine the association between mono use, and dual/poly use, with each categorical independent variable. Multivariable multinomial regression (MMR) was conducted to determine factors associated with mono use and dual/poly use. The data were presented as non-user versus mono user, and non-user versus dual/poly user. Independent variables with p=0.25 or less that were in univariate analysis were included in the model to determine the effect of each on the dependent variable (i.e. poly user of the tobacco product) after adjusting for the influence of other independent variables. Subsequently, all two-way interactions between independent variables in the final models were also examined, p>0.05 indicated that there were no significant two-way interactions. The overall fit of the MMR model was satisfactory, with no outliers or multicollinearity. All statistical analyses were performed at the 95% confidence level using SPSS software version 24.

Definition of tobacco users

Tobacco users were defined as those who had used any tobacco product (cigarettes, e-cigarettes, or shisha) at least once in the previous 30 days.

RESULTS

Sample characteristics

The response rate was 88.7% (n=13162/14832). Participants consisted of school-going adolescents aged 11–19 years. The distribution of respondents by gender was almost equal, i.e. male 51.1% and female 48.9%. Approximately two-thirds of respondents were Malay, the remaining consisting of Chinese (13%), and Sarawak, Sabah Native and Indian ethnicity (5.4– 5.7%). More than one-third of respondents were aged ≤15 years, and more than half (54.5%) were in rural areas (Table 1).

Table 1

Characteristics of the respondents who participated in the TECMA Study 2016 among school-going adolescents in Malaysia (N=13136)

VariablesEstimated populationn%
Gender
Male1881131658251.1
Female1803629655448.9
Age (years)
≤121369393413837.2
13–151434842527838.9
16–19880523372023.9
Ethnicity
Malay2433432924366.1
Chinese477966176413.0
Indian2136747485.8
Sabah native2117815455.7
Sarawak native1995584475.4
Other1470953854.0
Locality
Urban1677958768545.5
Rural2006801544854.5

Among the students, 14.5 % were tobacco users, while 8.0 % and 6.5 % were mono and dual/poly users, respectively. A significantly higher proportion of mono and dual/poly tobacco product users was observed among males, older age, Malay, Sabah and Sarawak natives, those who perceived less harm in tobacco product usage, not taught in school and exposed to SHS at home, outside the home and at school. In addition, the proportion of ever exposure to smoking promotion at the point of sale was also higher among mono and dual/poly tobacco users (Table 2).

Table 2

Prevalence of non-tobacco product users, mono tobacco product users and dual/poly users among school-going adolescents in Malaysia who participated in the TECMA study (2016) (N=13056)

VariablesNon-tobacco product userMono tobacco product userDual/poly tobacco product userp
Estimated populationn% (95% CI)Estimated populationn% (95% CI)Estimated populationn% (95% CI)
Gender
Male1390475487773.9 (72.3–75.4)296706101315.8 (14.5–17.1)19394861210.3 (9.3–11.4)<0.001
Female1704942629095.6 (94.8–96.4)531111912.9 (2.4–3.6)25575731.4 (1.0–2.1)
Age (years)
≤121244606374590.9 (89.7–91.9)903192816.6 (5.7–7.6)344671122.5 (2.0–3.2)<0.001
13–151162810441481.0 (79.2–82.8)16016953011.2 (9.8–12.7)1118633347.8 (6.6–9.2)
16–19708001300880.4 (78.5–82.2)9932939311.6 (9.8–12.9)731933198.3 (7.2–9.6)
Ethnicity
Malay2003150767582.3 (81.0–83.5)25837394310.6 (9.7–11.7)1719096257.1 (6.3–8.0)<0.001
Chinese452681167894.7 (92.9–96.1)18369553.8 (2.6–5.5)6915311.4 (0.9–2.2)
Indian19504167991.3 (88.6–93.4)12052485.6 (4.0–7.9)8580213.1 (1.9–5.1)
Sabah native17274843881.6 (77.4–85.1)220115610.4 (7.9–13.5)17021418.0 (5.5–11.7)
Sarawak native16353036781.9 (77.6–85.6)243715512.2 (9.2–16.0)11656255.8 (3.6–8.9)
Other12701532686.3 (81.7–89.9)146403710.0 (6.8–14.4)5439623.7 (2.3–6.0)
Locality
Urban1473469865687.8 (86.9–88.9)1257906107.5 (6.7–8.3)786984224.7 (4.1–5.3)<0.001
Rural1641948451181.8 (80.3–83.3)20402759411.2 (10.0–12.4)1408253437.0 (6.1–8.0)
School-based education of tobacco harms
Yes2177172805287.5 (86.5–88.5)2007817118.1 (7.3–9.0)1089813864.4 (3.8–5.1)
No18217881282.4 (78.3–85.8)270888312.3 (9.3–15.9)11816415.3 (3.5–8.0)<0.001
SHS exposure at home
Yes1047380351375.2 (73.3–77.0)19745885214.2 (12.8–15.7)14816947910.6 (9.3–12.1)<0.001
No2067315765290.2 (89.2–91.2)1523585526.7 (5.9–7.5)713542863.1 (2.6–3.7)
SHS exposure at public areas
Yes14492295159769 (75.3–78.4)25845288513.7 (12.5–15.0)1782506359.4 (8.4–10.6)<0.001
No1864349600492.6 (91.6–93.7)913653185.1 (4.4–5.9)412731302.3 (1.8–2.9)
SHS exposure at school
Yes742137276774.0 (71.9–74.0)14303050514.3 (12.7–18.0)11746044211.7 (10.3–13.3)<0.001
No2369786839088.5 (87.5–89.5)2058386947.7 (6.9–9.6)1010253213.8 (3.2–4.5)
Saw tobacco advertisement at point of sale
Yes1134356403882.3 (80.7–83.9)14053751010.2 (9.0–11.5)1028463667.5 (6.5–8.6)<0.001
No1981061713985.9 (84.7–87.0)2092806949.1 (8.2–10.1)1166793995.1 (4.3–5.9)

Approximately 3 in 5 dual/poly tobacco users used e-cigarettes first before using other tobacco products, 26.5% used cigarettes, and 11.1% used shisha first before moving to other tobacco products. Table 3 shows that more than half of dual/poly users used e-cigarettes and cigarettes (53.4%), and more than 1 in 5 respondents used all tobacco products (i.e. cigarettes, e-cigarettes and shisha), while more than two-fifths of dual/poly users used e-cigarettes and shisha, or shisha and cigarettes.

Table 3

Type of tobacco products first used and type of tobacco product used by dual/poly tobacco users in the TECMA Study (2016) in Malaysia (N=750)

VariablesEstimated populationSample95% CI
n%LowerUpper
Type of tobacco product first used
E-cigarette13940949162.556.867.8
Cigarette5904719326.521.532.0
Shisha247556611.18.115.0
Type of tobacco product used
Cigarette and e-cigarette11794140353.748.359.1
Cigarette and shisha258369911.88.715.8
E-cigarette and shisha236828910.88.014.4
Cigarette, e-cigarette and shisha5206317423.719.428.7

Multivariable logistic regression revealed that the odds of dual/poly use were significantly higher among males (AOR=14.73; 95 % CI: 9.11–23.81), Sabah and Sarawak natives (AOR=7.41; 95% CI: 3.48–15.79; AOR=3.27, 95% CI: 1.43–7.44, respectively), aged 13–15 years (AOR=3.50; 95% CI: 2.45–5.29) and aged 16–19 years (AOR=5.99; 95% CI: 4.04–8.87), perceived less health harmful effect of smoking (AOR=2.98; 95% CI: 2.15–3.85) and never been taught the health hazard of tobacco products in school (AOR=2.24; 95% CI: 1.45–3.46). However, a comparison between mono and dual/poly users revealed that only older age groups were significantly associated with higher odds of becoming dual/poly users (Table 4).

Table 4

Multinomial logistic regression and the associated factor(s) of becoming dual/poly user among school-going adolescents aged 11–19 years who participated in the TECMA study 2016 in Malaysia

VariablesMono user vs non-userDual user vs non-user
AOR (95% CI)AOR (95% CI)
Gender
Male9.30 (6.04–10.42)14.73 (9.11–23.81)
Female (Ref.)11
Age (years)
≤12 (Ref.)11
13–151.99 (1.52–2.60)3.50 (2.45–5.29)
16–191.98 (1.50–2.67)5.99 (4.04–8.87)
Ethnicity
Malay3.75 (2.28–6.14)5.93 (3.31–10.61)
Chinese (Ref.)11
Indian1.41 (0.72–2.71)2.22 (0.96–5.16)
Sabah native3.73 (2.01–6.91)7.41 (3.48–15.79)
Sarawak native3.71 (1.98–6.94)3.27 (1.43–7.44)
Other3.82 (1.84–7.91)4.21 (1.80–49.88)
Locality
Urban (Ref.)11
Rural1.33 (1.07–1.66)1.06 (0.82–1.31)
Smoking harmful
Yes (Ref.)11
No1.76 (1.39–2.23)2.98 (2.15–3.85)
School-based education of tobacco harms
Yes (Ref.)11
No1.83 (1.34–2.49)2.24 (1.45–3.46)
SHS exposure at school
Yes1.79 (1.41–2.27)2.59 (1.95–3.43)
No (Ref.)11
SHS exposure at public areas
Yes2.03 (1.54–2.68)2.49 (1.71–3.60)
No (Ref.)11
SHS exposure at home
Yes1.90 (1.48–2.45)2.45 (1.43–3.53)
No (Ref.)11
Expose to tobacco promotion at the point of sale
Yes0.97 (0.77–1.23)1.03 (0.78–1.36)
No (Ref.)11

DISCUSSION

This is the first study that investigated the prevalence and factors associated with dual/poly tobacco use among school-going adolescents in Malaysia. The study revealed that 6.5% of school-going adolescents used dual/poly tobacco products. The prevalence is significantly lower than the 19.1% prevalence of poly use reported among high school students in the USA20, and 20% globally9. However, it is almost similar to the 5.4% reported by Aleyan et al.21 and 7.0% Cooper et al.22 among youths in Canada and Texas, USA, respectively.

Conventional cigarettes and e-cigarettes were the most common products used by dual/poly users in this study. This finding is consistent with another study23. This finding might be due to the easy accessibility to these two products24. Cigarettes are widely available, widely used and have been in the market for a long time. For instance, cigarettes are sold in sundry shops, service stations and other premises in the country. At the same time, e-cigarettes are also available in premises selling vape/e-cigarette or online, compared to shisha, which are limited and are only available in certain types of premises. Therefore, putting restriction on adolescents to prevent them from patronizing these premises may reduce their usage of shisha among adolescents.

Most dual/poly users of tobacco products used e-cigarettes first and then started using conventional cigarettes or shisha. This is consistent with findings of a longitudinal study that reported that e-cigarette users were 6.17 and 4.98 times more likely to begin using cigarettes and combustible tobacco products (including hookah, cigars, or pipes), respectively25.

Many studies among adolescents, mainly in the US, have shown a higher prevalence of dual/poly tobacco use among males22. Our study had the same finding. Among the plausible reasons is that smoking is the norm among males in Malaysia. Many studies have shown that the smoking prevalence is significantly higher among adult and youth males in Malaysia. Lim et al.26 reported that the smoking rate among adult males was significantly higher compared to females [48.2% in males (95% CI: 47.1–49.4) vs 3.4% (95% CI: 3.1–3.8) in females, in 1996, 46.4% (95% CI: 45.5–47.4) vs 1.60% (95% CI: 1.5–1.8) in 2006, 43.9% (95% CI: 41.4–44.6) vs 1.00 (95% CI: 0.7–1.6) in 2011, and 43.0% (95% CI: 41.4–44.6) vs 1.40 (95% CI: 1.10–1.18) in 2015]. In addition, comparison of male and female youth (aged 13–15 years) status of current smoking from three nationwide cross sectional studies in 2003, 2009 and 2016 revealed a similar gender disparity among Malaysian youth, i.e. significantly higher among males, i.e. 26.1% (95% CI: 19.9–33.4) vs 2.4% in females (95% CI: 1.4–4.1) in 2016, 30.9% (95% CI: 29.6–33.3) vs 6.1% (95% CI: 4.6–7.4) in 2009, and 35.5% (95% CI: 33.2–38.0) vs 4.3% (3.4–5.4) in 2003. Therefore, the use of any tobacco product or other form of tobacco product by this group is the norm. Moreover, Malaysian society still holds on to traditions wherein the social norms governing female behavior may attribute to the reduced involvement of female respondents in smoking cigarettes and other tobacco products2.

The prevalence and odds of being dual/poly users of tobacco products were higher among respondents from older age groups (i.e. aged ≥16 years). The finding is consistent with various studies. For example, Willis et al.27 reported prevalence of dual/poly users of tobacco products were almost two times higher among adolescents in 10th grade compared to those in 9th grade. In addition, Osman et al.28 also reported a positive linear correlation between age and prevalence of usage, whereby 11% among adolescents aged 14 years were tobacco product dual/poly users, but the prevalence increased to 22% among adolescents aged 15 and 16 years, and further increased to 45% among those aged 17 years. This finding might be due to the majority of mono smokers (cigarette smokers) being among the older aged youths. Kowitt et al.20 found that susceptibility and tendency to use other tobacco products are among mono users of cigarettes. In addition, parents/guardians in Asian countries, including Malaysia, might give more protection to younger adolescents than to those from older age groups. The freedom enjoyed by older adolescents may increase the likelihood of involvement in risky health behaviors such as smoking. However, this needs to be investigated further in future studies.

We also found differences in the prevalence and odds of being mono or dual/poly users between ethnic groups and schooling locality. For instance, high prevalence and odds of mono use of tobacco products were reported among Malay, Sabah and Sarawak native adolescents, as well as in the rural areas. The odds of being dual/poly tobacco users were significantly higher among native Sabahan youths. This finding is interesting, as it could be because e-cigarettes and shisha are new products compared to cigarettes, therefore, they are less widely used. This is further strengthened by the low prevalence of their use among all youths from different ethnic groups and schooling areas. This suggests there is an opportunity to formulate and implement appropriate policies or measures to stem the use of these products, which in turn reduces the incidence and prevalence of dual/ poly tobacco use among youths in Malaysia.

Exposure to SHS at home, school and outside the house increased the odds of being a dual/poly user of tobacco products. The finding is congruent with the study of Mamudu et al.29 among youth in Tennessee. Exposure to tobacco use at home, school and outside the house increases tobacco use among youth because it normalizes tobacco use and makes the behavior socially acceptable. These might be among the plausible reasons for the findings obtained in this study. Thus, smoke-free home and/or vehicle policies would be required to reduce youth exposure and the negative implications of social acceptance of tobacco use. This affirms the US Surgeon General’s suggestion that the coverage of smoke-free policies should be extended to include e-cigarettes30.

The prevalence and the odds of being a dual/ poly user were significantly lower among those who perceived tobacco use to have adverse health effects. The finding was in-line with Osman et al.28 who reported 1.79 times higher odds of being a dual/poly user of tobacco among respondents who perceived tobacco use to be less harmful. In addition, the finding is also consistent with findings by Cooper et al.22 among youth in the US. This finding may be explained by the Health Belief Model31 that states that an individual who perceives the severity of behavior will stay away from the behavior. It is also in line with the Cognitive Dissonance Theory, in that people who want to persist with a negative behavior will tend to downplay the adverse effects to rationalize the behaviour32.

Exposure to tobacco product advertisements at point of sale was found to be insignificantly associated with mono and dual/poly use. This observation may be because these advertisements affect those who have yet to smoke more, making them susceptible to initiate smoking, as reported by an international study33. However, future studies need to examine whether the findings in this study are consistent for adolescents in this country and understand the cognitive mechanism behind it among dual/poly tobacco users.

In this study, we found that mono and poly users shared similar characteristics. This result has also been reported by a study in the US34. Other studies have also shown that mono users have a higher likelihood of becoming poly users22. Thus, it is imperative to identify budding smokers and initiate preventive measures to reduce smoking among youths.

Limitations

The study has several limitations. Firstly, as the data on tobacco use were self-reported, this may cause recall or response bias. Secondly, several variables such as drinking habit35 and peer smoking35, which had shown significant association with dual/poly users, was not investigated in this study. Thirdly, the cross-sectional study design does not allow the causal relationship to be established between independent and dependent variables. Fourth, only three types of tobacco products were investigated in the current study; this might underestimate the actual dual/poly users of tobacco products among school-going adolescents in Malaysia. Lastly, there is no biochemical validation to verify smoking status in this study. Therefore, there may have been under-reporting of tobacco use. Fifth, since the data were collected in 2016, they might not reflect the current situation of dual/poly use of tobacco among youth in Malaysia. However, a previous study in Malaysia showed the consistency between self-reported and exhaled carbon monoxide concentration when the anonymity and confidentiality of respondents were guaranteed36. In addition, the majority of Malaysian youth were enrolled in government primary and secondary schools, which enables the findings of the study to be generalized to the majority of youth in Malaysia.

CONCLUSIONS

The prevalence of dual/poly users was still low among Malaysian school-going adolescents. However, their characteristics were similar to those of mono users of tobacco products, and the majority used e-cigarettes before using other tobacco products. Therefore, proactive measures such as limiting access to e-cigarettes and identifying those non-tobacco users who potentially will use tobacco products should be implemented.