Disparities in tobacco use by adolescents in southeast, Nigeria using Global Youth Tobacco Survey (GYTS) approach

The study area, Enugu State, is one of five states in southeast region of Nigeria. It has 17 Local Government Areas (LGAs), administrative units of the state, five of which are urban and 12 are rural. Enugu State has 242 urban secondary schools and 393 rural secondary schools, of which 283 are public schools while 352 are private schools [27].

A cross-sectional study design was used to compare findings from school-based surveys of two groups of schools, defined by their urban or rural location (or strata) within Enugu State. A school survey was used because it is easy to administer, relatively inexpensive, provides reliable results and refusal is significantly lower than in household surveys. Also, some adolescents may not wish to discuss a potentially sensitive topic like tobacco use in their homes.

Different approaches for studying adolescent tobacco use exist, but the GYTS design was chosen because it is comprehensive and collects data on all key tobacco control indicators used to monitor youth tobacco use. It also uses a standardised methodology for the study procedures to enable comparison between studies. Some minor modifications were made where appropriate to fit the Nigerian setting. For example, country-specific values were included in the options to some questions in the questionnaires.

The study population comprised adolescents in 8th to 10th grades (Junior Secondary 2 and 3 [JS2 and JS3], and Senior Secondary 1 [SS1]), in both public and private secondary schools in Enugu State. Students were included in the study if they were in 8th, 9th or 10th grade in selected schools. All students in selected classes were eligible to participate in the survey. Students in the selected classes who were absent on the day of data collection or who refused to participate were excluded from the survey.

The minimum sample size required for the study was computed using the formula for determination of sample size for comparison of independent proportions [28]. Using power of 90%, significance level of 0.05, prevalence of 26.1% from previous GYTS in Nigeria [23], prevalence difference of 5%, and response rate of 80%, the minimum sample size calculated was 1890 students per group. This corresponded with the minimum sample size of 1875 students from 25 schools required for studies using GYTS methodology [29]. A sample of 2000 students from 25 schools in each study group was sought, resulting in a total of 4000 students from 50 schools for the entire study; 80 students per school.

All secondary schools in Enugu State were identified and stratified into urban and rural using the host LGAs. The 2014/2015 session school enrolment list, obtained from the Ministry of Education, contained the most current student enrolment information by grade and gender. This was modified by creating a separate enrolment list for each stratum. Two-stage cluster sampling was used to select the students independently within each stratum, with the probability of selection proportionate to school enrolment size.

In the first stage, 25 schools were selected from each stratum using systematic random sampling. The schools were the primary sampling units. In each stratum, the schools in the sampling frame were sorted in descending order by enrolment size. The cumulative population of eligible students in each school was calculated. To determine the sampling interval (SI), the total number of eligible students was divided by 2000 - the sample size for each stratum. A number, designated random start (RS), was chosen randomly between 1 and the SI. The school within which the RS fell was selected first, after which subsequent schools were selected by adding the SI to the RS. Selection of schools used the formula: RS; RS + SI; RS + 2SI; RS + 3SI, etc. Schools for which the cumulative population contained one of the serial numbers calculated from the series above were selected [29]. Microsoft Office Excel 2013 was used for sampling at this stage. Schools with a population size of less than 80 were excluded from each sampling frame before the 1st stage of sampling. The proportions of students excluded from the urban and rural sampling frames were 5.6 and 10.7% respectively. The urban sampling frame had 172 schools, with a population size of 60,601, while the rural sampling frame had 285 schools and a population size of 59,173.

In the second stage, classes were selected from each school using systematic random sampling. All students in the selected classes who were present on the day of survey administration were eligible to participate. There was no sample substitution for schools or classes that refused to participate in the survey because this would have altered the representativeness of the sample and thus would have led to biased results. Every effort was made to ensure the cooperation of selected schools and classes. (See Fig. 1 for a flow diagram of the sampling).

An anonymous, self-administered, semi-structured questionnaire adapted from the GYTS Core Questionnaire with optional questions Version 1.1 [30] was used to collect information about students’ socio-demographic characteristics, use of smoked and smokeless tobacco, and susceptibility to smoking initiation (Additional file 1). The questionnaire was first pre-tested with 162 students in three secondary schools that were not selected for the study: two urban (one public and one private) and one rural. These schools were located at least ten kilometres from the schools that were selected for the study. After correction of ambiguities identified in the questionnaires, the survey was administered in English in November and December, 2015. Teachers were absent during survey administration to enable the students to be free to answer the questions. During survey administration, one rural private school was found to be ineligible so no data were collected from that school.

Ever use of tobacco (cigarettes, other smoked tobacco, and smokeless tobacco) was defined as experimentation with any of the aforementioned tobacco products. Current use of cigarettes was defined as use on 1 day or more within the past 30 days. Current use of other smoked tobacco products (pipes, cigars, water pipes/shisha or bidis) or smokeless tobacco was defined as any use within the past 30 days. Any tobacco use was defined as use of one or more types of tobacco products, and both ever use and current use were examined. Dual tobacco use was defined as current use (within past 30 days) of any two tobacco products. Multiple tobacco use was defined as current use (within past 30 days) of cigarettes, other smoked tobacco and smokeless tobacco. Susceptibility to smoking initiation within the next year was defined as the absence of a firm decision not to smoke within the next year. This was determined if a student did not answer “definitely no” to the question ‘At any time during the next 12 months, do you think you will use any form of tobacco?’

Weights were initially calculated for each student record to account for the complex survey design and for differential non-response at school, class and student levels, using the GYTS approach [29]. The final weight for each record was a product of the school selection weight, class selection weight, and overall non-response adjustment factor. The school selection weight was the inverse probability of selecting a school. The class selection weight was the inverse probability of selecting a class in a school. The overall non-response adjustment factor was the product of school, class and student non-response adjustment factors. Weights were computed separately for each stratum.

Analyses were performed for all adolescents (ages 10–19 years), and for a subset of adolescent students (13–15 years of age) to enable comparison with previous studies.

Ethical approval for the study was obtained from the Health Research Ethics Committee of the University of Nigeria Teaching Hospital (NHREC/05/01/2008B-F-WA00002458-1RB00002323). Approval was also obtained from the Ministry of Education and the principals of the selected schools. Written assent was obtained from the students after a detailed explanation of the study objectives, procedures, risks and benefits. Written informed consent was also obtained from principals of the selected schools, who acted as the legal guardians of the students, as was done in previous GYTS studies in Nigeria.

Among all adolescents, there was a general pattern of higher use among students older than 15 years, with results reaching statistical significance for all three ever use measures (Table 3).

The present study found that almost one in three adolescents had ever tried smoking cigarettes and 13% of adolescents currently smoked cigarettes. These rates are much higher than those reported in Ibadan, southwest Nigeria, where 9.4% of students reported ever smoking cigarettes and 3.4% reported current use of cigarettes [21]. These differences in prevalence may be due to under-reporting of tobacco use by the students in Ibadan because their teachers were present during data collection. Similarly, two studies of female and male students in Anambra, southeast Nigeria reported lower prevalence of current cigarette smoking, 7.7% [25] and 8.7% [26] respectively. However, these studies were conducted in only public secondary schools. Therefore, tobacco use in private schools has remained unknown. In contrast, a study of secondary school students in Akwa Ibom, south-south Nigeria reported higher prevalence of current use of tobacco (34.8%) [22], but the investigators did not report their sampling method so the study may not be replicable. Furthermore, the prevalence of dual tobacco use in this study compares with findings in Limpopo, South Africa [31] but is higher than rates reported in the US national youth tobacco survey [32].

The present study’s findings of prevalence of ever use and current use of cigarettes are similar to findings from GYTS in Greece [33], where consistent GYTS procedures were followed. Yet, the results were lower than those reported in GYTS studies in Zambia [34] where 40.1% had ever smoked cigarettes, and Democratic Republic of Congo [19] where prevalence of current use of smokeless tobacco was 18%. It is noteworthy that these two African GYTS studies had response rates (RR) of 78.3 and 57% respectively, lower than the RR of 80% expected from GYTS studies. This may have affected the precision and validity of these results. Additionally, susceptibility to smoking initiation within the next year among non-smokers reported in this study is similar to findings from GYTS studies in Pakistan [35] (12%) and Philippines [36] (10.5%), indicating a possible rise in smoking prevalence in the subsequent year.

Notably, the prevalence rates of both ever use and current use of cigarettes were higher in this study compared with the 2008 GYTS in Nigeria which reported that fewer than 20% of students had ever smoked cigarettes and fewer than 10% of students currently smoked cigarettes [20]. These results may not be generalizable to all 13 to 15 year olds in Nigeria because the study excluded the southeast and northeast regions, which have contextual differences from other regions studied. However, the higher prevalence in this study may indicate a rising prevalence of adolescent tobacco use in Nigeria since the last GYTS was conducted nine years ago.

Findings in the present study show that prevalence of tobacco use was higher among students in rural schools compared to urban schools for all types of tobacco products. This finding could be explained by the presence of older students in rural schools compared to urban schools since tobacco prevalence was found to increase with increasing age. It is also possible that there are contextual differences in rural and urban locations which affect adolescent tobacco use but could not be identified in the present study. Similar findings have been reported by other studies, mostly outside Africa [37‐39]. The study in Osun, Nigeria [37] compared only one urban and one rural school, was inadequately powered due to small sample size, and may not be generalizable to adolescents in Nigeria. Among 13 to 15 year olds, it is remarkable that the prevalence of cigarette smoking, other smoked tobacco, dual and multiple use of tobacco, were all twice as high in rural schools compared to urban schools. This underscores the need for tobacco prevention interventions to also target adolescents in rural schools. In contrast, prevalence of smokeless tobacco use was comparable in both locations indicating that the driving factors for smokeless tobacco use may be similar in both locations. However, the finding that susceptibility to smoking initiation in the next year was higher in rural schools is of utmost public health importance since it suggests that prevalence of adolescent tobacco use will continue to be higher in rural schools in the coming year.

The prevalence of all types of tobacco use was higher among boys compared to girls, as has been similarly reported in other studies [34, 39‐44]. This is in contrast to the 2008 Nigerian GYTS report which showed no difference in prevalence of current cigarette smoking between boys and girls except in Kano [20]. It was surprising that boys reported higher use of smokeless tobacco than girls – unlike the finding in a similar study among Black South Africans [45]. The tobacco product most commonly reported by both boys and girls was cigarettes – highlighting the need for interventions that target cigarettes.

The present study’s finding that the prevalence of ever use of tobacco increased with increasing age, has been reported in previous studies [26, 39, 42]. On the contrary, current use of all types of tobacco was similar among the age categories studied (10–12 years, 13–15 years, and 16–19 years), consistent with the results reported in Zaria, Nigeria [46]. A possible explanation for this finding is that younger students could have been experimenting with tobacco at the time of this study. This is of concern because early exposure to tobacco can predispose adolescents to addiction and prolonged exposure leading to development of tobacco-associated diseases later in life.

In addition, this study showed no differences in current tobacco use by students in 8th to 10th grades, contrary to findings from other studies. This unusual finding was similarly observed in urban India where 6th graders used tobacco at higher rates than 8th graders [42]. A possible explanation for this finding in the present study is that 8th graders may have been experimenting with tobacco at the time of this study. The implication is similar to that of the observed age prevalence differences.

An important finding in the present study is that almost 90% of current smokers desired to quit smoking and had tried to quit in the last 12 months. A similar finding was reported by the 2011 GYTS in the Philippines [36]. Unfortunately, about 60% of adolescents in the present study reported they had never received help to stop smoking, unlike in other countries. In Nigeria, although some smoking cessation support exists in some hospitals, either in the form of physician advice or counselling, or the prescription of nicotine replacement therapy or anti-depressants, it is usually not covered by health insurance. There are no smoking cessation support currently available in primary healthcare facilities or in the community where it is easier for smokers to access them [47]. In addition, no specific smoking cessation support is available for adolescents in Nigeria to the best of our knowledge. The findings of this study underscore the need for culturally tailored smoking cessation programs for adolescents. These programs could use adolescents as change agents with their peers, since the most common source of help for smoking cessation reported in this study was friends. Such interventions could use peers or role models who are fluent in the native language to target rural schools.

The present study was school-based, therefore its findings may not be generalizable to all adolescents in Enugu, Nigeria. However, a significant proportion (71.1%) of adolescents are enrolled in school in Enugu State [48]. Additionally, because only students who were present in school on the day of data collection participated in the study, the findings may have been over-estimated or under-estimated depending on whether tobacco users were more likely to be in school on that day or not. This is unlikely, however, because the students were not given prior information about the study until the arrival of the study team on the day of data collection. Another limitation, common in research, is that the survey was self-administered and students may over-report or under-report their tobacco use behaviours. Although the extent of this type of bias cannot be determined from the present study, reliability studies in the US have shown good test-retest results for similar tobacco-related questions [49].