Tobacco use and its determinants in the 2015 Kenya WHO STEPS survey

Tobacco use and exposure is one of the leading risk factors for non-communicable diseases (NCDs), a major public health threat globally and a preventable cause of morbidity and mortality [1–10]. According to the World Health Organization (WHO), in 2015, over 1.1 billion people smoked tobacco, which represents around 15% of the global population [2]. Smoking was attributable to 11.5% of total global deaths in 2015 and 150 million disability-adjusted life years [10]. Nearly 80% of these individuals live in low- and middle-income countries which constitutes the majority of countries in the African region [3, 8]. Use of tobacco products and exposure to tobacco smoke kills around 6 to 7 million people each year [1, 2, 5]. Over 5 million of the deaths result from direct tobacco use and a further 600, 000 result from non-smokers being exposed to second-hand smoke [1, 9].

Although Africa has had low prevalence of tobacco use [11, 12], projections show increase with population growth [9, 13]. Around one in five adults in the region smokes tobacco [3, 9, 10, 14–16], with the highest rates in Sierra Leone (37.7%) and lowest in Sao Tome & Principe (6.75%) [15]. Among users, there is a disproportionally higher proportion among men as compared to women, at 21 and 3%, respectively. Eighteen percent of young people (13–15 years old) in Africa currently use some form of tobacco product [3, 17]. Studies show use of tobacco as early as under 7 years of age in western Africa [18] and in Kenyan secondary schools as early as 11 years old [19]. The landscape of tobacco use in Africa is changing with increasing use of alternative forms of tobacco being used such as e-cigarettes including use by women [10, 20–23].

According to the 2014 Kenya Global Adult Tobacco Survey (GATS), a nationally representative household survey as a component of global tobacco surveillance system, 2.5 million adults use tobacco products in Kenya, which constitutes 11.6% of the population [24]. The 2014 Kenya Demographic and Health Survey (KDHS) estimated 16% of men and 0.4% of women aged 15–45 use cigarettes [25]. In its 2014 global NCD report, the WHO estimated current tobacco use at 13% [26]. These figures are consistent with findings by Gathecha et al. in a review of previous studies on tobacco use in Kenya that demonstrated prevalence range between 3.8 and 19% [27]. The Kenya GATS showed there was a higher proportion of men that use tobacco products as compared to women at 19.1 and 4.5%, respectively, and that around 6% of users smoke tobacco daily while 1.8% report not to use daily [24]. Alcohol has been shown to be associated with increased tobacco use in sub-populations such as rural Kenya and Nairobi slums [28, 29]. The GATS Kenya survey uses the definition of smokeless tobacco products as those wholly or partly made from tobacco and do not need to be ignited for it to be consumed.

Kenya ratified the WHO Framework Convention for Tobacco Control (FCTC) in 2004, and implemented the Kenya Tobacco Control Act in 2007 [2, 6, 30]. The country has an agency dedicated to tobacco control, and attained the highest level of achievement for tobacco advertising, promotion and sponsorship deterrent efforts, which are both measures of success according to the WHO FCTC and MPOWER [31–34]. While gains have been made, given these guidelines, there is still need for greater attention, especially first establishing the levels and determinants of tobacco use [6, 8, 27, 35, 36].

The WHO STEPwise approach to surveillance of NCD risk factors (STEPS) survey is part of a global surveillance strategy in response to the growing need for data on country level trends in NCDs and injuries and the related risk factors [37]. The Kenya STEPs survey is the first nationally representative survey to collect comprehensive information on risk factors for NCDs and injuries [38]. The survey serves as an evidence base to strengthen NCD prevention and control initiatives in the country. The objective of our study was to describe patterns and determinants of tobacco use in Kenya, including current and former use. We also assessed the influence of alcohol use amongst tobacco users in a national sample, a novel addition to the literature. Finally, we examined the more novel use of e-cigarettes.

The WHO STEPwise approach to surveillance (WHO STEPS) study was completed in Kenya between April and June 2015 [38]. The STEPS instrument is a cross-culturally validated survey tool used to assess burden of leading NCDs and associated lifestyle risk factors in a nationally representative sample. A cross-sectional household survey approach was used, with a three-stage cluster sample design involving selection of clusters from the National Bureau of Statistics household-based sampling frame (NASSEP V), households and eligible individuals as further described in the Kenya WHO STEPS report [38]. The individual identified as the head of household at the time of contact responded to the survey. Written informed consent was obtained from the selected individual. Criteria for inclusion was individuals aged 18–69 years old. Those that refused consent were excluded from the study.

Most tobacco users had only ever attended up to primary school education (n = 434/605, 71.7%). Conversely, only 1.2% of current tobacco users had completed university education or higher (data not shown in table). The proportion of tobacco users in rural and urban areas are similar (21.5 and 21%, respectively), however, there is a predominance of tobacco smoking in urban areas and a predominance of smokeless tobacco use in rural areas. Of the seven respondents who reported e-cigarette use, most were male, younger than 40 years old, had attained some education, and lived in urban areas.

Men had nearly seven times higher odds of being tobacco users as compared to women (OR 7.63, 95% CI 5.63–10.33), see Table 3. After controlling for confounding, the strength of the relationship between age and tobacco use was less evident, however a trend showing an increase in use with age still remained, and those in the 50–59 year age group, were nearly twice as likely as compared to their younger counterparts aged 18–29 to be engaged in tobacco use (OR 1.76, 95% CI 1.14–2.74). Those with at least primary education were less likely to be currently using tobacco. Those that use alcohol have three times higher odds of being current smokers (OR 3.36, 95% CI 2.52–4.48). Of note, after controlling for confounding the relationship between tobacco use and heavy episodic drinking, in particular, was no longer evident also as shown in Table 3.

This study provides the first nationally representative estimates on the prevalence of tobacco use alongside other leading risk factors for NCDs, and disease outcomes related to these risk factors. Overall, 13.5% of the respondents currently use tobacco. This is slightly higher than what was has been reported in the Kenya GATS conducted in 2014 at 11.6% [25], and in some previous studies in the country as shown by Gathecha et al. [27]. The prevalence of tobacco use was also higher in the STEPS as compared to the 2015 GBD study, where for example daily tobacco use for males was 18.6% compared to 14.9%, respectively [10]. The difference between the STEPS survey and GATS could be attributed to the different age ranges of the population sampled. The Kenya GATS examined respondents aged 15 years and above, which may include an age group (15–18 years) that have not yet engaged in tobacco use therefore lowering the prevalence estimate. Alternatively, differences in the STEPS and GATS could be explained by sampling approach. While both the STEPS and GATS use sample frames that originate from the Kenya National Bureau of Statistics (KNBS), cluster selection, and randomization for participant selection differed [25, 38].

The high prevalence found in our study is of concern, and calls for urgent action as tobacco use is not only a major risk factor for health, it also is a major contributor to inequities in health and social development [42–44]. Prevalence levels are higher than in 14 of 30 other sub-Saharan African countries [15], and in other countries in the East African region such as Ethiopia (4.2%) and Uganda (9.6%). It is comparable to Rwanda (12.8%) and Tanzania (14.1%) [45–48]. It is therefore a call for renewed diligent implementation of the WHO FCTC and the Tobacco Control Act (TCA) by all stakeholders [3].

A national tobacco cessation quit line, nicotine replacement therapy and pharmacologic aids for cessation such as Bupropion are available. However, the availability of treatment for tobacco cessation is only at certain sites. Furthermore, coverage for treatment is not guaranteed. Finally, the enforcement of smoke-free zones is still lacking and no current monitoring for effectiveness of policies exists to date [27, 31, 35, 49]. Additionally, more can be done at the healthcare provider to level to educate patients on smoking cessation [50, 51]. Lack of understanding amongst the general population on the effects of tobacco use on chronic disease, such as cancer, continues to be a problem [27, 52].

The changing trend of tobacco use in Africa is also important to highlight. The Africa region is estimated to have the highest predicted increase in tobacco consumption between 2010 and 2025 [53, 54]. This is because the tobacco industry is turning its focus on the youth and especially in Africa which is at the same time experiencing large population increases [10, 12].

Consistent with studies done in other countries in sub-Saharan Africa [15, 16] and other LMICs [55], tobacco use among men is significantly higher than women (OR 7.11 CI 5.65, 8.93). Studies suggest that this phenomenon may be attributable to the tobacco industry consistently targeting men by portraying smoking as a masculine activity with advantages such as increased sexual prowess as well as the community perception towards smoking that is generally more tolerant of male smokers over female smokers [56–59]. Recognizing sex differences in tobacco use is important because it guides which tobacco control interventions should be prioritized in order to reach the group that is most affected [60]. Studies have shown that there should be a difference in interventions such as tobacco messaging and cessation depending on the sex targeted [60–63].

Our study found that majority of the smokers were in the age group 18–29 years and the mean age of initiation was early twenties. This is a vulnerable population that are being targeted by the tobacco industry especially in the African region [17, 20, 54, 64, 65]. The high prevalence of tobacco use among the younger population has been proposed to be driven by the high exposure to advertising, among other factors. For example, 59% of youth are currently being exposed to pro-cigarette billboard advertisements and 15% of these youth in the region are likely to initiate smoking [66]. Addressing tobacco use at this age group is critical as it will determine the future burden of tobacco use and its consequences. This calls for strengthened enforcement of the FCTC and TCA that already have provision on the banning of tobacco advertising, promotion and sponsorship.

In our study, 3.6% of the respondents reported consuming smokeless tobacco while the figure was 4.5% in the GATS [25] and 2.05% for men in the KDHS [15]. The use of smokeless tobacco may be influenced by smoking bans in public places, introduced in Kenya in 2007. Researchers have found that smokeless tobacco use is perceived to be less harmful than smoked tobacco products when these products cause nicotine addiction, cancers and other non-communicable diseases [67]. In our population, higher levels of education were protective against smokeless tobacco use. Educational interventions specific to harmful risks of smokeless tobacco use may have a significant role. There was a higher likelihood of smokeless tobacco use among rural residents than urban residents necessitating the need for intensified campaigns against smokeless tobacco in rural areas.

Alcohol use had a positive effect on current and daily tobacco use. This relationship was true for ever users of tobacco, binge drinkers and non-binge drinkers. This finding is corroborated in literature [41, 65]. The association between tobacco and alcohol use has been explained by many factors including genetics, neurobiological mechanisms, conditioning mechanisms and psychosocial factors [67, 68]. Our study found higher odds among binge drinkers (OR 8.7) than non-binge drinkers (OR 2.8) when exploring current tobacco use. There is, therefore, an added advantage in tailoring messaging to the public against joint tobacco and alcohol use [69, 70]. Additionally, it is prudent for stakeholders working in alcohol control and tobacco control to work collaboratively to maximize impact.

Respondents with no formal education were more likely to be tobacco users. Similar findings are prevalent in other LMICs comparable to Kenya [10, 15, 16, 55]. This has an implication on tobacco control public education. The Ministry of Health and other stakeholders should ensure that they employ a mix of communication channels when distributing anti-tobacco messages. Innovative, culturally appropriate and carefully designed messaging is critical to reaching at-risk groups including adolescents and those with limited ability to read [70–72]. In regards to health warnings on cigarette packaging, the Ministry of Health recently passed regulations for the implementation of graphic health warnings which will have an effect on smokers who cannot read [73].

These are the first findings at the national level on e-cigarette use in Kenya. Seven respondents reported to have ever used an e-Cigarette. The global estimates of ever use and current use of e-cigarettes are 54.7 and 19.4% among current tobacco smokers and 7.0 and 1.5% among non-smokers respectively [74]. While motivations for usage are not clear, tobacco cessation is associated with increased use of e-cigarettes whether or not users intended to quit smoking cigarettes [75]. In this aspect, the trend is encouraging however future targeted studies need to be done to understand better the drivers of e-cigarette use and to monitor prevalence of usage as it is predicted that e-cigarette use will increase in future [23].

The use of tobacco has seen concerning trends in the African region in recent decades, Kenya notwithstanding. The 2015 Kenya WHO STEPS was unique in providing primary data on the status quo of tobacco use in the country. Interesting trends by sociodemographic status highlight the importance of focusing on young people, men, and those with lower levels of education. Concurrent usage of alcohol is associated with higher odds of risky tobacco use. Targeting tobacco prevention strategies amongst alcohol users is therefore recommended, including revisiting the public smoke ban for effective implementation. Further data is needed on the use of smokeless tobacco, and its impact on smoked tobacco products, as well as on the novel use of e-cigarettes, with considerations for health and societal implications.