INTRODUCTION
The tobacco epidemic remains one of the most significant public health challenges worldwide, and South Africa is no exception1,2. According to the 2021 Global Adult Tobacco Survey (GATS), 29.4% of South Africans aged >15 years use tobacco products, including smoking tobacco, smokeless tobacco, or heated tobacco2. While there has been a downward trend in tobacco consumption reported worldwide, low- and middle-income countries continue to bear the heaviest burden1,3. Despite various efforts by governments and the World Health Organization (WHO) to curb this dangerous trend, tobacco use remains a pressing public health issue1,2.
Tobacco use and other unhealthy behaviors are prevalent in South Africa4,5. Data from the South African Demographic and Health Survey (SADHS 2016) highlight some concerning trends5 – 37% of men and 8% of women are tobacco users; 68% of women are overweight, with 41% obese; 31% of men are overweight, with 11% obese; 61% of men and 26% of women consume alcoholic beverages; and of those who consume alcohol, 28% of males and 5% of females engage in risky alcohol consumption5. Furthermore, the SADHS (2016) found that only 25% of households reported consuming at least one portion of fruit or vegetables (F&V) daily5.
The ‘3-4-50’ concept highlights the link between three key unhealthy behaviors, namely: tobacco use, inadequate physical activity, and poor diet6. These constitute major risk factors for four major chronic diseases (cardiovascular diseases, cancers, type 2 diabetes, and respiratory conditions) and ultimately account for over 50% of premature deaths worldwide6,7. Evidence from multiple studies suggests that unhealthy health behaviors tend to cluster together, increasing the overall risk of negative health outcomes and socioeconomic inequalities6-8. A Canadian study reported that smoking cessation is challenging when multiple unhealthy health behaviors co-exist, as the odds of dependence on nicotine increase by 23% with each additional unhealthy behaviour9.
A 2022 scoping review identified a rising trend in smokeless tobacco (SLT) use among a much younger population10. SLT users were also most likely to engage in multiple unhealthy behaviors, such as smoking cigarettes and consuming alcohol concurrently10-12. This trend was confirmed by the Centers for Disease Control and Prevention (CDC) in the USA, which reported that novel forms of SLT, along with the influence of friends and family, were identified as key elements influencing the use of SLT products among the youth11.
It is well established that all forms of tobacco products are harmful, with no safe level of exposure1. Therefore, promoting a quit attempt is a desirable clinical and public health priority1,8. It is shown that brief motivational interviews by healthcare providers increase the chances of a successful quit attempt by 30%, while quitting rates are increased by 87% with intensive counselling sessions, follow-ups, and pharmacotherapy13. Although most tobacco users express the desire to quit, many are unsuccessful, as the chances of quitting also decrease with higher nicotine dependence (HND) and the presence of concurrent risky health behaviours6,8,9.
Studies suggest that the likelihood of attempting to quit tobacco use is associated with being female, older, having higher formal education, exposure to smoke-free environments, lower level of nicotine dependence, awareness of the dangers of tobacco, and concern about health8,13,14. In addition, multiple studies have shown that past or present alcohol use disorder is negatively associated with making a quit attempt and has a greater chance of relapse15,16. Also, smokers with HND are less likely to attempt to quit8,17 while weight gain has been identified as both a barrier to quitting and a trigger of relapse18.
A large number of adult patients visiting the outpatient department (OPD) at a district hospital seek care for the treatment of non-communicable diseases (NCDs) related to health19. Information for South African is, however, limited on the relationship between quit attempts and behavioral health, particularly, whether a quit attempt is associated with the uptake of other healthy behaviors among ambulatory hospital patients, and the extent to which nicotine dependence modulates this. Based on these findings, we hypothesized that individuals who adopt one or more healthy behavioral habits are more likely to try to quit. Therefore, this study aims to evaluate the relationship between quit attempts, nicotine dependence, and health behaviors among tobacco users.
METHODS
Study design and research setting
This cross-sectional study was conducted between February and July 2023 at the OPD of DYDH among tobacco users aged ≥18 years. DYDH is a district-level healthcare facility located in Krugersdorp, west of Gauteng, South Africa. At the time, the hospital had a capacity of 245 beds, served as a referral center for 44 fixed primary care clinics, eight mobile clinics, four satellite clinics, and three maternity and obstetric units. The OPD operates weekdays, Monday to Friday, from 8:00 a.m. to 4:00 p.m.
Study population, sampling strategy, and recruitment
According to the district health information system, there were 26632 adult patient visits to the OPD of DYDH in 2019. In addition, a study by Bokoro et al.13 , in the same setting, reported a tobacco use prevalence of 20.0% in 202013. The required sample size for the study was calculated to be 336 respondents, using a 5% margin of error and a 95% confidence interval. To compensate for potential non-responses and incomplete data, the sample size was increased by 10%, resulting in a final target of 400 respondents.
On each working day, consecutive tobacco users were invited to take part in the study. Posters in English about the research study were displayed in the OPD with permission from the hospital management. A trained research assistant fluent in English as well as several local languages (Setswana, Afrikaans, Sotho, and Zulu), notified all patients about the ongoing study, and invited them to participate after the morning health promotion talks. Those who consented were assured that they would not lose their position in the queue, as they would be returned to their original position or the next position in the front of the queue once they completed their participation. The enrolment continued until the sample size was reached. All screened patients received a sticker on their files afterwards, to avoid multiple recruitment.
Data collection
A researcher-administered questionnaire, written in English, was used to collect the data. Each questionnaire had a code to ensure anonymity. The questionnaire was developed with input from relevant literature on tobacco use, nicotine dependence, quit attempts, and health behaviors4,5,20,21. Ethical approval for the study was granted by the Human Research Ethics Committee (HREC-Medical) of the University of the Witwatersrand, South Africa (clearance number: M201173). Additional permission was obtained from the Chief Executive Officer of DYDH. Informed consent was obtained from all respondents.
Measures and definitions
A 12-month quit attempt was defined among tobacco users as a ‘yes’ to the question: ‘During the past 12 months, have you tried to stop tobacco use?’22,23. A lifetime quit attempt was defined by any answer indicating ‘once’, ‘twice’, or ‘three times or more’ as opposed to ‘never’ in response to the question: ‘Have you ever attempted to quit tobacco use?’8.
To determine the nicotine dependence levels among smokers, the Heaviness of Smoking Index (HSI) was used. The HSI is a validated and effective measure of nicotine dependence24. It is derived from two items: the number of cigarettes smoked per day (CPD) and the time to first cigarette (TTFC) in the morning24.
Health behaviors included dietary consumption of fruit and vegetables (F&V), physical activity, and consumption of alcohol. Dietary consumption of F&V was measured as proposed in the South African Social Attitudes Survey (SASAS) of 20145, by asking respondents how often they consumed fruits and/or vegetables per day or week5. Adequate F&V intake was defined as the WHO guidelines of a minimum of five servings of F&V daily25.
Physical activity was measured using the Global Physical Activity Questionnaire (GPAQ), a tool developed by the WHO to monitor physical activity levels across different countries6,21. The GPAQ gathers data on physical activity across three domains: vigorous and moderate activity at work, transport-related activity, and vigorous and moderate activity during leisure time. Additionally, it records information on sedentary behavior, specifically the duration of time spent sitting21.
To assess alcohol consumption, the Alcohol Use Disorders Identification Test-Concise (AUDIT-C) tool was utilized. This is a brief but effective screening tool for identifying individuals who engage in harmful drinking or may have an active alcohol use disorder26. The AUDIT-C comprises the first three questions from the full Alcohol Use Disorders Identification Test (AUDIT). Harmful drinking refers to a pattern or quantity of alcohol consumption that increases the likelihood of adverse health outcomes and is recognized by the WHO as a distinct disorder26.
Covariates
Current tobacco users were defined as individuals who responded ‘daily’ or ‘less than daily’ to the question: ‘Do you currently use any tobacco products on a daily basis, less than daily, or not at all?’22. For tobacco use patterns, the GATS questionnaire was used to collect data on tobacco use patterns20. The GATS questionnaire is an internationally valid and reliable tool used to obtain information on tobacco use and quit attempts for both smokers and SLT users20.
Sociodemographic characteristics data collected included age, gender (male, female), marital status (single, married, or divorced/widowed/separated), employment status (employed including professionals, artisan and self-employed; and unemployed including students), education level (no formal education, primary education, secondary education, tertiary), and ethnicity in terms of self-identification as Black African, White, Colored (of mixed ancestry), (Black, White, Colored, Indians), and Indian/ Asian.
Study respondents were also asked to rate how they perceived their general health status (good, average, poor, or I don’t know)5,8. Respondents were also asked to specify the reason for their clinical visit. Smokers were categorized into nicotine dependence levels based on their scores from the HSI questions24. Respondents were considered adequately physically active if they engaged in at least 150 minutes of moderate-intensity exercise or 75 minutes of vigorous-intensity activity per week21. The AUDIT-C tool was used to assess alcohol drinking behavior, where a score of ≥8 indicated harmful behavior.
Potential confounders were identified prior to analysis based on existing literature. Sociodemographic variables (age, sex, marital status, employment status, education level, and ethnicity) were included due to their known associations with tobacco use behaviors and quit attempts. Perceived general health status was included as it may influence motivation to attempt to quit, while the reason for the clinical visit was included to account for differences in underlying health status. Tobacco use patterns were treated as primary exposure variables. All identified factors were adjusted for in a multiple-stepwise logistic regression analysis.
Data analysis
A statistician provided support with data management and analysis. The data were entered into REDCap and later analyzed using STATA 18. Descriptive statistics, including frequencies and percentages for the categorical variable, and means and standard deviations for the continuous variables, were calculated. Measures of central tendency and cross tabulation were also performed in terms of age, sex, marital status, employment status, education level, ethnicity, self-reported health status, clinical diagnoses, mean number of cigarettes, snuff used, and proportions of tobacco users engaged in each health behavior. The 12-month quit attempt and lifetime attempt variables were analyzed as binary (yes, no).
To analyze the associations between respondents’ demographic and clinical profiles, levels of nicotine dependence, health-related behaviors, and attempts to quit smoking, appropriate statistical tests were applied: chi-squared tests, t-tests, ANOVA, and logistic regression were used for testing categorical and numerical variables, as appropriate. Statistical significance was set at p<0.05. Multivariate logistic regression analyses were performed to examine factors associated with making a quit attempt. The 12-month and lifetime quit attempt (yes, no) was specified as the dependent variable. Independent variables entered into the multivariable models included sociodemographic characteristics, adequate fruit and vegetable intake, adequate physical activity and harmful alcohol use. Adjusted odds ratios (AORs) with 95% confidence intervals (CIs) were estimated.
RESULTS
Respondent characteristics
A total of 400 Respondents completed the questionnaires. Supplementary file Table 1 summarizes the baseline sociodemographic data, self-reported health status, and clinical conditions of the respondents. Most respondents were Black (65%), male (51.2%), single (46.5%), employed (43%), and had a secondary education (71.2%). The mean age of the respondents was 48.3 years. Less than half of the respondents (45.2%) described their general health status as good. The most common reason for a clinical encounter was a cardiovascular disease (CVD) (49%). The mean duration of reported health conditions among respondents was 6.9 years.
Tobacco use patterns
About 80.5% (n=322) were cigarette smokers (268 were exclusive cigarette smokers), and 19.5% (n=78) were SLT users. Most of the cigarette users smoked daily (91.9%), and preferred the manufactured type of cigarette (63%). The mean number of packs per year of smoking was 14.5. Hookah pipe only was reported by 6.8% (n=27), e-cigarette only use was noted by 2 participants, combined cigarette and other product use by 4.8% (n=19), and combined e-cigarette and other product use by 1.5% (n=6).
Most cigarette smokers had low nicotine dependence (LND) (60.9%), while 21.25% had moderate dependence, and 17.7% high nicotine dependence (HND).
Among SLT users, the majority (71.7%, n = 56) used snuff daily. The median duration of snuff use was 10 years (IQR: 3–20). On average, SLT users used three snuff dips per day (Supplementary file Table 2).
Only three respondents (0.75%) met the WHO guidelines for adequate consumption of F&V, of a minimum of five servings daily. More than half of the respondents (55%) consumed less than one serving of F&V daily. Compared to others, respondents with a tertiary education were more likely to eat at least one serving of F&V each day (OR=2.04; 95% CI: 1.19–3.47, p=0.009) (Supplementary file Table 3).
Regarding physical activity, most respondents (83%) reported engaging in either vigorous or moderate physical activity. Of these, only 37.6% met the WHO guidelines for a total of 150 minutes of weekly physical activity. On average, respondents worked for eight hours a day, with half spending three hours sitting or reclining. In the univariate analysis, females, pensioners, and those with a mean pack-years of smoking higher than 14.5 were negatively associated with engaging in adequate physical activity (p=0.001, 0.001, and 0.002, respectively). Respondents who described their general health as good were less likely to engage in adequate physical activity (OR=0.22; 95% CI: 0.11–0.46, p=0.001). Tertiary education, being Black, and not being a pensioner were all significantly associated with engaging in adequate levels of physical activity (p=0.02, p=0.028, p=0.001, and p=0.006).
The overall prevalence of alcohol use among the study respondents was 67.7% (n=271), with 70% of Blacks, 55% of males, and 45% of females reporting alcohol use. Harmful drinking was reported by 17.3% of respondents. Being Black or unemployed was significantly associated with harmful alcohol drinking (p=0.03 and p=0.01, respectively). Harmful drinking was negatively associated with being a female or married (p=0.00 and p=0.03, respectively).
Quit attempts
In the past 12 months, 25% of respondents had attempted to quit tobacco, while 54.5% reported having tried to quit at some point in their lives (Table 1).
Table 1
Twelve-month and lifetime quit attempts among tobacco users at DYDH, South Africa, February–July 2023 (N=400)
Among various health behaviors, consuming at least one serving of F&V daily was the only factor significantly linked to a quit attempt in the previous 12 months (OR=2.77; 95% CI: 1.63–4.69, p=0.001). Smokers diagnosed with cardiovascular disease were less likely to attempt quitting (OR=0.59; 95% CI: 0.35–0.99, p=0.044). Across all tobacco users, those who rated their general health as good and those with a respiratory illness were more inclined to try quitting (OR=1.94; 95% CI: 1.09–3.44, p=0.024; and OR=1.07; 95% CI: 1.18–3.62, p=0.001, respectively). Marital status showed a significant association, with individuals who were married or in other non-single arrangements being more likely to have made a quit attempt (OR=1.62; 95% CI: 1.05–2.50, p=0.028; and OR=2.69; 95% CI: 1.46–4.96, p=0.001, respectively), as detailed in Tables 2 and 3.
Table 2
Factors associated with the quit attempts of cigarette smokers (N=287) and smokeless tobacco (snuff) users (N=98), for 12-month and lifetime quit attempts of tobacco users attending DYDH, South Africa, February–July 2023 (N=400)
Table 3
Univariate logistic regression analysis of health behavior of tobacco users attending DYDH, South Africa, February–July 2023 (N=400)
Multivariate regression analysis (Table 4) revealed that individuals who reported good overall health and were unemployed, were more likely to quit smoking within the past year (AOR=3.0; 95% CI: 1.5–5.99, p=0.002; and AOR=3.11; 95% CI: 1.28–7.6, p=0.013, respectively). In relation to lifetime quit attempts, non-pensioners regardless of employment status were significantly more likely to have attempted to quit tobacco use (AOR=2.96; 95% CI: 1.23–6.23, p=0.014; and AOR = 2.96; 95% CI: 1.32–6.65, p=0.008, respectively). Individuals living with HIV were six times more likely (AOR=6.25; 95% CI: 1.93–20.2, p=0.002). Unlike Black or females who were less likely to quit (AOR=0.34; 95% CI: 1.20–1.58, p=0.001; and AOR=0.62; 95% CI: 0.39–0.98, p=0.04, respectively).
Table 4
Multivariate adjusted logistic regression analysis with 12-month and lifetime quit attempts as outcome among tobacco users attending DYDH, South Africa, February–July 2023 (N=400)
DISCUSSION
This study found a low quit rate with poor uptake of healthy behaviors among tobacco users. The findings reveal that most tobacco users were cigarette smokers, with the majority reporting low nicotine dependence. One quarter of respondents had attempted to quit in the past 12 months and almost half of them attempted to quit at least once in their lifetime. The engagement in other healthy behaviors was minimal, with only three respondents consuming ≥5 servings of F&V daily, a third achieved adequate physical activity, and as well abstaining from alcohol. Furthermore, being unemployed, eating at least one portion of F&V daily, having a respiratory problem, living with HIV, and perceiving their general health as good, were all associated with making a quit attempt in the past 12 months among smokers. In contrast, smokers with CVD and respondents who self-identified as female or Black individuals were less likely to attempt quitting. These findings hold significance at clinical, public health level and has the potential to improve outcomes of chronic diseases care, considering that tobacco use (in particular smoking) is a traversal risk factor among most NCDs.
Only a quarter of tobacco users attempted to quit in the past year, a worrying finding in a population with high chronic disease burden as nearly half reporting CVD. This is much lower than previously reported in other studies conducted in South Africa – SASAS data in 2017 (60%)23, in DYDH in 2019 (74%)13, the GATS South Africa in 2021 (40.5%)2, and the United States CDC in 2022 (53%)27. The low quit attempt rate observed in this study may partly be explained by the impact of the COVID-19 pandemic, particularly the temporary ban on cigarette sales in South Africa during 202028,29. Restrictions during the pandemic also limited patient–clinician interactions and reduced brief cessation advice, a key driver of quit attempts30. While the ban initially forced some smokers to reduce or temporarily stop smoking, evidence shows that many relapsed once the ban was lifted and many purchased cigarettes through informal channels28,29. Consequently, when data were collected in 2023, the residual effects of these factors likely contributed to the observed low quit attempt rate in the past 12 months28,29. Notwithstanding the above, this trend was not observed in the USA, where the CDC reported that 53.3% of adult smokers made a quit attempt in 2022, compared with the pre-COVID-19 rate of 55.1% in 201827.
Although cigarette smoking was the predominant form of tobacco use among respondents, the finding that approximately one-fifth reported using smokeless tobacco (SLT) is consistent with previous South African reports, particularly among Black females12,31. Despite representing a minority in this study, only fewer SLT users reported having made a quit attempt, either in their lifetime or within the past 12 months. This is clinically concerning, given that SLT use is associated with an increased risk of adverse health outcomes, including various cancers10,11, and that those who continue to use SLT may constitute a large number at population level. Notably, persistent SLT use occurs despite high awareness of its health risks, with approximatively four out of five South African SLT users reporting knowledge of associated harms (GATS survey 2021)2. This pattern may reflect the addictive potential of SLT products and may partly explain the low prevalence of quit attempts among these users. These study findings therefore underscore the necessity of prioritizing SLT users for tobacco use cessation treatments just as much as cigarette smokers10,11.
Black South Africans were less likely to attempt to quit smoking than other ethnic groups, a finding consistent with previous reports23,32. This finding may be explained by similar significant socio-economic disparities between the ethnic groups in South Africa, such that Black South Africans are predominantly in the lower socio-economic class and experience higher levels of psychosocial distress that in turn result in unhealthy coping mechanisms, including continued cigarette smoking3,8. This highlights the need to implement interventions that address the social determinants of ill health.
The level of tobacco dependence is known to influence both quit attempts and success8,17. Most respondents had low nicotine dependence (LND), aligning with a prior South African study showing that individuals with LND were more likely to intend to quit8. Smokers with LND are generally more amenable to non-pharmacological interventions such as counselling and behavioral therapy7-9 Clinicians, particularly in primary care, should therefore be trained in tobacco dependence treatments, including brief advice and motivational counselling, which are effective in promoting quit attempts and cessation9,13. Tertiary education was linked to LND, which agrees with previous South African studies9,13. A recent publication by the Society for Research into Nicotine and Tobacco suggested that tertiary education significantly reduces tobacco use in low- and middle-income countries3. This may indicated that socioeconomic disparities may translate into differing tobacco cessation interventions in different socioeconomic groups and the nuances may be important for effective lowering of tobacco consumption3.
Regarding F&V consumption, the majority of respondents did not meet WHO recommended intake. Only a very small proportion of participants met the WHO guideline, reflecting extremely low levels of adequate F&V consumption in this population. This finding is consistent with evidence from a recent national household survey in South Africa, which reported that only 0.6% of individuals achieved adequate F&V intake25. Collectively, these findings highlight that insufficient consumption of F&V remains a persistent public health concern in South Africa, reflecting the global trends7,18. Diet poor in F&V carries significant health implications for the prevention and management of NCD, particularly in a population already exposed to other unhealthy risky behaviors such as tobacco use. These findings highlight the necessity for combined interventions addressing diet and tobacco simultaneously. cessation7-9.
Regarding physical activity, individuals with tertiary education, Black respondents, non-pensioners, and those engaging in harmful drinking were found to be more physically active. A recent meta-analysis published in 2023 found that physical activity interventions were beneficial only in the short-term, as they helped reduce cravings and withdrawal symptoms in individuals with HND7,33. This effect, however, did not seem to contribute to long-term smoking cessation. Furthermore, another study found that smokers who were highly physically active were more likely to consume more F&V, have lower nicotine dependence, and were more likely to have successfully quit smoking compared to less physically active smokers34. Although not established in our study, many researchers support the integration of physical activity as a strategy for tobacco cessation and relapse prevention19,34.
Conjoint use of tobacco and alcohol is prevalent in South Africa13,35. In this study, being Black or unemployed was associated with higher likelihood of excessive alcohol consumption. These findings are consistent with other studies conducted in South Africa13,15,35 and may be attributed to the fact that Black South Africans are disproportionately unemployed, and while those employed are likely of other ethnic groups and may have alternative methods to relieve stress, anxiety, and depression related to work, the unemployed who tend to be Black3,4 may not. This inability to cope with stressors and negative emotions may promote continued smoking and act as a barrier to quit attempts3,4. Moreover, several studies have found that smokers with alcohol use disorders tend to have HND and experience difficulties in making a quit attempt16,19. Hence, it is strongly recommended that tobacco users consider stopping or substantially reducing alcohol consumption before or during a quit attempt16,19.
Strengths and limitations
This is a novel study in South Africa to examine the association between quit attempts and concurrent engagement in other healthy behaviors among tobacco users in a clinical setting. This study provides valuable insight into the association between quit attempts and other health behaviors, as well as the potential role of nicotine dependence.
However, several limitations should be considered. Due to the cross-sectional design, causal inferences cannot be made. Additionally, the reliance on self-reported data for key variables may have introduced information bias and potential misclassification of both exposures and outcomes, which could have affected the estimated associations. Although multivariate analyses were conducted to adjust for relevant confounders, residual confounding from unmeasured or inadequately measured variables cannot be excluded. Finally, as the study was conducted among tobacco users attending a hospital-based clinical setting in South Africa, the findings may not be generalizable to the broader population of tobacco users or to non-clinical settings. Despite these limitations, the results provide important insight into the relationship between quit attempts and other health behaviors, including the role of nicotine dependence, among tobacco users in a South African hospital-based setting.
CONCLUSIONS
With only one-quarter of respondents reporting a quit attempt in the past 12 months and nearly one-fifth demonstrating HND, these findings point to considerable gaps in access to tobacco cessation support in South Africa. While a causal relationship cannot be inferred from this cross-sectional study, these results suggest that improved access to evidence-based cessation support, including pharmacotherapy, is needed within the public health system. The observed co-occurrence of tobacco use with unhealthy diet, alcohol use, and physical inactivity also highlights the potential value of integrated approaches to behavior change. These findings suggest that making a quit attempt while concurrently adopting other healthy behaviors is a complex behavioral process. Future studies with larger sample sizes are required to better understand the influence of socioeconomic factors, levels of nicotine dependence, co-existing health risks, and individuals’ perceptions of their health status.
