Factors associated with quit attempt and successful quitting among adults who smoke tobacco in Ethiopia: Global Adult Tobacco Survey (GATS) 2016

INTRODUCTION Tobacco is the leading cause of preventable death in the world. Identification of factors associated with quit attempts and successful quitting can help strengthen tobacco cessation programs. In Ethiopia, no prior study of such factors exists. Our aim was to identify factors associated with quit attempts and successful quitting among adults who smoke tobacco in Ethiopia. METHODS We used the Ethiopian 2016 Global Adult Tobacco Survey (GATS) data (n=10150). GATS is a nationally representative household survey that collects data on sociodemographic and tobacco-related characteristics. We calculated prevalence of reported past 12 months quit attempts and successful quitting and performed logistic regression to obtain prevalence ratios with 95% confidence intervals. A p<0.05 was considered statistically significant. RESULTS Overall 42.0% of people who smoked tobacco made a quit attempt. Men were more likely (APR=3.9; 95% CI: 1.4–10.7) to make a quit attempt compared to women but were less likely to successfully quit (APR=0.6; 95% CI: 0.3–0.9). Those aware of the health harms of tobacco were 2.5 (95% CI: 1.1–5.5) and 3.9 (95% CI: 1.8–8.5) times as likely to make a quit attempt and successfully quit, respectively, than those unaware. Receiving healthcare provider advice to quit was not associated with quit attempts. CONCLUSIONS More than 4 in 10 people smoking tobacco in Ethiopia are making attempts to quit. Receipt of healthcare provider advice to quit is not yet associated with quit attempts in Ethiopia; however, awareness of the health harms of tobacco is a powerful predictor of quit attempt and success in quitting. Improved access to cessation support and expanded awareness of the health harms of tobacco are urgently needed to enhance both quit attempts and success across Ethiopia.


INTRODUCTION
Tobacco use is the leading cause of preventable death globally, killing more than 8 million people a year. More than 7 million of those deaths are due to direct tobacco use, while around 1.2 million are the result of secondhand smoke exposure 1,2 . Over 80% of people who use tobacco live in low-and middle-income countries (LMICs) where the burden of tobacco-related illness and death is heaviest 1,2 . In Ethiopia, an LMIC in eastern Africa, the most recent nationally representative data from 2016 show that 3.7% of individuals aged ≥15 years (6.2% men and 1.2% women) currently smoked tobacco products, and 29.3% and 12.6% of adults were exposed to SHS at the workplace and home, respectively 3 . This indicates that about two and a half million Ethiopian adults are at risk of tobacco-related morbidity and mortality from tobacco use 3 , and that numerous others may be at risk of morbidity and mortality from secondhand exposure. In Ethiopia, 42% of people who smoke tobacco reported attempts to quit in the past year, and over three-fourths (76%) of these attempted to quit without any assistance 3 . People attempting to quit tobacco use face challenges due to nicotine dependence, and there is a high incidence (up to 50%) of relapse after attempting to quit 4,5 ; therefore, it is crucial to establish cessation support for successful quitting 6 .
Evidence shows that availability of tobacco cessation services and cessation support, healthcare provider involvement, and access to pharmacotherapy play critical roles in quit attempts and successful attempts to quit 4,6-8 . Specific recommendations for strengthening tobacco cessation services are provided in the World Health Organization Framework Convention on Tobacco Control (WHO FCTC), an international treaty which Ethiopia has signed 6,9 . Recommendations, cessation services and smoking cessation pharmacotherapies remain severely limited in Ethiopia 10 . Only around a quarter of people who made a quit attempt in Ethiopia used cessation methods: 3.1% used pharmacotherapy, 14.7% used counseling or advice, 5.6% used traditional medicine, and 13.7% used other quit methods 3 .
Identification of factors that could facilitate cessation is important to make best use of available resources, particularly to inform the design of evidence-based smoking cessation interventions 11 . Several studies performed outside Ethiopia have identified factors that are associated with quit attempts and successful quitting including age, educational level, wealth index, receiving advice to quit from a healthcare provider, and awareness of smokingrelated health consequences [12][13][14][15] . No published studies have examined sociodemographic factors and other characteristics that are associated with quit attempts and successful quitting among people who smoke tobacco in Ethiopia. The aim of this study is to assess whether sociodemographic and other factors among people who smoke tobacco are associated with quit attempts and successful quitting, in order to provide the evidence base for development of well-tailored, effective tobacco control strategies and to strengthen cessation programs in Ethiopia.

Data source and study population
Secondary analyses of data from the 2016 Ethiopia Global Adult Tobacco Survey (GATS) 16 were performed. GATS, a component of the Global Tobacco Surveillance System (GTSS), is the global surveillance system for systematically monitoring adult tobacco use and tracking key tobacco control indicators. It is a nationally representative household survey of all non-institutionalized adults aged ≥15 years and provides weighted national estimates of tobacco use and tobacco control indicators 17 . GATS uses a multistage, geographically clustered sampling design and simple random sampling to identify households for participation 18 . A household questionnaire is administered to one adult in the household to determine the number of eligible household members. One household member is then randomly chosen to complete the individual questionnaire, which collects in-depth information on background characteristics, tobacco smoking, smokeless tobacco use, cessation, secondhand smoke, economics, media, and knowledge, attitudes and perceptions of tobacco.

Measures and variables
Definitions and denominators 'Quit attempt' was defined as an attempt to quit within the past 12 months among adults who currently smoked tobacco, or as smoking cessation within the past 12 months among adults who formerly smoked. 'Successful quitting' was defined as smoking cessation 12 months ago or more, among adults who reported ever smoking tobacco. 'People who currently smoked tobacco' were defined as respondents who reported smoking tobacco (cigarettes, cigars, pipes, gaya, and shisha/waterpipe tobacco smoking) daily or less than daily. 'People who formerly smoked' were defined as respondents who reported having formerly smoked tobacco (daily or less than daily) and who were not smoking at the time of interview. 'People who ever smoked tobacco' were defined as including both respondents who formerly and currently smoked tobacco.

Statistical analysis
Analyses were completed using SAS-callable SUDAAN 11.0.1 (RTI International, Research Triangle Park, North Carolina). Descriptive analyses including weighted percentages and 95% confidence intervals were calculated for quit attempt and successful quitting against each independent variable. We examined associations between dependent variables (quit attempt and successful quitting) and each independent variable using univariable logistic regression. We then conducted multivariable logistic regression to evaluate predictors of quit attempts and successful quitting. Both adjusted regression models controlled for gender, age group, place of residence, working status, education level, wealth index, awareness of health harms of tobacco, and ever used smokeless tobacco; the model for quit attempts also controlled for time to first smoke after waking and healthcare provider advice to quit smoking. Adjusted prevalence ratios (APRs) and corresponding 95% confidence intervals are reported. Statistical inferences are based on a significance level of p<0.05. According to the GATS Analysis Package, missing responses are excluded from calculations and indicators with sample size less than 25 have been suppressed 19 . We limited our analyses to the 929 respondents with information on either quit attempts or successful quitting. Table 1 shows the distribution of sociodemographic and other characteristics among people who made a quit attempt and people who successfully quit. Among people who made a quit attempt in the past 12 months, 94.6% (95% CI: 89.6-97.3) were male, more than three-fourths were aged 15  In the unadjusted model, all sociodemographic variables, except for primary education and low through high wealth index, were significantly associated with quit attempts (Table 2). Those with the lowest wealth index were less likely (PR=0.3; 95% CI: 0.2-0.6) to attempt to quit smoking compared to those with the highest wealth index. Those who were aware of health harms of tobacco were more likely to attempt to quit than those who were unaware (PR=3.6; 95% CI: 1.6-8.2). Those who smoked 6-30 minutes (PR=0.2; 95% CI: 0.1-0.4) and 31-60 minutes after walking (PR=0.4; 95% CI: 0.2-0.7) were less likely to attempt to quit smoking compared to those who smoked >60 minutes after waking. Healthcare provider advice to quit smoking and having ever used smokeless tobacco were not significantly associated with quit attempts.

Quit attempt and successful quitting by sociodemographic variables
In the fully adjusted model, the prevalence ratio of quit attempts among men was significantly higher (APR=3.9; 95% CI: 1.4-10.7) than among women ( Table 2). Those who were aged 15-30 years (APR=5.2; 95% CI: 1.1-24.9) and who were aware of health harms of tobacco (APR=2.5; 95% CI: 1.1-5.5) were more likely to attempt to quit than those aged ≥60 years and unaware of health harms of tobacco, respectively. All other demographic variables as well as time to first cigarette after waking, healthcare provider advice to quit smoking, and ever use of smokeless tobacco, were not statistically associated with a quit attempt in the fully adjusted model.

Association of selected demographic characteristics and successful quitting
The overall prevalence of successful quitting among adults who ever smoked was 27.8% (95% CI: 19.8-37.5); 26.8% (95% CI: 20.2-34.7) in men and 32.4% (95% CI: 12.8-60.9) in women (   smoked and were aged 31-44 years were less likely (PR=0.4; 95% CI: 0.2-0.9) to have successfully quit compared to those age ≥60 years (Table 3). Those who were aware of health harms of tobacco were more likely (PR=3.5; 95% CI: 1.0-12.1) to have successfully quit than those who were not aware. There was no significant statistical association in the unadjusted model for gender, residence, working status, education level, wealth index, or having ever used smokeless tobacco with successful quitting of tobacco smoking.
After adjusting for all independent variables, men were significantly less likely to have successfully quit than were women (APR=0.6; 95% CI: 0.3-0.9) ( Table 3). Those aged 31-44 years were less likely (APR=0.4; 95% CI: 0.2-0.8) to have successfully quit compared to those aged ≥60 years. Those who were aware of health harms of tobacco were significantly more likely to have successfully quit (APR=3.9; 95% CI: 1.8-8.5) than those who were not aware. There was no significant statistical association in the fully adjusted model for residence, working status, education level, wealth index, or having ever used smokeless tobacco with successful quitting of tobacco smoking.

DISCUSSION
This study identified factors associated with quit attempts and with successful quitting using Ethiopia's GATS 2016 data. Men were more likely to make a quit attempt, but less likely to successfully quit, than were women. People aged 15-30 years were more likely to have made a quit attempt, and those aged 31-44 years were less likely to have successfully quit than were people aged >60 years. Healthcare provider advice to quit smoking was not associated with quit attempts. Awareness of the health harms of tobacco was significantly associated with both quit attempts and successful quitting. To our knowledge, this is the first study based on the nationally representative data from Ethiopia to assess factors associated with quit attempts and successful quitting. Information from this study can be used to develop well-tailored, effective tobacco control strategies and strengthen cessation programs in Ethiopia.
Gender was associated with both quit attempts and successful quitting in this study; men were more likely to have a quit attempt but less likely to successfully quit than were women, consistent with prior studies conducted across multiple countries 13,[20][21][22][23] . The higher prevalence of quit attempts among men may be due to a higher prevalence among men of exposure to media containing warnings about tobacco use, or a higher prevalence of noticing health warnings on cigarette packages 3 , both of which may motivate quit attempts. Despite a higher prevalence of quit attempts among men, however, this study demonstrated a lower prevalence of successful quitting among men compared to women. This could be related to a potentially greater severity of nicotine dependence among men; for example, the number of cigarettes smoked by men is greater than among women in Ethiopia 3 . This is potentially consistent with a recent systematic review of predictors of quit attempts among adults across high-and middle-income countries, which identified that lower severity of nicotine dependence was the single most consistent predictor of successful quitting 24 . The only other factor related to nicotine dependence examined in this study was time to first smoke after waking, which did not demonstrate a significant association with quit attempts.
Age was also associated with quit attempts and successful quitting in this study, though inconsistently. In unadjusted models, all age groups were more likely to report quit attempts compared with the oldest age group (>60 years), though only those aged 15-30 years were more likely to have made quit attempt in the adjusted model. The higher prevalence of quit attempts among younger age groups might be related to the lower prevalence of daily smoking among younger people compared to older ones 3 , which is consistent with findings from a four-country survey which showed that younger people are more likely and daily smokers less likely to make quit attempts 25 . All age groups, compared with the oldest age group (>60 years) had an APR of less than 1 in relation to successful quitting, though the association was only significant for the age group 31-44 years. While the association was not significant across all age groups in our study, the general pattern is consistent with other studies which have found that the chance of successful quitting increases in older age groups 13,14,21,22,26  diseases increases with age, which may motivate people of older ages to quit smoking in order to reduce the negative health effects of tobacco use 27,28 .
Other factors we assessed included healthcare provider advice to quit smoking and awareness of the health harms of tobacco. Unexpectedly, healthcare provider advice to quit smoking was not associated with making a quit attempt in this study. This is inconsistent with studies from other countries which showed a positive relationship between quit attempts and healthcare provider advice 13,29,30 . Prior studies have shown that simple advice to quit tobacco coupled with the use of cessation aids increases the chances of both quit attempts and successful quitting 31,32 . Evidence also shows that healthcare provider advice to quit tobacco increased the utilization of different cessation services 33 . The lack of association between healthcare provider advice and quit attempts in this study could be due to the small number of study participants who had visited a healthcare provider (n=180), a possible lack of knowledge among healthcare providers on tobacco cessation counseling services, or limited availability of cessation aids and services in Ethiopia. Lack of training of healthcare providers on cessation services and counseling, and time constraints in routine healthcare service, have been identified as barriers affecting the standard counseling services at healthcare facilities 3,34,35 . In Ethiopia, the majority (three-fourths) of people who attempted to quit smoking did so without any assistance, and few received cessation aid services 3 . Health system strategies to get assistance to Ethiopians attempting to quit might include: integrating cessation counseling and support into standard healthcare visits; offering cessation support through quitlines, mobile apps, or cessation-specific clinics; and expanding cessation training for all health care providers 4,31,32,36 .
Awareness of the health harms of tobacco was of significant impact in this study: those who were aware of health harms of tobacco were 2.5 times and 4 times as likely to make quit attempts and successfully quit, respectively, than those who were not aware. This finding is consistent with previous research 13,14,21,22,26,29 , and underscores the importance of warning populations of the health harms of tobacco 6 . In Ethiopia, cigarette packs are required to display effective warning labels that contain messages about the harms of tobacco use and large warnings with all appropriate characteristics are displayed on cigarette packs 34,37 . Notably, however, less than half of those who currently smoked noticed health warning signs on cigarettes 3 , which indicates an opportunity to increase the visibility of these warnings. The impact of awareness of the health harms of tobacco on successful quitting also suggests potential benefit of broader education, outreach, and public communication to raise awareness of these health harms of tobacco across Ethiopia. Creating a high level of awareness on the health risks of all types of tobacco use across all age groups, sexes, and places of residence is one of the major objectives in Ethiopia's tobacco control strategic plan 37 . The strategies set forward to achieve this goal include ensuring effective pack warning labels, implementing counter-advertising campaigns through mass media, and regulation and disclosure of the content of tobacco products 37 . Public health campaigns aimed at improving public awareness on harms of tobacco use have begun in collaboration with civil societies, WHO, Ministry of Health, and regional health bureaus in Ethiopia. In addition, awareness creation activities were completed during WHO's 2021 World No Tobacco Day celebration, which had the theme 'Commit to Quit' tobacco use 38 .

Limitations
This study has limitations. First, the cross-sectional nature of the data limits our ability to draw conclusions about the directionality of the findings. Second, the study relied on self-reported, closeended answers to questions on sensitive issues like tobacco smoking status, tobacco smoking quit attempts, and utilization of cessation services; social desirability bias in responses is possible here as a result of perceived stigma. Third, due to the absence of motivation-related questions in the questionnaire, we were unable to assess motivational factors such as confidence in quitting, which could be an important determinant in quit rates and successful quitting 24 .
Fourth, some small sample sizes (people who had recently visited a healthcare provider, for example) may have limited the power of this study to detect some potential associations. Despite these limitations, to our knowledge, this is the first study to assess the

CONCLUSIONS
This study revealed certain sociodemographic and other factors that predicted quit attempts and successful quitting. Men were more likely to have a quit attempt but less likely to successfully quit, and awareness of health harms of tobacco was associated with quit attempts and successful quitting. Healthcare provider advice to quit smoking was not associated with quit attempts in our study. This study provides evidence to expand access to cessation support for people attempting to quit and for all people who use tobacco in Ethiopia, expanding healthcare provider and healthcare system capacity to encourage cessation, and suggests the potentially powerful impact of expanding public awareness of the health harms of tobacco.