Study design, setting, and participants

This study employed a case-control design to compare successful quitters and unsuccessful quitters regarding the smoking cessation methods they used, quitters’ behaviors during the nicotine withdrawal and relapse prevention periods, supporting factors and obstacles, and other influencing factors. We conducted our study in Nakorn Sri Thammarat and Trung provinces (both in the Southern part of Thailand Province) for two reasons. First, these two provinces had the fifth (with the prevalence of 27.4%) and the ninth (26.4%) highest smoking prevalence, respectively, out of 77 provinces in Thailand, based on Thailand’s national tobacco consumption survey in 2017¹⁴. Second, they were conveniently located for the research team to travel to collect the data. We employed a disproportionate stratified random sampling method. In each province, one big and one small district were randomly selected from groupings based on population size (simple random sampling within the two groups). Furthermore, we used the same sampling method to choose two big and two small sub-districts from each district. In each sub-district, we then applied a simple random sampling method to select one community. We collected data from all participants in 16 communities.

The study participants were daily smokers (smoking at least one cigarette per day) at some point in their lifetime. Furthermore, they had tried to quit smoking (stopped smoking for at least one day) at least once in their lifetime. There was a two-step recruitment process. In the first step, the community health volunteers screened all community members aged ≥15 years in 16 studied communities according to three inclusion criteria: 1) they had smoked at least 100 cigarettes over their lifetime, 2) had once been regular smokers; and 3) had at least once attempted to quit smoking. People who said ‘no’ to any one of these criteria were not eligible for the study. Those who said ‘yes’ to all these three questions were asked a fourth question regarding their current smoking behaviors. Those who currently smoked were identified as ‘unsuccessful quitters’ while those who had quit smoking for at least six months were identified as ‘successful quitters’. However, those who had stopped smoking for less than six months were also excluded from our study because they were not yet clearly successful quitters. If the eligible community members allowed it, the community health volunteers would give prospective study participants’ contact information to the research team to invite them to participate in the study. In the second step, the researchers interviewed the eligible participants using the structured questionnaire, after explaining the study protocol and obtaining their consent. The main variables we assessed were participants’ sociodemographic data, smoking behaviors, and smoking cessation experiences of the last quit attempt, which included the cessation method used, the quitter’s behaviors during the cessation attempt period, and other potential influencing factors for the successful quit attempt. Data were collected from August to December 2020.

Measures and data collection

The outcome measure in our study was the result of the last smoking cessation attempt. We defined ex-smokers who had quit smoking for at least six months as ‘successful quitters’, and we defined current smokers who had stopped smoking for less than six months in their last attempt as ‘unsuccessful quitters’. We measured various independent variables which included sociodemographic data, smoking behaviors, motivation for quitting, withdrawal symptoms after quit attempt, smokers’ techniques supporting quitting and preventing relapse, and other influencing factors.

We did test-retest reliability test for our questionnaire among 10 successful quitters and 10 unsuccessful quitters before implementing our survey. Our original questionnaire contained 110 sub-questions (in the 39-question questionnaire). There were 102 sub-questions for which the participants’ answers were perfectly matched for the test and the retest. Eight sub-questions had some discrepancies in participants’ answers; however, only five sub-questions were considered unreliable. We did not keep these five sub-questions in our final analyses.

All interviewers were trained and supervised by study researchers. Trained field staff collected data from participants using a 39-question structured interview questionnaire developed by the research team covering all the abovementioned variables. Each interview lasted approximately 10–15 minutes.

Data analysis

For univariate analysis (Table 1), we calculated means with 95% confidence intervals (CI) for the continuous variables (with t-tests to compare the difference of these variables’ values between the successful and unsuccessful quitters), or tabulated frequencies and percentages for the categorical variables using chi-squared tests. Furthermore, case-control data were analyzed using a multiple logistic regression (Table 2) to estimate multivariable-adjusted odds ratios (ORs) and 95% CI of the various independent adjusted variables, including age (16–44, 45–60, and ≥60 years); marital status (married, single/widow), occupation (agriculture, freelance labor, merchant, unemployed, and other), residence (urban, rural), yearly income (0–60000, 60001–120000, and >120000 Thai Baht), smoking behaviors (regular, occasional), tobacco product types used (cigarette, roll-your-own), and whether or not a doctor had recommended that the participant stop smoking because of smoker’s illness (yes, no), as contributors to a successful smoking cessation outcome. We called the logistic regression that included all the mentioned variables ‘Model 1’. We did a number of logistic regression modellings by adding one additional variable at a time to the full Model 1 from a few sets of factors influencing successful quit attempts. These sets of factors included the cessation method used, technique used if employing self-cessation, motivation to quit, and withdrawal symptoms after stopping smoking (Table 3); and techniques supporting quitting, techniques preventing relapse, and other factors (Table 4). All tables show only variables that have at least 10 observations for either successful or unsuccessful quitters, and are reliable questions. The reason we added one influencing factor at a time into our logistic regression analyses was that these factors are not independent. Statistical analyses were conducted using STATA version 15.0.

Smoking cessation methods and techniques

We found that almost all successful and unsuccessful Thai quitters in our study employed unaided methods in their quit attempts. An unaided method means smoking cessation accomplished without the assistance from others or involving the use of medication. Successful quitters with unaided methods employed a cold turkey technique (abruptly stopping smoking) significantly more often than a cut-down-to-quit technique.

Even though there is evidence suggesting that both behavioral counselling and smoking cessation medical interventions increase success in smoking cessation compared with self-help materials or no treatment³, evidence in both Thailand and internationally has shown that unaided methods are crucially important in the real world of smoking cessation implementation. Based on the Thai national smoking survey, among successful ex-smokers, 94.0% quit smoking by themselves, and only 6.0% used some form of help to quit⁴. These 6.0% consisted of: 1.9% using medication, 0.6% using the national quit help line counseling service, 1.8% receiving smoking cessation services from health personnel, and 1.7% using other methods⁴. Similarly, only 30% of tobacco users worldwide had access to tools that could help them stop smoking, although approximately 60% said they wanted to quit¹⁵. Other studies also showed that many people are able to quit by unassisted strategies such as the ‘cold turkey’ method or cutting down and then quitting¹⁶. This means that unaided methods are crucially important in the real world of smoking cessation implementation.

Demographic data and smoking behaviors

Based on our findings, successful quitters were those aged ≥45 years and unemployed. A Korean study, involving males, also found that middle aged smokers (aged 40–64 years) and older smokers (aged ≥65 years) had a higher chance of successful quitting than younger smokers (aged 19–39 years)¹⁷. The reason was that older male Korean smokers have experienced more illness than younger ones, leading to a higher quit attempt rate and successful smoking cessation rate. The illnesses which had a positive effect on a successful smoking attempt included cardiovascular diseases and higher blood pressure^8,17, as well as high body mass index¹⁷. Since expensive cigarettes reduce tobacco consumption³, unemployed people in our study may have had higher motivation to quit than their employed counterparts for financial reasons.

Our study found that occasional smokers were more successful in their quit attempts than regular smokers. Similarly, the International Tobacco Control Southeast Asia (ITC-SEA Thailand) survey showed that Thai regular and heavy cigarette smokers had a higher relapse rate than occasional smokers, and that relapse was common after one month of quitting¹⁸. Using roll-your-own (RYO) (rather than commercial cigarettes) was a barrier to quitting in our study. As seen from the Global Adult Tobacco Survey conducted in Thailand in 2009, Thai men who smoked RYO were more likely to report not being interested in quitting compared to those who smoked manufactured cigarettes as there were fewer financial costs associated with RYO products¹⁹. Overall, smokers who were of older age, unemployed, smoked occasionally, or used commercial cigarettes (rather than RYO) had a higher chance of successful smoking cessation than their counterparts.

Motivational factors

Our study found that tangible motivational factors were more influential than intangible motivational factors. For instance, a doctor’s recommendation to stop smoking due to a smoker’s illness, or having a grandchild or a child more heavily impacted a successful quitting attempt than quit attempts motivated by oneself, families, relatives, or friends.

Concordant with our finding of the importance of a smoker’s illness to motivate them to quit, the Surgeon General Report 2020 on smoking cessation concluded that there is sufficient global evidence to suggest that certain life events (e.g. hospitalization, surgery, and lung cancer screening) can trigger smoking cessation attempts, smoking cessation treatment uptake, and successful cessation³. Health concerns are also an influential factor for cessation²⁰. Often a new mother will ask the father to quit smoking²⁰, and since half of children and grandchildren of successful quitters were four years of age (data are not shown), ‘well-baby’ clinics may be good targets in health facilities to advise smokers to quit. Buczkowski et al.²¹ found that smokers largely quit smoking spontaneously without any planning involved.

Our study found that self-motivation predicted failed quit attempts: in fact, 38.5% of unsuccessful quitters and 27.1% of the successful quitters had self-motivation to quit. A possible explanation is that self-motivation is not a strong positive influencing factor for a successful quit attempt because it is often unstable, even over a short period of time (<14 days). As a result, most quit attempts end in relapse, with an estimated average of 30 or more quit attempts undertaken by a smoker before eventual success²². Hence, employing external, tangible motivators (e.g. doctors’ recommendation over a smokers’ illness, having a child or grandchild) to encourage smokers to quit is crucial in smoking cessation.

Withdrawal symptoms

Our study found that even though severe cravings (66.2% of successful quitters and 71.8% of unsuccessful quitters) and irritable mood/anger / stress (35.6% of successful quitters and 39.0% of unsuccessful quitters) were the most common nicotine withdrawal symptoms among our study participants, they did not actually cause failure in quit attempts (Table 3). In fact, we found that insomnia (6.0% of successful quitters and 14.0% of unsuccessful quitters) significantly reduced the chance of successful quit attempts, similar to the findings of Ashare et al.²³. This sleep-disturbance symptom, either as a smoker characteristic or as a nicotine withdrawal symptom, is common among smokers and predicts smoking cessation failure and must therefore be reduced for successful smoking cessation²³.

Quitters’ having a good appetite after quitting increased the chance of success of smoking cessation in our study; this, of course may lead to weight gain. A study found that an average weight increase of 3.8–6.8 kg, with a minimum significant weight gain of 1–16 kg, has been reported by patients who successfully stopped smoking²⁴.

Techniques supporting successful quitting during the nicotine withdrawal phase and techniques preventing relapse if quitters make it through the nicotine withdrawal phase

Our study found that exercise worked well in both the nicotine withdrawal and the relapse prevention phases, while avoiding smokers was a good strategy to employ during the relapse prevention phase.

Since chronic exposure to nicotine changes the mesocorticolimbic dopamine system, an area of the brain which processes rewarding sensory stimuli, the altered brain condition requires nicotine to be maintained in order the brain to produce reward and prevent withdrawal symptoms²⁵. Aerobic exercise increases dopamine and serotonin, leading to mood improvement²⁶, and it was shown to predict successful smoking cessation, both in the nicotine withdrawal and the relapse prevention phases of our study. Evidently, in our study, exercising during the nicotine withdrawal phase increased the chances of successful smoking cessation with an OR of 2.00, and with an OR of 13.87 if the quitters performed exercise during the relapse prevention phase.

Supporting our finding, Shiffman et al.¹⁰ examined 21539 smoking episodes and 26930 non-smoking episodes among 212 occasional smokers and 192 daily smokers and found that avoiding smokers could prevent indulgent smoking. In practice, a smoking cessation guideline should suggest that quitters do exercise in both the withdrawal and relapse prevention phases, and to avoid being around smokers during the relapse prevention phase.

Other influencing factors

Other success-influencing factors included self-efficacy in smoking cessation, feeling that tobacco is expensive, a smoking prohibition in public places, a smoking prohibition in workplace, and a desire to avoid the unpleasant smell of other people’s smoking once they had stopped smoking themselves.

Our study provides more individual-level evidence to confirm the effectiveness of population-level policy interventions in enhancing smoking cessation. The policies used to increase the smoking cessation rate include: raising cigarette prices, employing comprehensive smoke-free policies, implementing mass media campaigns, requiring pictorial health warnings on packaging, maintaining comprehensive nationwide tobacco control programs, and providing insurance coverage for smoking cessation treatment that is comprehensive³. Thus, our study confirmed that implementing the WHO’s best buys policy interventions – to reduce affordability by increasing excise taxes and prices on tobacco products, and eliminate exposure to secondhand tobacco smoke in all indoor workplaces, public places, and public transportation¹⁵ – not only reduced access to tobacco products and smoke but also increased the chances of successfully quitting smoking. Based on their qualitative study with 15 current smokers and 16 former smokers, Buczkowski et al.²⁰ summarized that two very crucial motivations for smoking cessation were a smoking ban at home and at work due to non-smokers’ wishes and rules, as well as the high cost of cigarettes. When smokers successfully quit smoking themselves, smelling the smoke coming from smokers can become an aversive effect and, in fact, strengthen their resolve not to relapse to smoking again²⁷.

Conversely, our study found that other factors that reduced quitting success included pleasure enhancement or the reduction in stress brought about by smoking, the social pressure to smoke, and associating smoking with another activity.

Buczkowski et al.²⁰ found that stress and the need to reduce it by smoking a cigarette and the need to experience the smoking-connected pleasure were reasons why smokers sometimes relapsed. A study showed that in an environment where there is social pressure to quit smoking, individuals practice abstinence from smoking for a longer duration²⁸. Keeping away from the influence of peers who smoke was helpful in smoking cessation attempts²⁹. Tiffany et al.³⁰ also highlighted the importance of behaviors during smoking cessation attempts, as those who tend to smoke directly after activities such as exercise or dinner may continue to smoke habitually without much craving prompting it.

Strengths and limitations

Strengths of our study include: 1) we assessed the characteristics of the last smoking cessation attempt rather than the characteristics of the quitter, which has resulted in a study which can provide us with insights regarding practical positive influencing factors that can be employed in the real world; 2) we examined the effects of a comprehensive set of influencing factors of successful smoking cessation and found a number of interesting positive influencing factors, which can be used in developing smoking cessation guidelines, programs, and policies; 3) we separated influencing factors in the withdrawal and the relapse prevention phases, and found that some factors functioned differently between two phases; and 4) we employed the case-control design which is appropriate for studying the real past smoking cessation experiences between successful and unsuccessful quitters. However, our study also has some limitations. They include: 1) there may be some recall bias embedded in the case-control study design; and 2) our study had too few samples who used quitting aids (medication or counseling), hence we could not compare the positive influencing factors between aided and unaided methods, or between the ‘stopping smoking abruptly (cold turkey)’ and the ‘cutting down and then quitting’ methods. Furthermore, having only small samples of people who engaged in some influencing factors made it impossible to analyze the effects of a few potentially interesting factors, including avoiding being around tobacco products or stressful experiences.