In January 2015, South Korea's government raised the cigarette tax, and the retail price of cigarettes abruptly increased by 80% compared to the previous year. This research aimed to determine the effect of this increase on smoking cessation among South Korean smokers.
We analyzed data collected by the 2013–2015 South Korea National Health and Nutrition Examination Survey of 15,203 South Koreans over 19 years old using regression analysis. We examined the recent non-smoking period of nonsmoking people, prepared according to the survey, and analyzed the recent smoking cessation ratio.
Among smokers, from 2013 to 2014, the smoking cessation rate was 7.2%, and it increased to 9.9% in 2015 after the increase in the cigarette tax. In 2015, the recent smoking cessation rate was higher among people over the age of 60 (odds ratio [OR], 2.67) compared to those between the ages of 40 and 49. The recent smoking cessation rate was higher among people with below elementary education (OR, 2.28) and above university education (OR, 1.94) compared to high school, higher for those with apartments (OR, 1.74) compared to general type residences, and higher among those with a household income in the low-middle quartile (Q2) (OR, 2.32) compared to the highest quartile (Q4).
This innovative policy including increase in cigarette prices affected smoking cessation, and its impact varied by sub-group of smokers in South Korea.
Consistent with the global trend and based on inflation, South Korea has gradually increased its cigarette tax by 10%–30% in an attempt to reduce the smoking rate
The South Korea National Health and Nutrition Examination Survey (KNHANES) is a large-scale annual survey conducted systemically by the South Korea Centers for Disease Control and Prevention (KCDC). The data in the KNHANES represents the entire general population of South Korea
The goal of this study was to determine the effects of the innovative policy including the increase in tobacco prices and expansion of smoke-free areas on cigarette smoking cessation among South Korean smokers, and which sub-groups of smokers more readily responded to the new policy.
We used data from the KNHANES, a national survey performed in 2013–2015. The survey used complex probability procedures to provide a sample representative of the South Korean general population, with stratification and multiple stages of cluster selection using age, sex, residence type, education level, and other variables. We followed the guidelines for reporting sample weight (sampling weights) and stratification as indicated by the KCDC; this information is available on the KNHANES website (
Among 20,482 subjects surveyed in 2013–2015, a total of 15,203 subjects ≥19 years of age responded to the health survey, including a questionnaire concerning smoking habits. Using regression analysis, 3,086 subjects who were current smokers and who had smoked more than 100 cigarettes in their lifetime or who had stopped smoking within the past six months were examined to detect the significant factors for recent smoking cessation. Recent smoking cessation was defined as having stopped smoking within the past six months: they answered, “I smoked cigarettes in the past but do not smoke now” to a question asking “Current smoking status,” and gave a variable period less than 6 months to a question asking “Non-smoking period.” The recent smoking cessation rate (%) was calculated as follows: subjects with recent smoking cessation×100 (%)/total enrolled subjects (current smoker and ever smoker including subjects with recent smoking cessation). This retrospective study was approved by the Institutional Review Board (IRB) of Yong-in Severance Hospital (No. 3-2017-0020). The requirement of informed consent was waived considering the retrospective nature of this study.
All data were collected by self-reported questionnaires, including diabetes status, hypertension status, marital status, education level, residence type, and household income. Household income was determined based on each subject's self-reported monthly household income in South Korea. Income level was categorized into four quartiles. Body mass index (BMI; measured as body weight divided by the square of the body weight, and expressed in units of kg/m2) was measured by the healthcare provider.
We used KNHANES stratification variables and sampling weights. Differences in categorized variables and continuous variables from before and after the policy change were analyzed by chi-square tests and t-tests, respectively. Univariate and multivariate logistic regression analyses were conducted to identify factors associated with recent smoking cessation.
The interaction odds ratio (OR) value refers to the relative size of the OR value after the increase of cigarette tax in preparation for a before-tax change. This was used to determine if there were significant differences in the critical factors; that is, if the interaction OR value is greater than 1, it is a positive effect that relatively increases the smoking cessation rate, and if it is smaller than 1, it has a negative effect that has a relatively lower smoking cessation rate.
We enrolled 10,219 subjects (representing a population of 23,879,830) and 4,984 subjects (representing a population of 12,071,069) in 2013–2014 (before the policy change) and in 2015 (after the policy change), respectively. Subjects from before and after the price increase policy were not significantly different in clinical characteristics, including sex, age, height, weight, BMI, diabetes, hypertension, marital status, education level, type of residence, and household income (
We analyzed data of 3,086 smokers to define the effects of policy change on smokers. Cigarette prices had not changed for ten years before the announcement of the new policy; however, in January 2015, the government abruptly increased the price of cigarettes (
We defined the significant factors for subjects who recently stopped smoking within the past six months (recent smoking cessation) among 3,086 smokers. In the univariate analysis, age, BMI, hypertension, and marital status were significant factors contributing to recent smoking cessation before the new policy. The multivariate analysis illustrated that marital status was a significant factor for recent smoking cessation (OR, 0.36; 95% confidence interval [CI], 0.28–0.74; p=0.002) (
In the univariate analysis, age, hypertension, education level, residence type, and household income were significant factors for recent smoking cessation after the policy change. The multivariate analysis illustrated that the significant factors for recent smoking cessation were: subjects ≥60 years old (OR, 2.82; 95% CI, 1.15–6.91; p=0.023) compared to those 40–49 years old; education levels of below elementary (OR, 2.28; 95% CI, 1.00–4.95; p=0.049) and above university (OR, 2.17; 95% CI, 1.08–4.34; p=0.029) compared to high school education; and subjects living in apartment residences (OR, 1.77; 95% CI, 1.02–3.08; p=0.043) compared to general type residences. In addition, subjects with low-middle quartile (Q2) household income were more likely to stop smoking compared to subjects with the highest quartile household income (OR, 3.03; 95% CI, 1.40–6.58; p=0.005). Although there is no significant difference, subjects with the lowest quartile (Q1) household income were more likely to stop smoking than subjects with the highest quartile (Q4) household income (OR, 1.68; 95% CI, 0.70–4.01, p=0.243) (
We obtained the interaction OR and p-values to determine whether there is a significant gap between factors before and after the policy change. Subjects between 19 and 29 years of age were more likely to stop smoking compared to those between 40 and 49 before the policy change (OR, 3.89). After the policy went into effect, the recent smoking cessation rate was still higher for people aged between 19 and 29 years of age (OR, 1.82), but it is decreasing. (interaction OR, 0.48; interaction p=0.197). However, in women, compared to men, the smoking rate increased (OR, 6.05) after the price increase (
We found that the innovative policy change by the South Korean government was successful in encouraging smokers to stop smoking. The stop smoking rate was 7.2% before the policy change but significantly increased to 9.9% per year after the policy change (an increase of 27.3%). In addition, many smokers (39.6%) commented that they had stopped or reduced smoking because of the 2015 policy.
Cigarette price increase policies have been considered as one of the most effective strategies for smoking cessation
Smoking rates are higher in subjects with lower socioeconomic status. In South Korea, subjects with lower household income are 1.3–2.3 times more likely to smoke cigarettes, compared to subjects with the highest household income
Although a low socioeconomic level is a well-known susceptibility factor for smoking cessation because of an increase in cigarette prices, higher BMI, diabetes, education level, and marital status are not well-known factors. Previous studies abroad have reported that young adults more easily respond to cigarette price increase policies
There are several ways to stop smoking. Electronic cigarettes (EC) are considered as one of the ways to stop smoking. In one review article, EC helped people to stop smoking for 6 to 12 months compared to a placebo (4% vs. EC 9%), but the result is rated ‘low’ by GRADE standards
The policy to expand smoke-free areas was implemented to reduce exposure to second-hand smoke. Previous studies have revealed the positive effects of smoke-free area policy on smoking rate and health
We are well aware that smoking is a significant risk factor for important diseases, including lung cancer and asthma
The KNHANES data are obtained from a well-designed national program with complex, multi-stage probability sample extraction. We also used a complex sample analysis, recommended by KNHANES; therefore, this data represents 46,946,471 South Koreans, which is close to the total population of South Korea
The study has some limitations. First, this is a cross-sectional observational study, not a longitudinal cohort study. Therefore, analysis of the smoking cessation rate was conducted using different populations in 2013, 2014, and 2015. As a result, we should be careful in interpreting the results of this study. Second, we used an operational definition of recent smoking cessation. This measurement of recent smoking cessation may not fully reflect the real recent smoking cessation caused by the policy change. Third, we could not analyze the potential variables that affect smoking cessation, including the severity of underlying respiratory disease and lung function, and distances between homes and cigarette stores
In conclusion, using a large amount of national scale data, this study confirmed that an innovative policy change positively affected smoking cessation of current smokers in South Korea. In addition, this positive effect was more pronounced in people who were married and people with a high BMI, diabetes, below elementary education level, and with low household incomes.
Many countries have adopted cigarette price increase policies, and these have had a significant impact on the number of individuals who have stopped smoking. This is the first study to determine the positive effects of the national cigarette price increase policy conducted by South Korea's government. The retail price of cigarettes abruptly increased by 80% compared to the previous year, and this increased the number of people who stopped smoking. This positive effect was more pronounced in people who are ≥60 years of age, married, and with low household incomes. The findings suggest that this innovative cigarette increase policy should be maintained over the long term to retain this positive effect on the cessation of cigarette smoking in South Korea.
Significant factors for recent smoking cessation before policy change (2013–2014) among women smokers
Significant factors for recent smoking cessation after policy change (2015) among women smokers
Significant factors for recent smoking cessation before policy change (2013–2014) among men smokers
Significant factors for recent smoking cessation after policy change (2015) among men smokers
Variable | Before increase policy (2013–2014) | After increase policy (2015) | p-value |
---|---|---|---|
Male sex, n (%) | 5,068 (49.6) | 2,487 (49.9) | 0.836 |
Age, yr | 0.806 | ||
19–29 | 1,982 (19.4) | 951 (19.1) | |
30–39 | 1,992 (19.5) | 927 (18.6) | |
40–49 | 2,125 (20.8) | 1,026 (20.6) | |
50–59 | 1,982 (19.4) | 981 (19.7) | |
≥60 | 2,135 (20.9) | 1,096 (22.0) | |
Height, cm | 164.2±0.1 | 164.3±0.2 | 0.195 |
Weight, kg | 64.7±0.1 | 64.0±0.2 | 0.170 |
BMI, kg/m2 | 23.9±0.1 | 23.7±0.1 | 0.296 |
Diabetes | 950 (9.3) | 428 (8.6) | 0.303 |
Hypertension | 2,482 (24.3) | 1,290 (25.9) | 0.140 |
Married | 7,918 (77.5) | 1,513 (76.3) | 0.343 |
Education level | 0.247 | ||
Below elementary | 1,716 (16.8) | 797 (16.1) | |
Middle school | 939 (9.2) | 433 (8.7) | |
High school | 3,218 (31.5) | 1,869 (37.5) | |
Above university | 3,606 (35.3) | 1,883 (37.8) | |
Type of residence | 0.694 | ||
General type | 5,251 (51.4) | 2,467 (49.5) | |
Apartment | 4,965 (48.6) | 2,516 (50.5) | |
Household income | 0.551 | ||
Lowest quartile (Q1) | 1,532 (15.0) | 757 (15.2) | |
Low–middle quartile (Q2) | 2,615 (25.6) | 1,156 (23.2) | |
High–middle quartile (Q3) | 2,983 (29.2) | 1,500 (30.1) | |
Highest quartile (Q4) | 3,085 (30.2) | 1,569 (31.5) | |
Enrolled number | 10,219 (100) | 4,984 (100) | |
Representing number | 23,879,830 | 12,071,069 |
Values are presented as percentage or mean±SD unless otherwise indicated.
Representing number of recent smoking cessation group | Representing number of current smoker group | Univariate analysis | Multivariate analysis | |||||
---|---|---|---|---|---|---|---|---|
OR | 95% CI | p-value | OR | 95% CI | p-value | |||
Sex | ||||||||
Male | 376,663 (85.7) | 4,952,193 (87.0) | 0.90 | 0.54–1.49 | 0.675 | |||
Female | 62,912 (14.3) | 742,908 (13.0) | 1 | |||||
Age, yr | ||||||||
19–29 | 191,027 (43.5) | 1,117,223 (19.9) | 3.89 | 2.09–7.24 | <0.001 | 1.97 | 0.92–1.02 | 0.052 |
30–39 | 91,489 (20.8) | 1,461,478 (25.7) | 1.43 | 0.74–2.74 | 0.287 | 1.11 | 0.57–2.18 | 0.753 |
40–49 | 62,734 (14.3) | 1,428,726 (25.1) | 1 | 1 | ||||
50–59 | 44,523 (0.1) | 1,001,436 (17.6) | 1.01 | 0.49–2.10 | 0.973 | 1.05 | 0.50–2.19 | 0.894 |
≥60 | 49,800 (11.3) | 686,236 (12.0) | 1.65 | 0.91–3.02 | 0.102 | 1.79 | 0.95–3.36 | 1.788 |
BMI | 23.33±0.31 | 24.10±0.09 | 0.94 | 0.89–0.99 | 0.022 | 0.97 | 0.92–1.02 | 0.194 |
Diabetes | 30,351 (7.6) | 553,483 (10.6) | 0.69 | 0.34–1.39 | 0.301 | |||
Hypertension | 60,761 (13.9) | 1,277,485 (22.5) | 0.56 | 0.37–0.92 | 0.023 | 0.82 | 0.49–1.39 | 0.472 |
Married | 205,430 (46.7) | 4,196,181 (73.7) | 0.31 | 0.21–0.48 | <0.001 | 0.36 | 0.28–0.74 | 0.002 |
EC user for the last month | 42,058 (42.1) | 319,607 (24.2) | 0.98 | 0.58–1.65 | 0.976 | |||
Asthma history | 12,589 (2.9) | 155,585 (2.7) | 1.05 | 0.37–2.94 | 0.926 | |||
Education level | ||||||||
Below elementary | 31,924 (7.3) | 596,618 (10.5) | 0.85 | 0.44–1.66 | 0.642 | |||
Middle school | 42,535 (9.7) | 498,676 (8.8) | 1.36 | 0.67–2.77 | 0.393 | |||
High school | 160,384 (36.5) | 2,560,539 (45.0) | 1 | |||||
Above university | 204,730 (46.6) | 2,039,266 (35.8) | 1.6 | 0.96–2.68 | 0.072 | |||
Type of residence | ||||||||
General type | 118,737 (43.3) | 1,428,219 (57.0) | 1 | |||||
Apartment | 155,243 (56.7) | 1,075,412 (43.0) | 1.17 | 0.77–1.76 | 0.469 | |||
Household income | ||||||||
Lowest quartile (Q1) | 69,654 (15.8) | 725,621 (12.7) | 1.14 | 0.60–2.15 | 0.695 | |||
Low–middle quartile (Q2) | 113,910 (25.9) | 1,504,570 (26.4) | 0.90 | 0.52–1.56 | 0.697 | |||
High–middle quartile (Q3) | 119,522 (37.2) | 1,849,396 (32.5) | 0.77 | 0.46–1.27 | 0.301 | |||
Highest quartile (Q4) | 136,487 (31.0) | 1,615,513 (28.4) | 1 |
Values are presented as number (%) or mean±SD unless otherwise indicated.
OR: odds ratio; CI: confidence interval; BMI: body mass index; SD: standard deviation.
Representing number of recent smoking cessation group | Representing of current smoker group | Univariate analysis | Multivariate analysis | |||||
---|---|---|---|---|---|---|---|---|
OR | 95% CI | p-value | OR | 95% CI | p-value | |||
Sex | ||||||||
Male | 235,204 (85.8) | 2,184,381 (87.1) | 0.89 | 0.46–1.71 | 0.718 | |||
Female | 38,778 (14.2) | 319,252 (12.8) | 1 | |||||
Age, yr | ||||||||
19–29 | 58,123 (21.2) | 542,971 (21.7) | 1.82 | 0.68–4.84 | 0.230 | 1.76 | 0.64–4.83 | 0.269 |
30–39 | 87,071 (31.8) | 589,032 (23.5) | 2.51 | 0.93–6.79 | 0.069 | 1.71 | 0.60–4.86 | 0.313 |
40–49 | 35,623 (13.0) | 605,034 (24.2) | 1 | 1 | ||||
50–59 | 48,422 (17.7) | 481,869 (19.2) | 1.71 | 0.64–4.53 | 0.282 | 1.91 | 0.72–5.08 | 0.197 |
≥60 | 44,741 (16.3) | 284,724 (11.4) | 2.67 | 1.29–6.32 | 0.026 | 2.82 | 1.15–6.91 | 0.023 |
BMI | 24.54±0.52 | 24.28±0.15 | 1.02 | 0.95–1.09 | 0.629 | |||
Diabetes | 32,055 (12.1) | 204,627 (9.0) | 1.39 | 0.65–2.97 | 0.393 | |||
Hypertension | 42,953 (16.7) | 661,728 (26.6) | 0.56 | 0.32–0.97 | 0.037 | 0.44 | 0.24–0.79 | 0.007 |
Married | 215,299 (78.6) | 1,674,891 (66.9) | 1.82 | 0.93–3.55 | 0.081 | |||
EC user for the last month | 24,310 (27.6) | 400,889 (41.9) | 0.77 | 0.47–1.27 | 0.302 | |||
Asthma history | 1,429 (0.5) | 65,782 (2.6) | 0.19 | 0.03–1.48 | 0.114 | |||
Education level | ||||||||
Below elementary | 42,189 (15.4) | 246,118 (9.8) | 2.28 | 1.10–4.73 | 0.027 | 2.28 | 1.00–4.95 | 0.049 |
Middle school | 15,181 (5.5) | 213,117 (8.5) | 0.95 | 0.38–2.35 | 0.907 | 0.80 | 0.32–2.00 | 0.629 |
High school | 86,897 (31.7) | 1,155,759 (46.2) | 1 | 1 | ||||
Above university | 129,712 (47.3) | 888,636 (35.5) | 1.94 | 1.04–3.64 | 0.039 | 2.17 | 1.08–4.34 | 0.029 |
Type of residence | ||||||||
General type | 118,737 (43.3) | 1,428,219 (57.0) | 1 | 1 | ||||
Apartment | 155,243 (56.7) | 1,075,412 (43.0) | 1.74 | 1.01–2.98 | 0.045 | 1.77 | 1.02–3.08 | 0.043 |
Household income | ||||||||
Lowest quartile (Q1) | 39,801 (14.5) | 401,236 (16.0) | 1.16 | 0.51–2.61 | 0.728 | 1.68 | 0.70–4.01 | 0.243 |
Low–middle quartile (Q2) | 101,409 (37) | 509,031 (20.3) | 2.32 | 1.17–4.60 | 0.016 | 3.03 | 1.40–6.58 | 0.005 |
High–middle quartile (Q3) | 68,135 (24.9) | 840,694 (33.6) | 0.94 | 0.46–1.94 | 0.874 | 0.92 | 0.41–2.07 | 0.838 |
Highest quartile (Q4) | 64,635 (23.6) | 752,670 (30.1) | 1 | 1 |
Values are presented as number (%) or mean±SD unless otherwise indicated.
OR: odds ratio; CI: confidence interval; BMI: body mass index; EC: electronic cigarette; SD: standard deviation.
Before increase policy | After increase policy | Interaction | ||||
---|---|---|---|---|---|---|
OR | p-value | OR | p-value | OR* | p-value† | |
Male sex | 0.90 | 0.675 | 0.89 | 0.718 | 0.987 | 0.975 |
Age, yr | ||||||
19–29 | 3.89 | <0.001 | 1.82 | 0.230 | 0.48 | 0.197 |
30–39 | 1.43 | 0.287 | 2.51 | 0.069 | 1.76 | 0.351 |
40–49 | 1 | 1 | 1 | |||
50–59 | 1.01 | 0.973 | 1.71 | 0.282 | 1.69 | 0.397 |
≥60 | 1.65 | 0.102 | 2.67 | 0.026 | 1.62 | 0.370 |
BMI | 0.94 | 0.022 | 1.02 | 0.629 | 1.09 | 0.073 |
Diabetes | 0.69 | 0.301 | 1.39 | 0.393 | 2.01 | 0.181 |
Hypertension | 0.56 | 0.023 | 0.56 | 0.037 | 1 | 0.994 |
Married | 0.31 | <0.001 | 1.82 | 0.081 | 5.79 | <0.001 |
Education level | ||||||
Below elementary | 0.85 | 0.642 | 2.28 | 0.027 | 2.67 | 0.051 |
Middle school | 1.36 | 0.393 | 0.95 | 0.907 | 0.7 | 0.538 |
High school | 1 | 1 | 1 | |||
Above university | 1.6 | 0.072 | 1.94 | 0.039 | 1.21 | 0.642 |
Type of residence | ||||||
General type | 1 | 1 | ||||
Apartment | 1.17 | 0.469 | 1.74 | 0.045 | 1.49 | 0.252 |
Household income | ||||||
Lowest quartile (Q1) | 1.14 | 0.695 | 1.16 | 0.728 | 1.02 | 0.975 |
Low–middle quartile (Q2) | 0.9 | 0.697 | 2.32 | 0.016 | 2.59 | 0.034 |
High–middle quartile (Q3) | 0.77 | 0.301 | 0.94 | 0.874 | 1.23 | 0.638 |
Highest quartile (Q4) | 1 | 1 | 1 |
*Ratio of OR after policy changes: if OR >1, it has a positive effect that relatively increases the smoking cessation rate; and if OR <1, it has a negative effect that has a relatively lower smoking cessation rate. †Interaction p-value <0.2 signifies changes in recent smoking cessation rates compared to before policy change.
OR: odds ratio; BMI: body mass index.