Ethnic disparities in prevalence and patterns of smoking and nicotine dependence in rural southwest China: a cross-sectional study.

OBJECTIVES: This study examines ethnic disparities in prevalence and patterns of smoking and nicotine dependence in rural southwest China.
DESIGN: This was a cross-sectional design.
SETTING: This study was conducted in rural Yunnan Province of China. PARTICIPANTS: 7027 consenting individuals aged ≥35 years among Han majority and four ethnic minority groups (Na Xi, Li Shu, Dai and Jing Po) participated in this study. Information about participants' demographic characteristics as well as smoking habits and an assessment of nicotine dependence with the Fagerstrom Test for Nicotine Dependence (FTND) was obtained using a standard questionnaire.
RESULTS: Males had significantly higher prevalence of current smoking than females (64.8% and 44.4%, p<0.01). Among current smokers, the prevalence of nicotine dependence was significantly higher in males compared with females (19.9% and 7.1%, p<0.01). Jing Po men and women had the highest prevalence of current smokers (72.2% vs 23.1%, p<0.01), whereas the highest prevalence of nicotine dependence was found in male Dai current smokers and female Li Shu current smokers (44.8% vs 32.5%, p<0.01). Filtered cigarettes were the most popular form of tobacco used across all five ethnic groups. Over 75% of tobacco users initiated smoking and regularly smoked during adolescence, and those of minority ethnicity smoked regularly at a younger age than those of Han descent (p<0.05). Individuals in all five ethnic groups with higher levels of education had a lower probability of current smoking status (p<0.05), whereas a negative association of level of education with nicotine dependence was only observed in current smokers in the Han majority and Dai ethnic minority groups. Among Han majority current smokers, higher annual household income was associated with a higher risk of nicotine dependence (p<0.05).
CONCLUSION: Future interventions to control tobacco use should be tailored to address ethnicity and socioeconomic factors. © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

This present research is the first study to examine ethnic disparities in prevalence and patterns of smoking in the southwest Chinese population.

Our large sample size enhances the validity of our results.

High response rate (93.7%) in this study.

Prevalence of smoking was based on self-reporting, and may therefore be subject to recall bias.

Introduction

Smoking remains a major public health problem worldwide. The WHO reported there are >1.1 billion smokers globally, and projects >1 billion tobacco-related deaths this century.1 Non-communicable diseases (NCDs) cause ~40 million deaths each year, and many NCDs have tobacco-related causes and/or contributing factors.2 3 Given these grave health hazards, WHO ranks tobacco smoking as a top 10 threat to human health for the 21st century.

Smoking behaviours were gendered from the outset: nearly all smokers were male at the beginning of 20th century. However, alongside economic and social development, the prevalence of female smokers increased in many countries.4 5 Moreover, the age at which people begin smoking has dropped lower over the last century in Asian countries like China.6 According to a previous study, smoking can lead to a 15.2% loss of health life years for females and 25.5% loss for males.7

As the world’s largest tobacco producer and consumer, China accounts for about 28% of smokers globally.8 According to China Statistic Yearly 2014, China produces 2560.4 billion cigarettes per year, with smokers numbering >300 million. In China, the death rate attributable to smoking was 16.5% in men and 1.7% in women in 2012.9

China is a multiethnic nation, home to 56 ethnic groups, each with unique customs. The largest ethnic group is Han, accounting for 92% of the total population. Many studies have explored racial and ethnic disparities in prevalence of smoking worldwide.10–12 In China, several studies have examined ethnic differences in prevalence of smoking and nicotine dependence,13 14 but the literature is sparse, especially with regard to ethnic minority communities in rural regions. Improved understanding of ethnic disparities in smoking and nicotine dependence is critical to effectively reducing and controlling its prevalence in China.

Yunnan Province stretches along China’s southwestern border and is a major tobacco plantation region. Indeed, tobacco production is one of the largest industries in Yunnan, with 105 counties currently cultivating tobacco. Consequently, residents have higher tobacco exposure than the national level.15 Yunnan also has China’s largest concentration of ethnic minorities, with 25 ethnic minority groups residing in the province, 15 of which only live in Yunnan. Traditions, heritage, culture, and lifestyles differ among these ethnic groups. Most of the ethnic minority population in Yunnan lives in rural or mountainous regions, lacks awareness of the harms of tobacco relative to the rest of the population, and as a result may have a higher smoking rate. However, little information is available on the prevalence and patterns of smoking and nicotine dependence among minorities in China, and the causes of those variations by ethnicity that are currently known are not well understood. Therefore, this study aimed to investigate ethnic disparities in prevalence and patterns of smoking and nicotine dependence among the Han majority and four unique ethnic minority groups (Na Xi, Li Shu, Dai and Jing Po) in rural southwest China.

Methods

Study area and population

The study population consisted of residents of Yunnan Province, China who responded to the Yunnan Provincial Community Health Survey (YCHS) from 2015 to 2017, a survey supported by the National Natural Science Foundation of China. This survey was conducted in one Han majority county and four distinct ethnic minority-populated regions of Yunnan Province using a stratified multistage sampling technique. In stage 1, all 129 counties were divided into four strata (valley area, dam area, semi-mountain area and alpine area) based on the geographical characteristics of the regions. One Han majority county was randomly chosen from valley area, and one unique ethnic minority autonomous county was randomly chosen from each stratum, for a total of five counties. In stage 2, each selected county was further divided into three categories—economically advantaged, economically average and economically disadvantaged—based on wealth distribution (per capita gross domestic product). One township was then randomly selected from each of these three chosen groups, for a total of 15 townships. In stage 3, three villages were chosen by probability proportional to size from each of the 15 townships. In the final stage, simple random sampling was conducted to select individuals aged ≥35 years from each chosen village. The YCHS initially selected 7500 individuals aged ≥35 years to participate, of which 7027 consented, representing an overall response rate of 93.7%.

Data collection and measurement

Twenty medical students from Kunming Medical University were selected and trained as interviewers for data collection. Each participant who gave informed consent was interviewed in person by one of the interviewers using a pretested and structured questionnaire. Information on demographic characteristics, individual socioeconomic status including ethnicity, household income and level of education, and self-reported smoking habits was obtained through the questionnaire.

Definitions

Participants who reported that they had smoked >100 cigarettes in their life thus far as well as currently smoked any kind of tobacco product on a daily basis were categorised as current smokers. Self-reported current smokers were assessed for nicotine dependence using six questions from the Fagerstrom Test for Nicotine Dependence (FTND).16 FTND has 5 grades and includes 6 questions, resulting in a total score from 0 to 10. Dependence was defined as a score ≥5 in this study. Specifically, an FTND score from 0 to 2 was defined as very low dependence; a score from 3 to 4 was defined as low dependence; a score equal to 5 was defined as medium dependence; a score from 6 to 7 was defined as high dependence; and a score from 8 to 10 was defined as very high dependence. Annual household income was divided into three categories: low (US$800). Level of education was also classified into three categories: illiterate, primary (grade 1–6) and middle (grade 7–9) or higher.

Statistical analysis

Data were double entered into an EpiData Entry V.3.1 electronic database and analysed using R V.3.2.0 software. Descriptive analysis techniques and χ2 tests were used in this study. Counts and percentages were used to present categorical variables.

The age-standardised prevalence of current smokers and nicotine dependence were computed by directly standardising to the overall sample. All statistical significance decisions were based on a two-tailed p value of <0.05.

Results

Table 1 shows the demographic characteristics of the participants. Participants in the study included 3434 (48.9%) males and 3593 (51.1%) females. Among the study participants, the percentage of male Han ethnic majority, Na Xi, Li Shu, Dai and Jing Po ethnic minorities was 50.6%, 45.9%, 49.4%, 49.1% and 49.2%, respectively. Han majority participants had the highest level of education and annual household income of the ethnic minorities studied, whereas those of the Li Shu ethnic minority had the lowest level of education and those of the Dai ethnic minority had the lowest annual household income level of the five ethnicities studied (p<0.01).

Table 1

Demographic characteristics of the study population

Characteristics	Han ethnic majority	Na Xi ethnic minority	Li Shu ethnic minority	Dai ethnic minority	Jing Po ethnic minority	All
	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)
Sex
Male	757 (50.6)	644 (45.9)	675 (49.4)	686 (49.1)	672 (49.2)	3434 (48.9)
Female	738 (49.4)	758 (54.1)	691 (50.6)	711 (50.9)	695 (50.8)	3593 (51.1)
Age (years)
35–44	311 (20.8)*	405 (28.9)	545 (39.9)	336 (24.1)	467 (34.2)	2064 (29.4)
45–59	604 (40.4)	535 (38.2)	538 (39.4)	575 (41.2)	586 (42.9)	2828 (40.4)
≥60	580 (38.8)	462 (33.0)	283 (20.7)	486 (34.8)	314 (23.0)	2125 (30.2)
Level of education
Illiterate	171 (11.4)**	194 (13.8)	559 (40.9)	549 (39.3)	406 (29.7)	1879 (26.7)
Primary (grade 1–6)	730 (48.8)	648 (46.2)	654 (47.9)	723 (51.8)	693 (50.7)	3448 (49.1)
Middle (grade 7–9) or higher	594 (39.7)	560 (39.9)	153 (11.2)	125 (8.9)	268 (19.6)	1700 (24.2)
Level of annual household income (US$)
Low	141 (9.4)**	147 (10.5)	693 (50.7)	637 (45.6)	608 (44.5)	2226 (31.7)
Medium	467 (31.2)	695 (49.6)	524 (38.4)	646 (46.2)	645 (47.2)	2977 (42.4)
High	887 (59.3)	560 (39.9)	149 (10.9)	114 (8.2)	114 (8.3)	1824 (26.0)

*P<0.05, **p<0.01.

As shown in tables 2 and 3, the overall prevalence rate of current smoking was 35.1%, and the rate of nicotine dependence among current smokers was 40.6%. Males had a significantly higher prevalence of current smoking than females (64.8% and 44.4%, p<0.01) in all five studied ethnicities. Moreover, with the exception of Li Shu and Jing Po ethnic minority groups (where smoking prevalence among females is substantial, with 11.3% and 23.1%, respectively), prevalence of current smoking was consistently high among males, and consistently nearly zero among females. The prevalence of nicotine dependence was much higher in male current smokers than female current smokers (7.1% and 19.9%, p<0.01). Among the five populations studied, males of the Jing Po ethnicity had the highest prevalence of current smoking, whereas male current smokers of Na Xi descent had the highest prevalence of nicotine dependence (p<0.01). Females in the Jing Po ethnic minority group also had the highest prevalence rate of current smoking (p<0.01). Participants aged 35–44 years in the Han, Na Xi and Dai ethnicities had the highest prevalence of current smokers (p<0.05). Education was inversely correlated with probability of smoking in all five ethnicities studied: those with higher levels of education had a lower likelihood of currently smoking (p<0.05). Among Han and Dai current smokers, participants with higher levels of education also had a lower probability of nicotine dependence. Among Han current smokers, higher annual household income was associated with a higher risk of nicotine dependence (p<0.05).

Table 4 indicates patterns of smoking among current smokers in rural Yunnan Province. Filtered cigarettes were the most popular form of tobacco used among all five ethnicities studied (p<0.01), particularly in the Na Xi ethnic minority (98.4%). The second most used form of tobacco varied by ethnicity: hookah/water pipe in Han (18.2%) and Na Xi populations (0.7%), pipe tobacco among Dai (12.7%) and Li Shu (5.3%) ethnic minorities, and chewing tobacco in the Jing Po population (27.9%). The Jing Po ethnic minority had the lowest age of smoking initiation, but the age of smoking initiation and of regular smoking for all five ethnicities studied was 12–20 years (p<0.01). Na Xi, Li Shu and Dai ethnic minorities had a younger average age of regular smoking than the Han population (p<0.05). More than 85% of smokers in the study smoked in public places (defined as having smoked in a public place in the prior 7 days); the Han population had the highest rate of smoking in public places (p<0.05).

Table 4

Patterns of smoking among current smokers in rural regions of Yunnan Province, China (%)

Variables	Han ethnic majority	Na Xi ethnic minority	Li Shu ethnic minority	Dai ethnic minority	Jing Po ethnic minority	All
Use of various forms of tobacco
Filtered cigarettes	69.2**	98.4	86.7	93.3	60.8	80.0
Unfiltered cigarettes	17.6	0.0	1.0	3.2	3.9	5.0
Hookah / water pipe	18.2	0.7	2.1	0.2	0.8	4.1
Pipe tobacco	2.4	0.5	5.3	0.4	0.3	1.7
Hand-rolled cigarettes	2.2	0.4	5.0	12.7	13.3	7.3
Chewing tobacco	0.0	0.0	0.0	0.2	27.9	7.3
Age of smoking initiation (years)
≤11	1.3	2.3	0.6	6.9	7.7**	3.9
12–20	79.5	80.8	84.6	75.8	60.1	75.2
21–34	17.3	15.0	13.0	14.0	25.6	17.6
≥35	1.9	1.9	1.8	3.3	6.6	3.3
Age of regular smoking (years)
≤11	0.3*	1.0	0.4	4.3	5.0	2.3
12–20	66.8	75.2	79.0	68.7	56.9	68.4
21–34	31.0	21.5	18.0	23.8	29.4	25.3
≥35	1.9	2.3	2.6	3.3	8.7	4.0
Location of smoking in the previous 7 days
Home	13.2*	11.3	9.2	11.2	10.9	10.9
Public spaces (schools, hospitals and so on)	86.8	88.7	91.8	88.8	89.1	89.1

*P<0.05, **p<0.01, the χ2 tests for independence were conducted for each pattern of smoking across the ethnic groups.

Discussion

The findings indicate a high prevalence rate of smoking and nicotine dependence in rural southwest China, particularly among male adults of ethnic minority descent. Further, the data reveal significant ethnic differences in prevalence and patterns of smoking: both current smoking and nicotine dependence vary by ethnic minority group.

The present study also found that filtered cigarettes were the most popular tobacco form in all five studied ethnicities in rural southwest China. This aligns with previous Chinese studies.15 17 Furthermore, the finding that the second most common tobacco form among the Han population was hookah/water pipes is also consistent with previous research.15 Chewing tobacco was particularly popular in the Jing Po ethnic minority. This is possibly due to the fact that the majority of the Jing Po minority lives in semi-mountainous and mountainous areas where tobacco with a stronger taste and higher tar content is preferred. These ethnic differences indicate culture plays a role in shaping smoking behaviours.

The present study also revealed that most current smokers in the five studied ethnic groups initiate smoking as well as begin regular smoking during adolescence, a finding which accords with previous studies,18 19 indicating smoking is a serious problem for teenagers in rural southwest China. In addition, all four ethnic minorities studied had a lower age of smoking initiation and of regular smoking than the Han population. In particular, the Jing Po ethnic minority had the highest proportion of initiation of smoking and regular smoking behaviours at <11 years of age. These findings thus underscore an urgent need for governmental interventions to improve health education and to increase efforts to prevent adolescents from smoking initiation, especially in ethnic minority areas.

In the present study, over 86% of current smokers in the five studied ethnic groups smoked in public places over the 7 days prior to the collection of data, a significantly higher rate than observed in a prior Chinese study (63.3%) as well as studies from other Asian nations, including India (29.9%), the Philippines (32.6%) and Thailand (27.2%),20 indicating public places have become more crowded with smoking in rural southwest China. The findings suggest it is essential to raise awareness of the harms of smoking and to implement comprehensive smoke-free policies to reduce levels of exposure to secondhand smoke in public places.

In the present study, males had a markedly higher prevalence rate of current smoking and nicotine dependence than females among all five ethnicities studied, while prevalence of current smoking in males in all four studied ethnic minorities was higher than in the Han population. That males smoked significantly more and were more likely to have a nicotine addiction than females accords with other studies.21 22 However, the overall prevalence rate of current smoking and nicotine dependence among males was higher in the present study than in both previous Chinese studies as well as studies in other countries.18 23 The findings thus provide further evidence that smoking and nicotine dependence remains a widespread challenge for rural Chinese male adults, particularly in ethnic minority populations.

Our finding that individuals with a lower educational level had a higher prevalence of current smoking and nicotine dependence in all five studied ethnicities accords with other studies of the relationship between education and tobacco use.21 24 Namely, while those with high levels of education tend to have healthier lifestyles, those with low levels of education are more likely to believe false information about tobacco, lack awareness about tobacco hazards, and pay insufficient attention to the harm of tobacco and their own health conditions, leading to a comparatively high level of smoking and nicotine dependence.24 However, our study indicated that the differences in prevalence of current smoking by education are modest (only 5%–10% differences). This result differs from previous studies where large differences in prevalence of current smoking by education were observed (over 20% differences).25 26 The reasons behind this inconsistent effect of education on current smoking levels require further research.

In the present study, whereas individual annual income had no association with current smoking among the five studied ethnic groups, it was positively correlated with nicotine dependence in the Han majority population. This result is consistent with previous Chinese studies,15 but differs from some western studies’ findings that those with lower incomes had a higher probability of nicotine dependence.27 The reason for this dichotomy is unclear, and further investigation is needed to examine the exact nature of the association between income and nicotine dependence.

The findings of the present study may be limited. Prevalence of smoking was based on self-reporting and was not validated with biochemical tests. It may therefore be subject to recall bias, leading to an underestimated prevalence of smoking in the study population.

Conclusion

This study reveals ethnic differences in prevalence and patterns of smoking and nicotine dependence in rural southwest China. The results of our study underscore the necessity of considering ethnicity as a factor in future tobacco control programme, and highlight the need for future tobacco cessation interventions to target men and those with lower education levels.

Table 2

Age-standardised prevalence of current smoking by socioeconomic status (%)

Characteristics	Han ethnic majority	Na Xi ethnic minority	Li Shu ethnic minority	Dai ethnic minority	Jing Po ethnic minority	All
Sex
Male	59.2**	67.1**	59.6**	66.9**	72.2**	64.8**
Female	1.1	0.8	11.3	0.7	23.1**	7.1
Age (years)
35–44	36.5*	34.9*	32.1*	38.5*	41.3*	36.2*
45–59	34.3	31.5	35.1	35.1	49.5	35.3
≥60	23.1	27.2	39.2	26.9	52.2	31.0
Level of education
Illiterate	38.7**	35.8*	36.1*	35.2*	49.6*	39.5*
Primary (grade 1–6)	31.0	32.5	34.2	31.3	45.9	35.3
Middle (grade 7–9) or higher	26.8	28.7	31.4	31.7	40.2	30.6
Level of annual household income
Low	24.2	27.2	35.6	37.4	46.4	33.8
Medium	28.7	33.5	33.8	36.7	48.7	36.2
High	28.1	29.1	34.2	33.8	43.2	34.4
All	30.2	31.3	35.1	33.2	47.4*	35.1

*P<0.05, **p<0.01, the χ2 tests for independence were conducted for prevalence of current smoking across socioeconomic status in all five studied ethnicities.

Table 3

Age-standardised prevalence of nicotine dependence among current smokers by socioeconomic status (%)

Characteristics	Han ethnic majority	Na Xi ethnic minority	Li Shu ethnic minority	Dai ethnic minority	Jing Po ethnic minority	All
Sex
Male	42.7**	44.8**	38.9*	42.6**	44.4**	42.8**
Female	0.0	0.0	32.5**	0.0	19.9	22.1
Age (years)
35–44	47.4	44.7	33.7	35.4	40.8	39.9
45–59	37.3	42.5	42.3	45.8	38.5	41.0
≥60	44.7	45.3	36.9	43.1	35.0	40.8
Level of education
Illiterate	48.0*	45.0*	39.7*	50.4*	36.5*	41.9*
Primary (grade 1–6)	42.4	41.5	33.6	42.5	32.8	38.6
Middle (grade 7–9) or higher	40.6	38.9	28.9	35.3	27.3	37.4
Level of annual household income
Low	34.2	43.2	35.1	37.7	38.3	40.0
Medium	35.6	42.8	42.2	36.2	36.0	39.2
High	44.5*	42.7	37.6	37.4	37.6	42.8
All	42.0	44.0	37.9	42.1	38.	40.6

*P<0.05, **p<0.01, the χ2 tests for independence were conducted for prevalence of nicotine dependence across socioeconomic status in all five studied ethnicities.