Prevalence, pattern and sociodemographic differentials in smokeless tobacco consumption in Bangladesh: evidence from a population-based cross-sectional study in Chakaria.

BACKGROUND: The health hazards associated with the use of smokeless tobacco (SLT) are similar to those of smoking. However, unlike smoking, limited initiatives have been taken to control the use of SLT, despite its widespread use in South and Southeast Asian countries including Bangladesh. It is therefore important to examine the prevalence of SLT use and its social determinants for designing appropriate strategies and programmes to control its use.
OBJECTIVE: To investigate the use of SLT in terms of prevalence, pattern and sociodemographic differentials in a rural area of Bangladesh.
DESIGN: Population-based cross-sectional household survey. SETTING AND PARTICIPANTS: A total of 6178 individuals aged ≥13 years from 1753 households under the Chakaria HDSS area were interviewed during October-November 2011.
METHODS: The current use of SLT, namely sadapatha (dried tobacco leaves) and zarda (industrially processed leaves), was used as the outcome variable. The crude and net associations between the sociodemographic characteristics of respondents and the outcome variables were examined using cross-tabular and multivariable logistic regression analysis, respectively.
RESULTS: 23% of the total respondents (men: 27.0%, women: 19.3%) used any form of SLT. Of the respondents, 10.4% used only sadapatha,13.6% used only zarda and 2.2% used both. SLT use was significantly higher among men, older people, illiterate, ever married, day labourers and relatively poorer respondents. The odds of being a sadapatha user were 3.5-fold greater for women than for men and the odds of being a zarda user were 3.6-fold greater for men than for women.
CONCLUSIONS: The prevalence of SLT use was high in the study area and was higher among socioeconomically disadvantaged groups. The limitation of the existing regulatory measures for controlling the use of non-industrial SLT products should be understood and discussion for developing new strategies should be a priority. Published by the BMJ Publishing Group Limited. For permission to use (where not already granted under a licence) please go to http://www.bmj.com/company/products-services/rights-and-licensing/.

Data were of very high quality as they were collected as part of a regular health and demographic surveillance system by a team of highly trained interviewers with very strong supervision for quality control.

Multivariable logistic regression analysis using Generalised Estimating Equation (GEE) resulted in accurate estimates of the effects of the independent variables on the use of SLT.

Use of SLT was assessed based on self-reported data and may suffer from under-reporting.

No casual inference could be made as the study was based on a cross-sectional survey.

The study findings cannot be generalised for Bangladesh.

Introduction

The tobacco epidemic is one of the biggest public health threats the world has ever confronted, which is believed to account for about 6 million deaths a year.1 Although tobacco use is a problem in both the developed and developing world, the problem is more acute in developing countries. The use of tobacco also varies between low-income and high-income countries with regard to the different forms of tobacco consumed.2–4 Strict regulatory measures help high-income countries to control the level of tobacco use, whereas in low-income countries there is an absence of regulation and/or weak implementation of existing regulations.5 In addition, the availability of crude tobacco products in processed and/or unprocessed forms further aggravates the situation in a country such as Bangladesh, which perhaps exposes the disadvantaged section of the population more to the health hazard of smokeless tobacco products (SLT).6 In Bangladesh the two major forms of SLT products are zarda (industrially processed tinned tobacco leaves) and sadapatha (dried whole tobacco leaves, mostly done at home), both of which are usually consumed with betel leaves.7 8 As these products (especially sadapatha) are not industrially manufactured or processed products, they are outside the regulatory framework and do not have any packaging on which to show any regulatory warning about their nicotine content and adverse health effects.8 The consumption of such SLT products may not be perceived to be as harmful as smoking, which increases the consumption of SLT among the general population, especially the illiterate and economically disadvantaged.

As reported in many studies, SLT products have a high level of toxicity, mutagenicity and cancer-causing elements and roughly 4200 low-grade chemicals9 10 which would disproportionately affect health. SLT products are also associated with several adverse health outcomes including cancer,11 12 cardiovascular disease,13–15 oral pain,16 chronic disease (hypertension, diabetes, tuberculosis, chronic obstructive pulmonary disease (COPD))17–19 and life-threating complications in reproductive stages of women, with adverse pregnancy outcomes (eg, stillbirth, low birth weight).20 21 Despite the adverse health hazards of SLT, unlike smoking tobacco, very few initiatives have been undertaken by government and non-governmental agencies to control its use.

The prevalence of SLT use in most of the South Asian countries is much higher than in sub-Saharan Africa, Central and Western Asia and other developed countries.22–24 In South and Southeast Asia SLT is used in diverse forms, particularly in Bangladesh where an increasing array of SLT products such as sadapatha is widely available.4 25 26 Moreover, Bangladesh is one of the main cultivators of tobacco in the world, and tobacco leaves are readily available for use as SLT.4 The most common forms (zarda and sadapatha) are easily accessible to the general population. However, the prevalence and pattern of SLT consumption among the general population is unknown. Although several studies have been conducted to assess the prevalence, pattern and determinants of smoking tobacco in Bangladesh, similar studies on SLT—particularly on zarda and sadapatha—are very limited.6 It is against this background that this study investigated the prevalence, pattern and sociodemographic differentials in SLT consumption, mainly zarda and sadapatha, using data from a cross-sectional study in Chakaria, a rural area in Bangladesh. The findings of the study will help in designing appropriate policies and programmatic interventions to protect people from the harmful effects of SLT consumption.

Materials and methods

Study area

The study was conducted in Chakaria upazila, a subdistrict of Cox's Bazar District, which is located in the southeast of Bangladesh. The population density of Chakaria is 782 per square kilometre, which is a little lower than the national density of 964 per square kilometre.27 28 Over 3393 acres of agricultural land have been used for growing tobacco, which could have been used to produce other crops.29 The study area has been one of the most conservative areas in terms of religious belief and openness to modern ideas with a very low level of secular education. Ninety-one per cent of the population in Chakaria are Muslims while the remainder are either Hindu or Buddhist.30 The literacy rate among the population aged ≥7 years is 47.6%, with a similar level for men and women.31 In Chakaria, pulmonary tuberculosis, acute respiratory infection, stroke, diabetes mellitus and COPD have been the leading causes of death in all age groups.32

Study design and data collection

Data for this study came from a baseline survey carried out as part of a study on reducing tobacco use through an awareness building-based intervention. The study was conducted in 30 randomly selected villages out of 183 during October and November 2011. Twenty-two female interviewers who had at least 10 years of schooling and who came from the locality formed the data collection team with two experienced supervisors. To ensure the quality of the data, a supervisor revisited 5% of the respondents, chosen randomly, within 2 days of data collection by the field workers. Later, the supervisors and data collectors together sorted out any inconsistencies found in the collected data. All the completed questionnaires were manually checked for completeness and for any inconsistencies before the data were entered in the computer database. Finally, 6178 randomly selected individuals aged ≥13 years from 1753 households were successfully interviewed.

Definition of variables

The dependent variables used in this study are the use of SLT (yes/no) in the form of sadapatha (dried tobacco leaf available as loose leaves) or zarda (processed tobacco leaves as flakes produced industrially and available in a tin). The respondents were first asked whether they currently use either of the SLTs and, if yes, at what age they started. Sadapatha is commonly sundried without any additive ingredients and can be produced at the household level while zarda is made from tobacco using lime spices, vegetable dyes and areca nut. Sadapatha is more raw in nature than zarda and is cheaper.

Sociodemographic characteristics of the respondents other than age were collected at the time of the interview. Age was obtained from the existing database maintained by the health and demographic surveillance system. The independent variables included age, education, marital status, occupation and household socioeconomic status which were used as categorical variables in logistic regression analysis. Household socioeconomic status was determined by the wealth index computed from ownership of a standard set of household items observed by the interviewers using principal components analysis.33 34 In principal component analysis the first component (a linear combination of the assets weighted by the coefficent or factor loading) that possesses maximum variability in the data was used to calculate the score or wealth index. The wealth index was then used to categorise the individuls into five equal groups (or quintiles) where the first quintile is the poorest 20% of households and the fifth quintile is the wealthiest 20% of households.35 36

Data analysis

Both bivariate and multivariate statistical analyses were performed to explore the sex and socioeconomic differentials in the consumption of SLT. The 95% CI for the true prevalence of SLT use was calculated adjusting for clustering at the household level. Association between the dependent and independent variables was assessed by χ2 test. Considering the prevailing variation of tobacco use between men and women in Bangladesh, further analysis was conducted to investigate the sex differential in the consumption of the two forms of SLT. To examine the net sex differentials adjusting for the effect of the independent variables, multivariable logistic regression analysis was carried out. As all members aged ≥13 years in a household were interviewed, multiple respondents were included in the study. This situation might have resulted in correlated observations within a household requiring a special statistical technique to avoid this problem. Two of the commonly used techniques in such a situation are the marginal model with generalised estimating equation (GEE) and a cluster-specific model.37 38 We used the GEE model for these data and estimated the model using STATA software for Windows (V.13.1).

Results

Initiation of SLT use

Both men and women started to use SLT in their 30s, with an average age of 35.0 years (mean±SD 35.1±13.6 years). Although the use of SLT started somewhat earlier in men than in women (33.7±14.8 years vs 37.5±10.9 years), in general the use of zarda started earlier in men than in women (28.3±13.1 years vs 35.0±12.8 years) while the use of sadapatha started earlier in women than in men.

Prevalence and pattern of SLT use

The results presented in table 1 show that 22.9% of the respondents in the study area reported using any form of SLT. Of these, 10.4% used only sadapatha, 13.6% used only zarda and 2.2% used both. The use of any form of SLT was significantly higher among men, people aged >45 years, illiterate, ever married, day-labourers and people from poorer families. When the two forms of SLT (ie, sadapatha and zarda) were separately analysed it was found that the use of zarda was more prevalent in men than in women (22.4% vs 6.0%) while the use of sadapatha was higher in women than in men (14.4% vs 5.9%). The sex differentials in the use of zarda and sadapatha were statistically significant (p<0.001).

Table 1

Prevalence of smokeless tobacco use by sociodemographic characteristics in Chakaria, 2011

		Sadapatha		Zarda		Smokeless tobacco
Characteristics	n	% (95% CI)	p Value	% (95% CI)	p Value	% (95% CI)	p Value
Sex			<0.001				<0.001
Men	2877	5.9 (5.1 to 6.8)		22.4 (20.8 to 24.0)	<0.001	27.0 (25.4 to 28.7)
Women	3291	14.4 (13.3 to 15.6)		6.0 (5.2 to 6.8)		19.3 (18.1 to 20.6)
Age			<0.001				<0.001
13–24	2339	0.4 (0.2 to 0.8)		8.1 (7.0 to 9.3)	<0.001	9.2 (8.1 to 10.5)
25–44	2330	8.8 (7.6 to 10.1)		17.6 (16.1 to 19.2)		25.6 (23.8 to 27.4)
45–64	1122	28.3 (25.7 to 31.2)		18.2 (16.0 to 20.6)		41.4 (38.5 to 44.5)
65+	377	29.4 (25.1 to 34.2)		9.6 (7.0 to 13.0)		36.1 (31.5 to 40.9)
Education			<0.001				<0.001
None	1812	23.6 (21.6 to 25.6)		15.6 (13.9 to 17.4)	0.003	36.2 (33.9 to 38.5)
1–5 year(s)	2219	6.4 (54.4 to 7.6)		15.0 (13.6 to 15.6)		21.4 (19.7 to 23.2)
6+ years	2136	3.4 (2.7 to 4.4)		10.4 (9.2 to 11.9)		13.2 (11.8 to 14.7)
Marital status			<0.001				<0.001
Unmarried	1992	0.8 (0.5 to 0.12)		9.3 (8.1 to 10.7)	<0.001	10.5 (9.2 to 11.9)
Married	3831	13.4 (12.2 to 14.6)		15.9 (14.8 to 17.2)		27.6 (26.2 to 29.1)
Divorced	24	8.3 (2.1 to 27.9)		20.8 (8.9 to 41.4)		33.3 (17.6 to 53.9)
Widowed	321	35.8 (30.7 to 41.2)		12.2 (9.0 to 16.2)		42.7 (37.4 to 48.1)
Occupation			<0.001				<0.001
Day-labour	910	9.1 (7.4 to 11.3)		30.3 (27.4 to 33.4)		39.0 (35.8 to 42.3)
Farmer	573	9.6 (7.5 to 12.3)		22.7 (19.4 to 26.3)		29.1 (25.6 to 33.0)
Job/business	730	5.8 (4.3 to 7.7)		24.8 (21.7 to 28.2)		29.9 (26.6 to 33.4)
Housewife	2028	15.6 (14.1 to 17.2)		6.8 (5.8 to 8.0)		21.3 (19.6 to 23.1)
Unemployed	1740	8.3 (7.0 to 9.7)		4.9 (3.9 to 6.0)		12.1 (10.6 to 13.8)
Others	187	2.1 (0.8 to 5.0)		15.5 (11.0 to 21.4)		16.0 (11.4 to 22.1)
Asset quintile			0.444		0.028		<0.001
Lowest	1240	11.1 (9.4 to 13.1)		15.5 (13.5 to 17.7)		26.4 (23.9 to 28.9)
Second	1214	11.1 (9.4 to 13.1)		13.2 (11.4 to 15.3)		23.1 (20.8 to 25.5)
Middle	1249	10.4 (8.8 to 12.2)		14.5 (12.4 to 16.8)		23.8 (21.2 to 26.5)
Fourth	1203	10.6 (8.9 to 12.5)		13.8 (11.9 to 16.0)		22.4 (20.0 to 25.0)
Highest	1226	9.0 (7.4 to 10.8)		10.9 (9.2 to 12.9)		18.8 (16.7 to 21.0)
All	6168	10.4 (9.6 to 11.3)		13.6 (12.7 to 14.5)		22.9 (21.8 to 24.0)

In addition to the results described above, the strong sex differential can also be explored using the disaggregated level of zarda and sadapatha use among the current SLT users. For example, among current SLT male users, 77% used zarda, 12% used sadapatha and 11% used both, while in female users, 24% used zarda, 68% used sadapatha and 8% used both (figure 1).

Figure 1

Percentage of different forms of smokeless tobacco use among current users by sex.

The results from multivariable logistic regression analysis revealed that the odds of being a SLT user among illiterate people, older people, ever married and day-labourers was significantly higher than for those who were literate, younger, unmarried and with an occupation other than a day-labourer (table 2). No significant sex differential was observed in the use of SLT. However, a significant sex differential was observed (p<0.001) when the separate model was fitted for sadapatha and zarda. The results indicated that the odds of being a sadapatha user were 3.5-fold greater for women than for men. The opposite scenario was observed in the use of zarda with an odds ratio of 3.6 for men to women. To investigate whether the observed sex difference was modified by other background factors such as age, education, marital status, occupation and socioeconomic status, a model with an interaction term consisting of sex and each of the background factors along with their main effects was fitted. The insignificance of the interaction terms suggests that sex differentials in the use of sadapatha and zarda were not modified by the other background factors (results not shown).

Table 2

Results from multivariable logistic regression analysis for smokeless tobacco consumption

	Sadapatha		Zarda		Smokeless tobacco
	OR (95% CI)	p Value	OR (95% CI)	p Value	OR (95% CI)	p Value
Sex		0.000		0.000		0.231
Men	1		3.6 (2.5 to 5.0)		1.2 (0.9 to 1.5)
Women	3.5 (2.5 to 5.1)		1		1
Age		0.000		0.008
<24	1.0		1.0		1
25–44	15.7 (7.4 to 33.5)		1.5 (1.2 to 2.0)		2.2 (1.8 to 2.8)
45+	58.4 (26.8 to 127.3)		1.3 (1.0 to 1.8)		4.2 (3.2 to 5.5)
Education		0.000		0.158		0.000
None	2.2 (1.6 to 3.0)		1.2 (1.0 to 1.6)		2.0 (1.6 to 2.4)
1–5 year(s)	1.5 (1.1 to 2.1)		1.2 (1.0 to 1.5)		1.6 (1.3 to 1.9)
6+ years	1		1		1
Marital status		0.598		0.001		0.009
Unmarried	1		1		1
Married	1.3 (0.7 to 2.5)		1.4 (1.1 to 1.9)		1.1 (0.9 to 1.4)
Divorced	0.6 (0.1 to 2.8)		2.9 (1.0 to 8.7)		1.4 (0.5 to 3.5)
Widowed	1.3 (0.6 to 2.6)		2.6 (1.6 to 4.2)		1.8 (1.2 to 2.5)
Occupation		0.1714		0.000		0.000
Unemployed	1		1		1
Day-labour	1.4 (0.9 to 2.0)		3.9 (2.8 to 5.2)		3.5 (2.7 to 4.5)
Farmer	1.1 (0.7 to 1.7)		2.2 (1.6 to 3.1)		1.6 (1.2 to 2.2)
Job/business	1.1 (0.7 to 1.7)		3.0 (2.2 to 4.1)		2.7 (2.0 to 3.5)
Housewife	0.9 (0.7 to 1.2)		1.8 (1.3 to 2.7)		1.5 (1.2 to 1.8)
Others	0.4 (0.1 to 1.2)		2.8 (1.8 to 4.5)		1.6 (1.0 to 2.4)
Asset quintile		0.5941		0.231		0.144
Lowest	1.0 (0.7 to 1.4)		1.0 (0.7 to 1.3)		0.9 (0.7 to 1.2)
Second	1.2 (0.8 to 1.6)		1.0 (0.7 to 1.3)		1.0 (0.8 to 1.2)
Middle	1.1 (0.8 to 1.5)		1.2 (0.9 to 1.6)		1.2 (0.9 to 1.5)
Fourth	1.2 (0.9 to 1.6)		1.2 (0.9 to 1.6)		1.2 (0.9 to 1.4)
Highest	1		1		1

Discussion

This study examined the prevalence, pattern and sociodemographic differentials in the use of SLT in two common forms, sadapatha and zarda. The findings of the study are somewhat consistent with the findings from the Southeast Asian region and other low-income countries.2 6 39 40 The higher use of these SLT products among the disadvantaged population in the study area clearly indicated the vulnerability of this section of the population to the hazard of using SLT. Use of SLT products in the study area and in Bangladesh has been deep rooted in the Bangladeshi culture of hospitality. SLT products are the most common form of hospitality for both casual family visitors and invited guests. Paan (betel leaves) is always served in Bangladesh with betel nuts and quite often with either of these two forms of SLT. In addition, there are innumerable ready-to-eat paan retailers in the country who always offer SLT with paan, making it very convenient to use. Use of SLT is also believed to be helpful during the early period of pregnancy by reducing vomiting and increasing the appetite.41

SLT products are less expensive than manufactured cigarettes. Furthermore, the use of SLT is more acceptable among women than tobacco smoking because it does not involve public exposure of the behaviour, as is the case with smoking products like cigarettes. There is an element of age and social hierarchy for smoking in Bangladesh: younger people and those from a lower socioeconomic group do not smoke in front of older people and those from a higher socioeconomic group. Breaking this hierarchy for smoking with subtle threshold levels, which are hard to define, is considered impolite. Thus, SLT is a very safe and convenient alternative to tobacco with smoke. This aspect of social conformity of SLT use makes it easy to be adopted by the younger generation and the process continues.

Control of the use of SLT in Bangladeshi society will therefore require a massive social awareness programme using the education system and various electronic media alongside the tobacco control programme. The traditional approach of labelling packages with warnings will not work for sadapatha because it is not industrially produced and does not have any packaging—they are presented as dried tobacco leaves. Although zarda is presented in packages, the ready-to-use paan retailers put a pinch of zarda or a piece of sadapatha in the paan if the customer requests it. Users therefore do not see the packet of zarda unless the whole packet is bought for domestic use. An additional challenge with sadapatha is that it can be home grown and therefore it is easy to use. It is very convenient for paan eaters to get a pinch of zarda or sadapatha.

Study limitations

Our study has some limitations that need to be mentioned. First, SLT consumption was self-reported and may not be accurate. Evidence from other studies suggests that SLT users could be missed when self-reporting is used.42 Second, the study was based on a cross-sectional survey allowing only assessment of association and no causal relationship. Third, the study was not based on a statistically nationally representative sample and does not necessarily portray the national situation and may not be generalisable.