| Abstract|| |
Background: Tobacco use is reported to be rampant in urban slums in developing countries. Demographical variations in tobacco use between males living in urban slums vs those living in non-slum areas in India has not been reported, and this study was undertaken to address this issue.
Materials and Methods: Secondary data analysis of National Family Health Survey-3 (NFHS-3) was undertaken to study demographical variations in tobacco use between urban slum dwellers and non-slum dwellers in eight Indian cities. Demographic determinants for use of smoking and chewing forms of tobacco in the two groups were analyzed. SPSS® version 16.0 (SPSS Inc., IL, USA) was used for statistical analysis.
Result: The study population comprised 6887 (41.8%) males from slum areas and 9588 (58.2%) from non-slum areas of eight urban cities. Cigarette/beedi smoking was the commonest form of tobacco use among the study population. Pan masala use was the least common form of smokeless tobacco use, next only to snuff. There was a high statistical significance observed within the various demographic parameter studied in both the slum and non-slum dwelling males in study population. However, on studying the differences between the two groups, it was observed that statistical significance of P≤.001 was observed with age (15-49), secondary education, religion, household structure and marital status. The difference between the two groups in the mean number of cigarettes/beedis smoked was not statistically significant (P=.598).
Discussion and Conclusion: Male slum dwellers are a distinct urban population, whose health needs assessment requires a different approach than that for non-slum dwellers who often can afford the services that an urban Indian city can offer.
Keywords: India, smokeless tobacco, tobacco use, urban slum
|How to cite this article:|
Rooban T, Joshua E, Rao UK, Ranganathan K. Prevalence and correlates of tobacco use among urban adult men in India: A comparison of slum dwellers vs non-slum dwellers. Indian J Dent Res 2012;23:31-8
Tobacco use is a global public health problem. Globally, about 1.1 billion people smoke and it has been estimated that by 2025 the number would rise by 1.6 billion. Worldwide, tobacco-attributable deaths were 4.83 million in 2000. The number is projected to reach 6.4 million in 2015 and 8.3 million in 2030. In low- and middle-income countries such deaths will increase from 3.4 million in 2002 to 6.8 million in 2030. 
|How to cite this URL:|
Rooban T, Joshua E, Rao UK, Ranganathan K. Prevalence and correlates of tobacco use among urban adult men in India: A comparison of slum dwellers vs non-slum dwellers. Indian J Dent Res [serial online] 2012 [cited 2015 May 25];23:31-8. Available from: http://www.ijdr.in/text.asp?2012/23/1/31/99034
Rapid urbanization is a global phenomenon. Over the last two decades, many urban areas have expanded dramatically owing to many reasons, including rapid population growth, urban migration, and continued global economic integration.  The United Nations Expert Group in 2002 operationally defined 'slum' as a human settlement that has the following characteristics: 1) inadequate access to safe water, 2) inadequate access to sanitation and other infrastructure, 3) poor structural quality of housing, 4) overcrowding, and 5) insecure residential status. In 2007, the term 'slum' included 43% of all urban populations in all developing countries, and 78% of the urban population in the least developed countries.  It is reported that more than one billion people live in slum areas, mostly in developing countries. According to estimates the slum population will double by 2030. Recent reports have shown that the urban population, which was 3.29 billion in 2007, will become 4.58 billion by 2025. The data proceeds to indicate that most of this urban growth (1.29 billion) is bound to occur in the less developed regions and mostly in Asian cities, which will accommodate 1.21 billion people by 2025. 
In 2001, 28% of the total population in India was living in urban areas, and this was projected to increase to about 50% (605-618 million) by 2021-2025 owing to several factors, including urban migration and population explosion. Demographic trends show that while the average urban growth rate stabilized at 3% over the past decade (1991-2001), the slum growth rate doubled. An alarming feature of the growth of the urban population is the proportion of people living in poverty (urban poor includes those poor people livening in slum and non-slum areas), with official estimates placing it at 32%. Projections suggest that while the urban population will double in the next 10 years, the urban poor will double in just 5 years. It is evident that the urban poor have the worst of both worlds-they adopt an urbanized lifestyle, which places them at a higher risk for non-communicable diseases, and they have poor access to healthcare, which is partly related to their poor purchasing ability. 
It is reported that socioeconomic status is a major factor determining tobacco use. Several studies have shown that the poor and uneducated are at an increased risk of tobacco use.  A harmful consequence of smoking, documented in Bangladesh, is the damage done to poor families when tobacco gets priority over food and other essentials. It has also been reported that those with the lowest standard of living are likely to be the heaviest smokers.  An unpublished study found that the average age at which Mumbai street-children started using tobacco was 11·3 years, and the most frequently used products were raw tobacco, gutkha (a blend of tobacco and flavorings), and cigarettes/beedis. 
Tobacco use is reported to be rampant in urban slums in developing countries.  However the demographical variation in tobacco use in male Indian urban slum dwellers vs non-slum dwellers has not been reported, and therefore this study was undertaken to address this issue.
| Materials and Methods|| |
We undertook a secondary data analysis of the National Family Health Survey-3 (NFHS-3) to study the variation in tobacco use between urban slum and non-slum dwellers. NFHS-3 during 2005-2006, conducted a multi-level, nationally representative, cross-sectional, household sample survey with a questionnaire for men, women, and children. A uniform sample design was adopted in all the states.  The survey questionnaire had three questions addressing chewable tobacco use; these questions were 'Do you currently smoke cigarettes or beedis?' 'Do you currently smoke or use tobacco in any other form?' and 'In what other form do you currently smoke or use tobacco?' The choices for the last question were: Cigar/pipe; pan masala; ghutka; other chewing tobacco; and snuff. The answer for the last question was recorded in a 'yes/no' format in the database.
The sample households for the study were selected using a two-stage sampling design. In the first stage, census enumeration blocks (CEB) were selected and, in the second stage, households within the CEBs were selected. In each city, slum and non-slum primary sampling units (PSU) were selected independently from the respective lists of slum and non-slum CEBs. On special request, for the NFHS-3, the Registrar General of India made available CEB-wise data for the eight cities for which slum/non-slum estimates were provided. Two separate lists of all the slum and non-slum CEBs in all the wards of each city served as two separate sampling frames at the first stage of selection. From each sampling frame, slum and non-slum PSUs were selected using proportional to population size (PPS) sampling. The house listing carried out in each of the selected CEBs served as the sampling frame for the selection of households.  Only the unweighted data (slum/non-slum) collected by the supervisor in the field study were considered for the present study. The definitions employed by NFHS-3 for the different demographical variables were used for this study also.  The definition of 'slum' used in the NFHS-3 was employed in this study too. 
With regard to tobacco use, the study subjects were categorized into four groups: Those with no habits, those habituated to smoking tobacco use (i.e., cigarettes, beedis, cigar, pipe, or other form), those habituated to smokeless tobacco use, and those habituated to multiple tobacco products. These formed the dependent variables. The independent variables were age-group, education, religion, household structure, marital status, occupation, and standard of living (SLI) index. The SLI index is derived using a scoring system where the house, facilities associated with the house, and physical items belonging to the household are given scores, which are summed. The result is measured against a set of SLI cut-offs. Households with a score 0-14 are classified as having a 'low' SLI, a score of 15-24 indicates a 'medium' SLI, and scores ≥25 indicate a 'high' SLI. As with House Type, if any of the variables from which the scores are drawn are missing, 'don't know', or 'other', the SLI for that household is then set to missing, wealth index (index of the economic status of households based on 33 characteristics). Each household asset is assigned a weight (factor score) generated through principal components analysis, and the resulting asset scores are standardized in relation to a normal distribution with a mean of zero and standard deviation of one. Each household is then assigned a score for each asset, and the scores are summed for each household; individuals are ranked according to the score of the household in which they reside. The sample is then divided into quintiles  along with city and slum/non-slum dwelling status as determined by the field supervisor formed the independent variables.
Body mass index (BMI) of the subjects was calculated and grouped as in the NFHS-3 survey. Usage statistics (e.g., the number of cigarettes/beedis smoked during the 24 hours preceeding the interview) were also obtained.
SPSS® version 16.0 (SPSS Inc., IL, USA) was used to carry out statistical analysis. Descriptive variables are presented for the demographic variables. Overall prevalences of various forms of tobacco use were computed for various demographic variables as point estimates. The results were cross-tabulated and the chi-square test was employed to find the significance of relationships between parameters. One-way ANOVA was employed to find the mean of BMI with regard to any form of tobacco use and dwelling status. Binary logistic regression by entering and simple categorical method was employed to calculate odds ratio (OR) and 95% confidence intervals (CI). P≤.001 was taken as significant.
| Result|| |
The study population comprised 6887 (41.8%) slum dwellers and 9588 (58.2%) non-slum dwellers from eight Indian urban cities (Delhi, Meerut, Kolkotta, Indore, Mumbai, Nagpur, Hyderabad, and Chennai).
The basic demographics of the study population have been described in detail earlier.  Tobacco use in the population is shown in Graph 1. Cigarette/beedi smoking was the commonest form of tobacco use among the study population. Pan masala use was the least common form of smokeless tobacco used, next only to snuff.
[Table 1] depicts the demographics of the study population. Age, education, religion, household structure, and marital status have been cross-tabulated against types of tobacco use. There was a high statistical significance observed within the various demographic parameters studied in both the groups in the study population. However, on studying the differences between the two groups, it was observed that statistical significance of P≤.001 was observed with age (15-49), secondary education, religion (except Christianity), household structure, and marital status.
[Table 2] depicts the socioeconomic demographics of the study population; the occupation, standard of living index, and wealth index were considered. There was a statistically significant difference within the groups. The difference of slum and non-slum dwelling males who were professionals (including technical and managerial), standard of living index (medium and high) and richest, were significantly different (P≤.001).
Bivariate logistic regression for any tobacco use for slum and non-slum males revealed that, among slum dwellers, the OR for tobacco use was 5.46 in the 20-29 year age-group and 5.73 in the 40-44 year age-group (as compared to the 15-19 year age-group). In non-slum dwellers, the odds for tobacco use was highest in the 25-34 years age-group. Among slum dwellers, the OR increases from 20-29 years and drops in 30-34 years to climb up steadily till 5.73 in 40-44 years and tapers down. In contrast, in non-slum dwellers, the OR rises from 3.87 in the 20-24 years age-group to 5.29 in 25-29 years age-group, plateaus in the 35-44 years age-group, and then increases to 5.96 in the 45-49 years age-group. The odds for tobacco use decreased with increasing education in both slum and non-slum males. 'Ever married' males were more likely to use tobacco than 'never married' males; this was statistically significant in both slum and non-slum dwellers [Table 3]. As compared to a non-slum male, a slum-dwelling male had an OR of 1.38 (95% CI: 1.3 to 1.47; P=.000) for any tobacco use. For smokeless tobacco use the OR was 0.68 (95% CI: 0.64 to 0.73; P=.000) and for smoking tobacco it was 1.03 (95% CI: 0.95 to 1.12; P=.43). For multiple tobacco habits the OR was 0.71 (95% CI: 0.64 to 0.78; P=.000).
|Table 3: Bivariate logistic regression with odds ratio, confidence interval and significance for any tobacco use between slum and non-slum dwelling males|
Click here to view
With regard to the types of tobacco use, the 15-19 years age-group were more likely to use smokeless tobacco in this study population. In smoking tobacco usage, slum males had increasing OR with increasing age till OR was 6.87 while in non-slum, it was 4.95. Education was seen to be significantly associated with smoking tobacco in both slum dwellers and non-slum dwellers [Table 4].
|Table 4: Bivariate logistic regression for smokeless and smoking tobacco among slum and non-slum dwelling males|
Click here to view
The mean number of cigarettes/beedis smoked by a male during the 24 hours preceding the interview was 7.77 ± 7.89 among slum dwellers (n=2067) and 7.65 ± 7.16 among non-slum dwellers (n=2516); the difference was not statistically significant (P=.598).
[Table 5] shows the comparison of the mean BMI for the study population using one-way ANOVA. A high statistical significant difference (P=.000) of mean BMI was observed between tobacco users and non-tobacco users as well as slum and non-slum dwellers. Only non-slum males exhibited significant difference between tobacco and non-tobacco users (P=.000). Graph 2 shows the distribution of BMI among the study population.
|Table 5: One-way ANOVA results comparing BMI of any tobacco use and the study population|
Click here to view
| Discussion|| |
It is said that any form of social marginalization (including poverty) may translate into poor health status.  Among the poor, the economic status, environment, and lifestyle all contribute to the causation of disease and poor outcomes. , Communicable diseases are common in slums owing to the crowding and poor sanitation, while noncommunicable diseases related to lifestyle are increasing among the Indian urban poor. [3,8] Usage of tobacco is more common among urban slum males than in non-slum males in all the studied cities.  Though the number of tobacco users is comparatively high in slums, no large-scale study has been done in India till date to study the difference in demographic characteristics of tobacco users in slum dwellers as compared to non-slum dwellers as has been done in countries like Bangladesh. 
In India, the people living in urban slums have high levels of poverty, low SLI, and lack of education,  which is similar to the situation in other developing countries.  Several studies have reported of higher smoking rates in areas of marginalized, or overcrowded areas,  social norms, cultural beliefs and neighborhood characters.  Our findings demonstrate that men residing in the urban slum areas are more likely to use tobacco (in any form) than their counterparts living in the urban non-slum areas [Table 1], [Graph 1].
Age is an important demographic characteristic that is associated with tobacco use. Smokeless tobacco use is more popular in the younger age-groups than smoking forms, which are predominantly observed in older age-groups. This finding has been documented earlier by several studies from India  and our findings were no exception [Table 4]. Similarly, education has important implications. As shown in [Table 3], as the level of education increases there is statistically significant decrease in the use of tobacco. In this study we found that smokeless tobacco use was more common than smoking tobacco use among the better educated subjects; smoking tobacco use increases as level of education decreases. Similar findings have been published earlier, and this trend is observed across India. ,
Tobacco use is more common among the poorest of slum and non-slum dwellers. Habitual tobacco use causes harm to both users and their families and places a huge burden on families that are already reeling under poverty.  A reduction of even the small expenditure on tobacco can have a beneficial impact on the standard of living of such poor families by reducing interruption of food procuring capability and increasing access to health care and thereby can decrease the prevalence of the malnutrition among the family members.  From [Table 5], it can be seen that slum males have low BMI as compared to their non-slum counterparts. Among slum males, only 14.13% of 'any tobacco' users and 15.93% of 'no tobacco' users had BMI>18.5, while among non-slum males the corresponding figures were 19.35% and 23.93%, respectively. Thus, inadequate health knowledge, limited access to health care, and high prevalence of malnutrition can aggravate the consequences of tobacco usage among populations living in urban slums. 
This study shows that, in India, the pattern of tobacco consumption varies in different regions. Such variations have been reported in literature earlier. 
Pan masala, which is processed arecanut along with lime and condiments, is not supposed to contain tobacco, but is generally sold in pouches that have similar packaging as gutka, which contains tobacco. Some types of pan masala do contain tobacco and a high number of pan masala users in our study perceived it as a tobacco product. ,
Difference in tobacco use between urban slum and non-slum males as compared to non-tobacco users
- Among slum dwellers, tobacco use is most common in males in the fifth decade of life, while it is most common in the third and fourth decades of life among non-slum dwellers.
- More number of slum-dwellers with primary education and secondary education use tobacco than their non-slum counterparts (with the same level of education).
- Low SLI was associated with greater tobacco use in non-slum males, while relatively higher SLI was associated with tobacco use in slum males.
Difference in smokeless tobacco use between urban slum and non-slum males as compared to non-tobacco users
- The younger age-groups in both groups were more likely to use smokeless tobacco.
- Higher educated males were more likely to use smokeless forms of tobacco than uneducated or little educated males.
- In the slum areas, 'poor' males used the smokeless form of tobacco more often, while in non-slum males the 'richest' used it more often.
- SLI did not greatly affect smokeless tobacco use among slum males while, among non-slum males, those with medium and high SLI used tobacco more often.
Difference in smoking tobacco use between urban slum and non-slum males as compared to non-tobacco users
- Males in the fifth decade of life were more likely to use tobacco in both groups.
- Among slum males, the 'poorest' of were more likely to use tobacco, while 'poorer' males were more likely to use tobacco among non-slum dwellers.
- Among slum males, those with low SLI were more likely to use tobacco than others, while among non-slum males, those with medium and high SLI more often tended to use tobacco.
Though this secondary data analysis suffers from several limitations, including use of a non-weighted sample and a relatively small study population of 16518 people to represent millions of people, it provides a robust estimate of the magnitude of the problem India faces in terms of tobacco use and the non-communicable diseases it is associated with. The cross-sectional design of this survey was a limitation which prevents us from commenting on the cause-effect relationships of the significant associations and the difference between slum and non-slum males tobacco use along with absence of the duration, intensity and frequency of tobacco use. Though odds ratio is used as a measure in the present study, its use in large-scale epidemiological and clinical studies have been questioned. 
The study has some strengths: For example, a representative sample of adult males from all the eight urban megacities in India have been analyzed for the first time. As the study population was selected scientifically, this study could be considered representative for the urban areas in India. Moreover, this is the first Indian study, to our knowledge, which identifies some pattern of difference of tobacco use between males of the urban slums and non-slums.
| Conclusion|| |
Male slum dwellers are a distinct urban population, whose health needs assessment requires a different approach from that of non-slum dwellers who often can afford the services that an urban Indian city can offer. It is to be borne in mind that focused poverty alleviation solutions may not be able to address the immediate health needs of this population. The rate of growth and the sheer size of Indian urban slums will negate many of the achievements of the health sector. Accurate characterization of the determinants of chronic and acute tobacco-related diseases in slums requires long-term prospective population-based surveillance. The findings of this study could be of use in designing a much-needed educational program for tobacco control in the slum and non-slum population. Better awareness programs could be designed based on the profile of the typical tobacco user shown by this study across India.
| Acknowledgment|| |
The authors would like to thank Macros International for their willingness to share the data. We would like to thank Indian Society for Dental Research; Dr. S. Ramachandran, Principal, Ragas Dental College; and Prof. Dr. A. Kanagaraj, Chairman of Jaya Group of Institutions, Chennai, for their constant support and encouragement.
| References|| |
|1.||Khan MM, Khan A, Kraemer A, Mori M. Prevalence and correlates of smoking among urban adult men in Bangladesh: Slum versus non-slum comparison. BMC Public Health 2009;9:149. |
|2.||Riley LW, Ko AI, Unger A, Reis MG. Slum health: Diseases of neglected populations. BMC Int Health Hum Rights 2007;7:2. |
|3.||Anand K, Shah B, Yadav K, Singh R, Mathur P, Paul E, et al. Are the urban poor vulnerable to non-communicable diseases? A survey of risk factors for non-communicable diseases in urban slums of Faridabad. Natl Med J India 2007;20:115-20. |
|4.||Sharma DC. Tobacco use among India's street children raises concern. Lancet Oncol 2009;10:844. |
|5.||Samarasinghe D, Goonaratna C. Tobacco related harm in South Asia. BMJ 2004;328:780. |
|6.||Bhat PN, Arnold F, Gupta K, Kishor S, Parasuraman S, Arokiasamy P et al. Morbidity and Health care. In National Family Health Survey (NFHS-3), 2005-06: India: Volume I, 1 st ed, Mumbai: International Institute for Population Sciences (IIPS) and Macro International; 2007. p. 426-9. |
|7.||Bhat PN, Arnold F, Gupta K, Kishor S, Parasuraman S, Arokiasamy P, et al. Morbidity and Health care. In National Family Health Survey (NFHS-3), 2005-06: India: Volume II, 1 st ed, Mumbai: International Institute for Population Sciences (IIPS) and Macro International; 2007. p. 1-23. |
|8.||Gupta K, Arnold F, Lhungdim H. Health and Living Conditions in Eight Indian Cities. National Family Health Survey (NFHS-3), India, 2005-06. Mumbai: International Institute for Population Sciences; Calverton, Maryland, USA: ICF Macro; 2009. p. 24-33. |
|9.||Bhat PN, Arnold F, Gupta K, Kishor S, Parasuraman S, Arokiasamy P, et al. Morbidity and Health care. In National Family Health Survey (NFHS-3), 2005-06: India: Volume I, 1 st ed, Mumbai: International Institute for Population Sciences (IIPS) and Macro International; 2007. p. 43. |
|10.||Madhiwalla N. Healthcare in Urban Slums in India. Natl Med J India 2007;20:113-4. |
|11.||Pearce J, Hiscock R, Moon G, Barnett R. The neighbourhood effects of geographical access to tobacco retailers on individual smoking behaviour. J Epidemiol Community Health 2008;63:69-77. |
|12.||Ray CS, Gupta PC. Bidis and smokeless tobacco. Curr Sci 2009;96:1324-34. |
|13.||Rooban T, Elizabeth J, Rao UK, Ranganathan K. Sociodemographic correlates of male chewable tobacco users in India: A preliminary report of analysis of national family health survey, 2005-2006. Indian J Cancer 2010;47:s76-s85. |
|14.||Reddy SS, Ali KH. Estimation of nicotine content in popular Indian brands of smoking and chewing tobacco products. Indian J Dent Res 2008;19:88-91. |
|15.||Pepe MS, Janes H, Longton G, Leisenring W, Newcomb P. Limitations of the odds ratio in gauging performance of a diagnostic, prognostic or screening marker. Am J Epidemiol 2004;159:882-90. |
Department of Oral and Maxillofacial Pathology, Ragas Dental College and Hospital, Chennai
Source of Support: None, Conflict of Interest: None
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]