UICC GLOBALink
The International Tobacco-Control Network

Tobacco use in Nairobi, Kenya


   

Paul Wangai, Jr. Consultant Physician,
Chairman Tobacco Control Committee,
Kenya Medical association
Nairobi, Kenya

Munyua Waiyaki, Senior Lecturer
Kenyatta University
Nairobi, Kenya

Annie J Sasco, Chief
Programme for Epidemiology for Cancer Prevention
International Agency for Research on Cancer (IARC)
for World Health Organization (WHO)
Lyon, France

All correspondence to: Dr. Paul Wangai, Jr.
PO Box 62610
Nairobi, Kenya
Email: 100077.2020@compuServe.Com

ABSTRACT

Objective: To investigate tobacco use among Nairobi City Residents in Kenya.

Method and subject: Cross?sectional survey using established census tracts for Nairobi. 1000 people invited to participate.

Results: Response rate of 61.4% was low. Study reveals high level of tobacco use (67% of men, 32% women who responded) - mainly cigarette smoking (92% ). Low education, low socio?economic status, being employed and increasing age attributed to most smokers.

Conclusion: Strategy to control tobacco use in Kenya could include education about dangers, ban on tobacco advertising and reduce access through high taxes. Regional cooperation is essential to make strategies effective.

(Afr j Med Pract, 2000;7(1):13-20)

INTRODUCTION

Diseases related to lifestyle are on the increase. Among these tobacco use ranks high. Despite declining use in most "developed" nations the "developing" nations are recording an escalating trend. The risks of tobacco are not well known by the public in these developing countries where uncontrolled tobacco promotion makes it easy for health warnings to be ignored.

Of great concern is the increase of tobacco among adolescents, youth, adults and women. Children living in urban areas and congested periurban areas are introduced to tobacco use very early. It is a common sight to see children collecting discarded cigarette butts from the streets for smoking.

Availability of tobacco single sticks at very low cost without an age limit, makes it easy for children to develop the smoking habit. Secondly, peer pressure for teenagers to be "cool" "grown Up" and "western" associates these norms with tobacco use. Thirdly, adults openly smoking at home further legitimize tobacco use to gullible impressionable children who assume. They see smoking as part of maturity.

To further understand the tobacco use habits of urban and periurban East Africans, this study was conducted in Nairobi. The study was designed to examine and understand tobacco use habits found among residents of the greater Nairobi City area. Generally, the study sought to answer.

  • Male/female ratios of tobacco users
  • Types of tobacco use in different gender, occupation, age groups educational levels
  • Frequency and quantity of cigarettes smoked
  • Children smoking
  • Methods of tobacco use risks
  • Knowledge of tobacco use risks
  • Attempts at cessation
  • Multiple drug use associated with tobacco use

SUBJECTS AND METHODS

The study subjects came from the population of greater Nairobi City area in Kenya. Nairobi is the largest, most populous capital city in Kenya. It has about 5 million inhabitants.

It is a multi-ethnic, multi-racial city at an altitude of over 5000 feet above sea level. The heart of the city is the business centre; surrounded by greater areas of residential estates characterised by both high quality homes and dilapidated slums. Various socio-economic status strata live in fairly well defined estates.

Due to high concentration of institutions of learning, the city has a high noticeable student population attending primary and secondary schools, colleges and universities. There are also large numbers of school drop outs and employment seekers.

It is because of this relatively vast and heterogeneous populace, that Nairobi becomes appropriate for tobacco use.

the purpose of the study, Nairobi was divided into five defined sections:

  1. EAST - Encompassing residential estates of high middle class e.g. Buru Buru area
  2. WEST - Encompassing high class shopping area e.g. Westlands and Parklands
  3. NORTH - Encompassing the sprawling low middle income areas e.g. Githurai and Kenyatta University
  4. SOUTH - Encompassing the slum areas of low class status e.g. Kibera
  5. CENTRAL - Encompassing central business district region of the government and private offices and businesses.

SAMPLE SELECTION

Questionnaires [n=1,000) were distributed to randomly selected individuals i.e. 250 from each of the 4 sections. The interview schedules were conducted by trained interviewers.

The randomization followed well established census tracts from the National census bureau of the government, to ensure stratification of respondents. This is a carefully developed census division of the city of Nairobi. In this way respondents represented society at large in all areas such as residential homes, schools, offices, shopping areas, bars, streets, industrial and business premises.

Each questionnaire was used by a trained interviewer to facilitate acquisition of quality from respondents. Any unwilling respondents were encouraged to answer as much as they could, as well as state reasons for their unwillingness.

INTERVIEWER TRAINING

A total of 5 Research Assistants were trained by the Principal Researchers using the developed questionnaire for field testing. The questionnaire underwent 3 modifications on the basis of field testing response.

Research assistants comprised of two PhD students who are lectures and 3 BSc students [full time]. These research assistants further trained a team of 50 interviewers, who con ducted the survey under their supervision. They in turn were supervised by the Principal Researchers.

In addition to these, there were several hired guides to assist interviewers and Research assistants especially in the slum residential areas, that are dangerous to walk around unescorted.

The Principal Researchers ensured that all interviewers knew how to complete the questionnaires, objectives of the research, expected courtesy to interviewees and avoiding coercion of respondents verbally or by any incentives [financial or otherwise]. No promises to respondents was to be given in any form.

RESULTS AND FINDING

Of the 1000 questionnaires filled, 614 respondents willingly gave useful information. 410 males and 204 females, with ages ranging from 12-17 years. The other respondents were unwilling or reluctant to give any useful information.

1) INTERPRETATION OF FINDINGS

The number of respondents from all sections were lower than expected. This unwillingness of respondents was due to suspicion that we were "establishment agents" on a mission to harm them in some way.

As a result, only 61.4% of the sample respondents provided useful information. It is noteworthy that Kenya at the time was going through a stormy political transition period into pluralistic democracy. This had turbulent impact on all areas of life. We attributed this climate to the low respondent rates we saw.

2) CHARACTERISITICS OF THE RESPONDENTS

A description of the 614 respondents in the study by 5 key variables considered to be associated with smoking habits are given in table 1. As indicated in this table, respondents who had received adult education were grouped together with those in the primary level category.

The marital status variable categorized the respondents according to whether they were married single, divorced or widowed. Eight occupational categories were created to facilitate the description of the respondents by occupational status. The technicians category included mechanics, electricians etc; while those in the professional category included Lecturers, school teachers, medical doctors, executives, accountants and those in the banking sector.

House Workers, casual workers, messengers, guards, truck and commercial public taxi drivers [called matatus"] were categorized unskilled workers. The skilled category included supervisors, secretaries, typists, salesmen and work superintendents; while soldiers, businessmen, policemen and those who were retired were considered under the category of "others". Students, unemployed and clerical officers were placed in these stated respective categories. Cases in which the occupants were not stated were excluded from the analysis.

Table I: Characteristics of the respondents

Sex % N=610
Male 65.1 397
Female 34.9 2l3
Age    N=594
>20 4.8   
20-29 38.9   
30-39 32.9   
40-49 17.0   
50+ 6.9   
Education    N=614
None 7.2   
Primary + Adults 23.6   
Secondary 30.3   
Above Secondary 38.9   
Marital Status    N=611
Married 53.5   
Single 39.8   
Divorced 4.4   
Widowed 2.3   
Occupation    N=589
Technical 2.9   
Professional 15.3   
Unskilled 19.4   
Clerical 3.7   
Skilled 6.6   
Others 18.4   
Unemployed 18.7   
Student 14.8   

Examining the percentages in the table reveal that the male respondents in the study were nearly twice as many as the females. Males constituted slightly over 65% of the respondents. Most of the respondents were in the age groups between 20 and 40 years. As shown in the table the two age groups constituted over 70% of the respondents while the combined proportion of the respondents aged less than 20 years and over 50 years is about 12%. Slightly less than 54% of the respondents were married while about 40% were single. The remaining 7% were either divorced or widowed.

The sample was biased towards those with higher education. Nearly 70% of the respondents had attained secondary or above level of education while only slightly more than 7% were in the 'none' category. Respondents with either primary or adult education were 23.6%.

The distribution of the respondents for the occupation variable revealed a much more even distribution for the various categories.

Table 2 (a): Overall Prevalence Rates by Sex in %

Sex No of users All %
Male 265 394 66.8
Female 68 213 31.9
Total 333 610 54.6

The unskilled constituted slightly over 19% of the respondents, those in the professional category 15.3%, the others and unemployed were 18.7% respectively. About 15% of the respondents were students. Distributions for the three remaining categories were as follows: Technicians [12.9%]; Clerical [3.7%] and the skilled [ 16.6% ].

To calculate prevalance of smokers, the first step was to develop a table containing the numbers of respondents falling in each of the categories and the corresponding users. Prevalance was then calculated as a ratio of the users in each of the category against the corresponding total number of respondents for the category. These ratios were next converted into percentages. The calculation of these rates is illustrated using the data summarized in table 2 [a]. As shown in this table, out of the 610 respondents in the study, 333 were smokers. Thus the prevalance rate was simply calculated as a ratio of 333 and 61 and then multiplied by 100 to obtain the percentages. The other rates summarized in the subsequent tables were similarly obtained. As in the results presented in the subsection above, the calculation of these rates was based on the cases in which the information was available. Thus, the cases falling under the "not stated" categories were not included in the calculations of these rates.

Examination of the overall prevalance rates as summarized in table 2[a]. above, suggest that the prevalance of smoking in Nairobi is quite high with over half of the respondents indicating that they were smokers. As shown in this table the prevalance of smoking is nearly 55%. These results further show, as expected, that smoking is more prevalent [about double] among men compared to women.

Table 2(b): Prevalance by types of cigarettes smoked by sex in %

Type Men Women All
   % % %
Cigarettes 91.9 92.4 91.7
Cigars 0.0 0.0 0.0
Pipe 1.1 10.0 1.0
Snuff 3.3 1.5 3.0
Chewed 0.3 3.0 1.2
Combination 3.3 3.0 3.3

Nearly all of those reported to be smoking were cigarettes. As shown in table 2 [b] the overall usage of cigarettes was about 92% with little variation between the sexes. Those taking snuff were 3% with the proportion of men taking this drug being higher [3.3% ] compared to the proportion of women.[1.5%].

The majority of users (nearly 89%) smoked on a daily basis while the rest were not daily smokers. Men were more likely to be daily smokers (91.5%) compared to women.

Nearly 41% of the smokers indicated (that they had been smoking for between 1 and 10 years while those who had smoked for less than 1 year were 6 6% . The proportion of those who had smoked for over 20 years was 10.7%. A substantial proportion [25.5%] did not specify the length of use. Table 2 [c] shows that men had smoked for longer than the women.

Table 2(c): Duration of smoking in years

Length of Use Males Females All
< 1 5.6 10.3 6.6
1 - < 5 21.1 34.5 23.8
5 - < 10 18.1 12.1 16.9
10 - < 15 11.6 8.6 11.0
15 - < 20 5.6 5.2 5.5
20 - < 25 8.2 3.4 7.2
> 25 3.0 5.2 3.5
Years not indicated 25.7 12.0 25.5

For example, the proportion of females who duration of smoking was between 1 and 5 years was 34.5% compared with 21.1% for the males. At higher duration of use the proportions for the males are persistently higher except for those who had smoked for more than 25 years.

Smoking appears to be associated with demographic and social economic factors as shown by the results of the three variables examined in this analysis namely: age, education and [h] presents the results for education and occupation, respectively.

Table 2(d): Quantity of smoking for those who smoke on a daily basis in %

No. of cigarettes Males(%) Females(%) All(%)
< 1 5.3 3.4 4.9
1 - < 5 17.4 36.2 21.0
5 - < 10 32.8 19.0 30.2
10 - < 15 24.7 20.7 23.9
15 - < 20 3.2 3.4 3.3
> 20 16.6 5.2 14.4

Results in table 2 [f] suggest that smoking increases with age. Prevalance increases from less than 45% for those aged less than 20 years compared to 83% for those aged more than 50 years. While this pattern appears to hold for men; prevalance does not show any pattern with age among women.

Table 2(f): Prevalance % by Age

Age Group Males(%) Females(%) All(%)
> 20 60.0 28.6 44.8
20-29 54.8 28.6 47.6
30-39 70.3 36.5 57.3
40-49 78.7 30.0 59.4
50 + 80.6    82.9

The results for the prevalence by the attainment level of education as shown in table 2 [h] suggest that a higher education level is associated with lower levels of smoking. This trend appears to hold for both men and women.

Table 2 [h] in which the prevalence rates with occupation are reported suggest an inverse association between smoking and occupational status. For example, smoking prevalence rate is only slightly more than 42% among the professionals compared to over 90% for the unskilled. The low figures among the unemployed and students probably reflect lack of financial resources among these groups.

Table 2(g): Prevalance % by education

Education Males(%) Females(%) All(%)
None 100 - 100
Primary school + 89.8 56.8 81.4
Secondary school 69.2 30.4 54.8
Above secondary school 47.9 13.4 33.9

Table 2(h): Prevalance % by occupation

Occupation Males(%) Females(%) All(%)
Technicians 100 - 100
Professional 49.0 19.5 42.2
Unskilled 77.5 - 90.4
Clerical 72.5 0.0 61.9
Skilled 81.3 69.6 74.4
Others 81.1 5.4 55.9
Unemployed 73.1 22.4 46.4
Students 45.6 0.0 35.6

The Male: Female ratio was 2:1 Most were married. Males were predominantly head of households. Level of education was predominantly high school and primary level of education.

The major age group of current heavy smokers for males was 20?40 years and for females 17-35. This shows an earlier smoking pattern for females compared to males.

The quantity used frequently was 2-10 cigarettes for males and 2-6 for females. There was no significant association to level of education for males but a greater tobacco use prevalence in more educated females [college and university level].

Table 3: Tobacco use prevalence in %

   Past Users Current Smokers Non Smokers
Males 10.3 66.8 23
Females 7.5 31.9 61

Students, unemployed and artisan males had higher rates; while bar maids, businesswoman and secretary/typist occupations were highest in females.

DISCUSSION

The population in Nairobi greater suburb and environments has a very high tobacco use prevalence of 66.8% in men and 31.9% of women. This probably has to do with affluence, western influence and changed cultural values. It is a very high prevalence compared to those previously reported in other African Studies.

There was a high prevalance in educated women who were working in an environment where they dealt with customers. This is a new phenomenon in East Africa where upto now, tobacco use was uncommon in women. With westernization and modernization this pattern is now changing. It calls for an aggressive tobacco preventive in adolescent health programmes as part of primary prevention.

The study findings are critical to tobacco control success in East Africa. They bring out factors that influence tobacco users among the youth.

Many of the tobacco users want to stop the habit but need assistance. Our tobacco control group need to provide such opportunities.

Many respondents were unwilling to cooperate with interviewers because they suspected us being 'government officials' looking for drug addicts or peddlers. They did not say that they would not cooperate directly; but rather circumvented issues and made the exercise futile by frustrating the interview process. Others demanded payment in exchange for information.

This survey is a landmark study after the only other one of 100 or so respondents in a 1978 work by D R Syme.

REFERENCES

  1. Harris Eze, A.O.C., " Smoking Habits in Nigeria" East African Medical Journal 1993;70:12.
  2. Onadeko, B et al "Smoking Patterns in Students of Nigeria" African Medical Science 1987;16:9-14.
  3. Nyong'o A. "Lung Cancer in Tropical Africa." East African Medical Journal 1995;92:5
  4. Betrand, E. "CHD in Black Africans: An overview " East African Medical Journal 1995;72:1:1
  5. Nyong'o, A "Smoking and Reproductive Health": Cigarette Smoking as a risk factor." East African Medical Journal. 1994;71:1:1
  6. Steyn K. et al "Smoking in the Black Community of South Africa." East African Medical Journal, 1994;71:12:12
  7. Syme, D R Tobacco Survey in selected High School in Kenya. Personal communication 1978.