Smoking is decreasing in developed countries [13–15], but increasing in developing countries including Turkey [1, 2, 16–18].
One of the targets of the WHO 'Health for All in the 21st Century' policy is to reduce the proportion of smokers to lower than 20% in over 15s and to 0% in under 15s (Target 12) . The rate was found to be 29.5% in over 15s in our study and this shows how far we are from this target. Also, passive smoking was present in 70% of the houses. These rates show the extent and importance of the problem in the region.
Smoking rates vary between the different regions of our country, as found in various studies [19–21]. In a study of the rural areas of Izmir, in the south-west of the country, the rate of current smokers was 13.2% in women and 64.3% in men aged 20 years and over . In adolescents in Kocaeli, in the north-west, this rate was found to be 50.3%  compared to 32.1% for the same age group in our study. In another study in Aydın, again in the south-west, the rate was found to be 48.3% . A study performed in 1988 showed that 62.8% of men and 24.3% of women aged 15 years and over were smoking in Turkey. The smoking rate of the total population was 43.6% . These rates are higher than in our study. All this data shows that smoking is not only a big problem in the SEAP region but also in the rest of the country. However, it is pleasing that the rate of smoking in women in the SEAP region is approximately half that of the country.
Previous data on the prevalence of smoking in the South-east Anatolian Project (SEAP) Region is lacking. Due to this limitation, smoking trends in the region are not obvious. However the Ministry of Health reported it to be 29% for the whole south-east Anatolian region in 1993  and this result is very close to the prevalence found in this study. Another study conducted in Elazig (a city near the SEAP region) in 1997, found the smoking prevalence to be 13.4% amongst females and 52.9% amongst males . These findings indicate that the prevalence of smoking in the region has not increased with time. It is also reported that the previously increasing rate of cigarette consumption in Turkey fell in recent years .
In this study, factors associated with smoking were examined. Gender, age, type of residence and education level were determined as variables affecting smoking by logistic regression. Employment, marital status and ethnicity have also been found as variables associated with smoking but these associations did not remain significant after other factors were controlled.
Gender has been determined as the main variable affecting smoking by logistic regression. The smoking rate was 6.7 times higher in men than in women when the other variables were fixed. Cigarette consumption was also 3 times higher in men (the mean total was 18.0 packs/year in men, 6.5 packs/year in women). In a study performed in rural Izmir this mean was found to be 16.4 packs/year in men  compared with 20.1 packs/year in men in the rural SEAP region in our study. The result of our study is higher. Although a fewer number of men in rural areas of the region smoke than urban men, their total consumption of cigarettes is much higher. In summary, these results show us that smoking is a serious problem amongst men in the SEAP region and smoking-cessation campaigns should therefore target men here. However, it has been shown that south-eastern European regions are in stage III of the smoking epidemic and thus in the next few decades we expect an increase in smoking in women and a decrease in men .
Age was another factor affecting smoking. When the other variables were fixed, smoking was seen 3.8 times more in the 35–39 age group than in the 15–19 age group. Smoking rates were 7.8% in women and 28.8% in men in the 15–19 age groups. This indicates that men, especially, start smoking earlier. A significant increase in the level of current smokers was determined in the study in both men and women of the 20–24 age group. This suggests that the 15–19 and the 20–24 age groups are important targets for anti-smoking campaigns. These results show similarities with the results from 1993 .
It was important that higher smoking rates were associated with advancing age. However, the rate of smoking was seen to be decreasing in women over 45 years of age and in men over 50. One of the reasons for this decrease is thought to be due to people giving up smoking at this age due to smoking related health problems. The level of cessation of smoking was 2.5 times higher in this age group (10.1%) than younger ages (4.3%).
In this study, the smoking rate was seen to be 1.7 times higher in urban regions than rural regions. This was the case in both sexes (p < 0.001) and especially in women, with three times more urban women smoking than rural women. This indicates that urban women are a suitable target for future smoking-cessation campaigns. A higher level of education and economic independence can be speculated to be the contributing factors to this observation. Higher socio-economic levels are obtained by education. Smoking rates increased with education attainment in women. The highest smoking rates were in women who had graduated from secondary school or higher. Smoking was seen to be 1.7 times more prevalent in this group than in those who were classed as illiterate. Similar results were obtained regarding men. Higher education levels should be expected to reduce smoking rates. In a study in Estonia, a negative relationship between educational level and smoking has been determined . A different result in our study is worrying for the effectiveness of education in Turkey. In several other studies conducted in Turkey, smoking was found to be highest in the well-educated groups in agreement with our study. [19, 21, 22].
Another result of this investigation was the high percentage of respiratory tract complaints in current smokers. The 'presence of a cough of more than 15 days duration' and the 'presence of sputum' was examined and these complaints were found significantly higher in smokers. The rate of 'presence of a cough of more than 15 days duration' was 8.7% in non-smokers and 15.8% in current smokers. The rate of 'presence of sputum' was 8.3% in non-smokers and 14.7% in current smokers (p < 0.001). These results are evaluated as an indicator of the negative effects of smoking on the respiratory system. It is thought that the higher rate of these complaints in men is due to them having a total cigarette consumption rate of three times higher than in women. Similarly, Steyn et al.(2002) reported that a dose-response was observed between the amount smoked and the presence of respiratory disease.
Tobacco smoke includes a lot of agents that cause lung cancer and chronic obstructive lung diseases and also exacerbates asthma [3, 6, 8]. Tobacco smoke also decreases birth weight and increases sudden infant death and allergies [3, 4]. It affects not only the smokers themselves but also others around or living with them. This effect has been shown in many studies [4–8, 27]. Continuous smoke exposure (passive smoking, involuntary smoking) affects the health of a household negatively, especially that of children and pregnant women. It has been found that passive smoking reduced the birth weight of infants of non-smoking mothers . Passive smoking is accepted as one of the important factors in the frequency of respiratory tract infections in children. Therefore, the presence of passive smoking in the household was examined in this study. There was a regular daily smoker in approximately two-thirds of the houses. It would normally be expected that smoking rates would decrease in houses where children or pregnant women were living. However it is both important and worrying that our study showed this not to be the case in the households we examined, with smoking being present in two-thirds of houses where children or pregnant women lived.
The effect of passive smoking on respiratory tract complaints was also examined in this study. The percentage of respiratory tract complaints in non-smokers was similar whether or not they experienced passive smoking within the household. However, it would normally be expected that non-smokers experiencing passive smoking within the household would have a high percentage of respiratory tract complaints. It is likely that persons are not only exposed in their home, but also in cars, public transport and other community settings. Such exposure may preclude the ability to determine health effects due to the limited variance in the exposure variable. It can be speculated that the duration of exposure must also be evaluated in order to show the difference; perhaps long-term exposure is important in passive smoking.