The current alcohol use among males and females was 48.1% and 1.2% respectively. According to this study 65.7% were lifetime abstainers. A similar large study, conducted by the WHO for the World Health Survey (WHS -2003) in 2003 reported prevalence of 83.4%, 32.5% and 1.8% for overall abstinence, male current drinking and female current drinking respectively . Therefore, our study carried out five years later shows an overall increase of alcohol use especially among males. However, the WHO GENACIS study conducted in 2002 – 2003 with a smaller sample size (n = 1201) showed a higher prevalence of current drinking both in men (53.1%) and women (6.4%) . We recruited those who have abstained from drinking within the preceding six months as former drinkers. In other studies this was usually one year. This can underestimate the prevalence of current drinking in our study. The pattern of ethnic specific drinking in our study is similar to the data of the WHO GENACIS study.
According to the WHS -2003, the overall, male and female lifetime abstinence in India was 89.6%, 80.2% and 98.4% respectively . Another large study in India among those above 10 years of age in 2005 reported male and female current drinking prevalences as 7.9% and 1.0% respectively. Compared with Sri Lanka, in Bangladesh which is a predominantly a Muslim country in South Asia, higher prevalence of lifetime abstinence were reported both for men (87.4%) and women (99.7%) (WHS -2003). In sharp contrast to the Sri Lankan data, the developed countries have much higher prevalence of alcohol use both in men and women; North America (Male 73%, Female 58%), Europe (Male 90%, Female 81%) and Western Pacific (Male 87%, Female 77%) . The prevalence of alcohol use is lower than the Sri Lankan figures in predominantly Muslim countries in the Middle Eastern region (Iran and Saudi Arabia: Male 18%, Female 4%; Pakistan and Afghanistan: Male 17%, Female 1%) . In a recent study from urban China, the prevalence of current drinking has been reported as 68% . The above comparisons made with local and international data were done to identify a general trend and are interpreted with caution as the definitions used for drinking and populations studied, varied among these studies.
Our study showed that the middle aged males and young females have higher frequencies of drinking. The age specific alcohol consumption among males in our study is similar to the WHO GENACIS study conducted in Sri Lanka in 2002 and many other regional studies [9, 28–30]. However, the higher prevalence of alcohol use among the young women in our study differs from the WHO GENACIS study and other regional studies [9, 28, 30] except for one study from Hong Kong .
According to our data, the majority of Sri Lankan adults; 79% of men and almost all women who consume alcohol drink less than 7 units per week on average. These data are similar to the previous findings from the WHO GENACIS study . Among men 37.9% drink < 7, 2.6% between 7 – 14, 2.4% between 14 – 21 and 5.2% drink > 21 units of alcohol per week. The fact that there was no gradual decline in the frequency between the different categories of severity but a jump from the lowest to the highest would demonstrate the difficulty of maintaining drinking within non-hazardous/safe limits (male <21 and female <14) except for very light drinking due to the addictive properties of alcohol. Similar to our data, a low prevalence of hazardous drinking has been reported from a study from Hong Kong . In contrast, much higher levels of hazardous drinking have been reported from the UK (31% of men and 20% of women consumed more than 21 units and 14 units of alcohol respectively) . In the Australian workforce risky drinking (14–42 units per week) was seen in 7.4% of males and high risk drinking (>42 drinks per week) in 3.3% .
In the present study the highest prevalence of drinking in males was seen in those with medium level of education. In women, the highest current drinking was in those with the highest level of education. In contrast the hazardous drinking in males was highest in those with the lowest level of education. Our data for current drinking and education in males compares with the previous data from the WHO GENACIS study but was different in women in whom the highest prevalence of drinking was in those with middle level of education . Similar data on the association of hazardous drinking and lower levels of education has been reported in many previous studies both from Asia and outside [28, 29, 32, 33].
Among the different income groups, current drinking was higher in those with higher levels of income both in men and women. In contrast among males, hazardous drinking was equally high in both the higher and lower income groups with middle income groups having the lower prevalence. This has not been described in the previous studies on alcohol in Sri Lanka but similar pattern has been described both for current drinking and high risk drinking among the Australians . In contrast to our data, higher prevalence of current drinking was reported in the low income groups in India . Among the occupation groups both in men and women senior officials and managers had the highest prevalence of current drinking. Men in the same category had the highest prevalence of hazardous drinking followed by those engaged in elementary occupations and blue collar occupations. Hazardous drinking was very low among male professionals, clerks and the small number of participants from the armed forces. The higher prevalence of drinking in senior officials and managers and lower prevalence of hazardous drinking in those in the armed forces may partly be due to the bias created by the small number in these categories. In the study among the Australian work force higher prevalence of hazardous drinking was reported in the blue collar workers and white collar workers compared to professionals .
The current smoking prevalence is 38.0% among males and 0.1% among females older than 18 years in Sri Lanka . Our study shows a strong association of smoking with both current drinking and hazardous drinking among Sri Lankan males. A significantly high proportion of quitters have given up drinking. Combined interventions for prevention of both smoking and alcohol use may have added benefits in this population. This has not been examined in the previously published studies from Sri Lanka but has been shown in studies from other countries [29, 34, 35].
The main strength of this study is the recruitment of a representative sample via cluster sampling with good response rate (91%). There are several limitations. Due to the prevailed war situation in the North and the East, the data on Sri Lankan Tamil and Moor ethnicities living on the other parts of the country may not represent the real picture on those ethnic groups. Comparisons and conclusions on certain ethnicities and occupations could have been affected due to the effect of small sample size. We classified those who have been drinking within the preceding six months as current drinkers. In other studies this was usually one year. This can underestimate the prevalence of current drinking compared to other studies. Alcohol is a social taboo especially among women in Sri Lanka. Therefore, there is a possibility of under reporting of current drinking among women although we used experienced and trained interviewers. Anecdotally there are small isolated pockets of high prevalence of female drinking in the North Western province in Sri Lanka. These pockets were not selected in the random selection process in our study. However since this is a very minor proportion of the overall Sri Lankan population, the effect on the overall results would be minimal. The alcohol content in illicit beverages is not standardized and, this may affect the accuracy of the calculation of units. Unavailability of drinking status of the 9% non-responders, recording only the total units of alcohol without the types of alcoholic drink and the drinking frequencies during data collection were among the other limitations.
The findings shown in this study have several implications in the management and prevention of alcohol use among adults in Sri Lanka. Nearly 50% of the males are current drinkers. This is unacceptably high in a predominantly Asian Buddhist culture which discourages use of alcohol according to the religion and cultural norms. The high prevalence of drinking in the higher social strata (higher education, income and occupational groups) both among men and women shows increasingly high acceptance of drinking even among women. These patterns can set unacceptable examples among wider community in Sri Lanka. Although highly educated people tend to avoid hazardous drinking, higher percentages of them continue to be current drinkers. In contrast higher percentages of poor and less educated indulged in hazardous drinking although the overall prevalence of alcohol consumption was lower than the highly educated. We feel that this can trigger a vicious cycle of increasing poverty, violence, fragmentation of families and other harmful consequences of alcohol abuse in the poor and less educated segments in the society compared to the educated. All highly educated were not rich, as income levels do not always increase with the level of education in Sri Lanka (professionals work in the government sector). This explains why hazardous and current drinking was evenly spread over all income categories. Multiple factors may have contributed to higher current drinking in the urban educated groups. They have a higher disposable income, more opportunities for consumption in terms of social gatherings, and also use alcohol as a means of socializing during business meetings.
The associations between smoking and drinking indicate that those who continue one habit continues the other and those who give up tend to give up both. This phenomenon needs to be taken into consideration in future preventive programmes on smoking and drinking cessation and prevention.