OUP user menu


Wei Hao, Zhonghua Su, Binglun Liu, Kui Zhang, Hanqing Yang, Shaozhong Chen, Meizi Biao, Chun Cui
DOI: http://dx.doi.org/10.1093/alcalc/agh018 43-52 First published online: 22 December 2003


Aims: To understand drinking patterns, health status related to drinking and the level of unrecorded alcoholic beverage consumption for the general population living in five areas of China in 2001. Methods: By cluster sampling, 24 992 community residents aged 15 years or older were interviewed by trained psychiatrists using structured questionnaires provided by WHO. Results: The 1-year drinking rate was 59.0%, and the point prevalence rate of dependence was 3.8%. The average annual consumption of pure alcohol was 4.47 l. The 1-year morbidity from gastritis/ulcer in the whole sample was 7.9%, which associated nonlinearly to alcohol intake, and heart disease and cerebral infarction/cerebral haemorrhage showed V-shaped curve relationships. Conclusions: The rate of alcohol use was higher in men than in women, and the annual alcohol consumption per capita was higher than that in the 1990s in the selected areas. Alcohol consumption plays a role in the development of alcohol-related physical diseases.

(Received 1 April 2003; first review notified 8 September 2003; in revised form 16 October 2003; accepted 24 October 2003)


Excessive drinking and alcohol-related problems are serious issues for public health world-wide. National and regional epidemiological surveys in China since the 1980s have shown that mean annual alcohol consumption and the prevalence rates of alcohol-related problems have increased steadily (Collaborative Group on Mental Disorders, 1986; Collaborating Research Group on Alcoholism and Related Problems, 1992a,b; Zhang et al., 1996; Liu et al., 1999; Wang et al., 1999; Tao et al., 2002). In 1994, we (Hao et al., 1998a,b) conducted a survey on alcohol consumption and alcohol-related problems in selected areas: the city of Zixing and its adjacent rural area in Hunan Province, the city of Fuyang and its adjacent rural area in Anhui Province, Chengdu and Panzhihua cities and their adjacent rural areas in Sichuan Province, the city of Yanji and its adjacent rural area in Jilin Province and the city of Jinan and its adjacent rural area in Shandong Province. The results of the survey showed that the male, female and total drinking rates were 84.1, 29.3 and 59.5%, respectively, and the annual consumption of pure alcohol per person was 3.60 l among individuals aged 15 or over in these communities. In order to describe current drinking frequency, mean yearly alcohol consumption, and consumption of unrecorded alcoholic beverages, and to understand health status and explore the relationship between health status and drinking in the general population in China, we conducted the present study on drinking, drinking patterns and health status related to drinking in the general population of five of the six areas in China mentioned above, under the sponsorship of the World Health Organization (WHO).


Samples and study populations

Sampling took place in five provinces: Sichuan Province in southwest China, Jilin Province in the northeast, Anhui Province in the east, Shandong Province in the north and Hunan Province in the central south. They were the city of Chengdu (CD) and its adjacent rural area in Sichuan Province (the Panzhihua site was not selected because it was in Sichuan Province also); the city of Yanji (YJ) and its adjacent rural area in Jilin Province; the city of Fuyang (FY) and its adjacent rural area in Anhui Province; the city of Jinan (JN) and its adjacent rural area in Shandong Province and the city of Hengyang (HY) and its adjacent rural area in Hunan Province. In all five areas, blocks in cities and villages in rural areas were selected with a probability proportionate to size from the primary units, which were selected randomly. Within those areas households were selected in clusters. All respondents from these five areas were individuals aged 15 years or older. All of the respondents gave written informed consent and the study protocol was approved by the Ethics committee of Central South University. Completion of the questionnaires was also voluntary.


An alcohol use and health status survey questionnaire provided by WHO, with minor modifications for social and cultural background, was used for assessing demographic characteristics (e.g. gender, age, ethnic group, occupation, current marital status, education level, family income per person per month, self-assessment on family income), drinking patterns and effects (frequency of drinking, amount of alcohol per occasion, beverage preference, acute alcohol intoxication, cost of drinking in the past week and heavy drinking), health status-related drinking, and physical and psychological diseases suffered in the past year. A copy of the survey questionnaire can be obtained on request from the first author.

An alcohol related-problems screening test (APST) was developed based on the Diagnostic Interview Schedule (DIS) (which was used in the author's survey of 1993; Hao et al., 1998a). Each item in the test has two possible answers, ‘yes’ or ‘no’. Items 1 and 2 deal with drinking amount and frequency, and items 3–11 are for acute intoxication, craving, physical dependence, withdrawal symptoms and social dysfunction problems. If two ‘yes’ answers were given (one of which must relate to items 1 or 2), a positive result was obtained.

The structured interview on drinking-related problems focused mainly on the diagnosis of alcohol-related psychiatric disorders based on DSM-III-R (American Psychiatric Association, 1987) criteria and was also used in the author's survey of 1993. A hierarchical diagnostic system was adopted in diagnosing abuse and dependence, uncomplicated alcohol withdrawal and withdrawal delirium, and alcohol amnestic disorder and dementia associated with alcoholism.

The survey took place during the period September 16–October 31, 2001.

Quality control

Interviewers were psychiatrists with at least 5 years clinical experience and experience of epidemiological surveys, and they were trained using a standard training manual for 7 days prior to the interviews. Lecture presentations were used to explain the purposes of the survey and the variables of the questionnaires, the conception of psychoactive substance misuse and dependence, and the diagnostic criteria of psychoactive substances in DSM-III-R. A pilot study was conducted after training in each site.

In order to maximize the rate of response, interviewers were trained in techniques for gaining entry to a household and maintaining rapport with respondents. Instructions also focused on guaranteeing respondents' anonymity, public relations with the community and seeking the help of community leaders. After each interview, the questionnaire was checked by the interviewers. If certain items had been missed, the participant was requested to complete them.

Statistical methods

For categorical data and ordinal categorical data, chi-squared statistics with Yates' continuity correction test and the chi-squared test of linear trend (Su, 2002) were used respectively. Student's t-tests, dependent and independent, were used for comparison of group means and one-way analysis of variance (ANOVA) was used for comparison of group means and the post hoc test for multiple comparisons. Stepwise regression models (Huang, 1995) were developed using the self-rating of health status and the variables on sociodemographic characteristics, drinking and smoking. Two tailed P-values were considered to be significant at <0.05. SPSS version 11.0 for Windows was used for all data analyses.


There were 25 643 selected participants (6309 in CD; 5349 in YJ; 5257 in FY; 4110 in JN; 4618 in HY) and of these 25 052 responded (6158 in CD; 5223 in YJ; 5178 in FY; 3952 in JN; 4541 in HY). Therefore, the completion rates were 97.5%. Non-response was due to the following: refused (0.8%); unable to contact at address (1.1%); and others (0.6%). A total of 25 052 respondents were interviewed, of whom 24 992 (99.9%) provided valid data.

Demographic characteristics

As shown in Table 1, data from 24 992 cases were gathered in the five areas. There were significant differences among the five areas in age (F = 105.483, P < 0.0001), education level (F = 826.194, P < 0.0001) and marital status (F = 463.914, P < 0.0001). The ratio of men (n = 13 992) to women (n = 11 000) in the total sample was 1.00:0.79, with the proportions in four areas (except JN) consistent with that of the country census data of 2000. The average age of the respondents was 40.2 years (SD = 15.6). The mean age of men (mean 40.4, SD 15.4) was significantly higher than that of their female counterparts (mean 39.6, SD 15.8; P < 0.0001). Han national constituted 87.0% of the sample, Korean comprised 12.6%, with the proportion of Korean people being significantly higher in YJ (χ2 = 13 582.166, d.f. = 4, P < 0.0001). 75.4% of respondents were married. The education level of the men (mean 9.1 years, SD 3.8) was significantly higher than that of the women (mean 8.0 years, SD 4.1) with the average years of education being 8.6 (SD 4.0) among all respondents (t = 22.182, P < 0.0001). The average family incomes in CD, YJ, JN and HY were significantly higher than that in FY (F = 770.259, P < 0.0001).

View this table:
Table 1.

Demographics in five sites

6155522151773950453924 992
    Male311251.0269451.6237652.8256464.9288663.613 99256.0
    Female299349.0252748.4244147.2138635.1165336.411 00044.0
Mean (year)40.7043.4237.7439.5738.5440.08
    Han group609799.9204439.1515099.5391199.0453499.521 73687.0
Marital status
    Married462675.8405577.7398877.0282471.5362179.819 11576.5
    Never married105217.282315.8109921.294423.974516.4466318.7
Educational level
    Primary or below192231.54939.4232744.9149337.8117926.0741429.7
    College or above83513.7101819.51252.447612.166614.7312012.5
Mean (year)8.5810.806.887.589.168.64
    Mild physical labour3936.453810.31673.22055.23577.916606.6
    Heavy physical labour681.11803.41142.2731.81212.75562.2
    Private business owner1432.166412.772914.11503.83928.620788.3
    Unemployed (able/unable to work)1692.83877.42384.6120.31964.310024.0
Families Income per person per month
    Mean (RMB)485.91437.09174.21491.11455.83405.71

Drinking frequency in current drinkers

Drinking status was ascertained with the two questions: ‘Have you had a drink that contains alcohol in the past 3 months?’ and ‘Have you had a drink that contains alcohol in the past 12 months?’ All respondents were required to answer the first question. If a negative answer was obtained, they were required to continue to answer the other question. In this study, ‘current drinkers’ or ‘3-month drinkers’ referred to those individuals who reported that they had had a drink in the past 3 months. ‘1-year drinkers’ included current drinkers and those who had had a drink at least in the past 12 months. The abstainers were those who had not had any alcohol beverage in the 12 months prior to the interview. The proportion of 3-month drinkers in the study was 63.8, 18.3 and 43.8%, for males, females and the total sample, respectively, and that of 1-year drinkers was 74.9, 38.8 and 59.0%, respectively. Both the 3-month and the 1-year drinking rates were higher in males than in females (χ2 = 51 793.795, d.f. = 1, P < 0.0001; χ2 = 3319.949, d.f. = 1, P < 0.0001). Table 2 shows the drinking rate in different age groups. Both 3-month and 1-year drinking rates increased with age, reached a peak at 41–50 years old in males and 36–40 years old in females and then declined.

View this table:
Table 2.

Three-month and 1-year drinking rates by age

Three-month drinking rateOne-year drinking rate
Age (years)Male (%)Female (%)Total (%)Male (%)Female (%)Total (%)

Table 3 contains data on the reported frequency of drinking among current drinkers by sex. These data show that more men than women were frequent users of alcoholic beverages, the difference was statistically significant (χ2 = 2354.612, d.f. = 12, P < 0.0001). A majority (76.7%) of female drinkers and 33.8% of male drinkers had alcohol beverages once a week or less, while 26.1% of male drinkers and only 5.4% of female drinkers drank at least once a day.

View this table:
Table 3.

Drinking frequency in current drinkers (%)

Men (8923)Women (2010)Total (10 933)

Alcohol consumption

If the respondent had had a drink that contained alcohol in the past 3 or 12 months prior the interview, he or she would be asked to answer following the two questions: ‘What kind of alcohol beverage did you drink and what was the average amount of these beverages on each occasion’ and ‘How often have you had alcoholic beverages’. The respondents were provided possible responses: (1) daily or 6 days per week; (2) 3–5 days per week; (3) one to two times per month; (4) two to three times per month; (5) once a month; (6) less often. As the alcohol content of different types of beverages is known, the amount of alcohol from each session could be converted to the total amount of pure alcohol consumed. In China, strong distilled spirits have an ethanol content of approximately 50–55% (v/v), weaker distilled spirits and paddy wine 35–44% (v/v), beer 4.5% (v/v) and grape wine, yellow wine, and rice wine 12–18% (v/v).

The average (±SD) annual consumption per capita in pure alcohol among all respondents aged 15 years or above was 4.47 ± 10.93 l. Men drank 13.4 times more than women (P < 0.001). The average annual consumption in pure alcohol for male, female and total 1-year drinkers was 10.1, 1.5 and 7.6 l, respectively. Heavy drinkers defined as those who had more than 50 ml (40 g) or more pure alcohol per day accounted for 6.7% (1674/24 992) of the sample, and consumed 55.3% of the total alcohol consumption.

Unrecorded alcoholic beverage consumption

In China, home-brewed alcohol beverages, tax-free beverages and counterfeited alcohol beverages (unscrupulous merchant-produced illicit alcohol beverages marketed as famous brands), were regarded as unrecorded alcohol beverages. The 3-month use rate and amounts of the unrecorded alcohol beverages consumed were higher in both HY and CD than in the other sites (χ2 = 2779.664, d.f. = 4, P < 0.0001).The 3-month use rate of the unrecorded alcohol beverages was 7.1% (1761/24 992) among all the respondents, the most frequent types of unrecorded alcohol beverages being rice wine and paddy wine. The proportion of unrecorded alcohol beverage accounted for 14.9% of total alcohol consumption, and those in CD and HY sites were 25.5 and 29.9% of each site's consumption, while the other three sites accounted for less than 1% each.

Types of beverages consumed

Preferences expressed by men and women are shown in the Table 4. Beer was the first choice of the all respondents, more than a half of the 1-year drinkers having consumed beer in the past year. Strong distilled spirits and weaker distilled spirits were the second and third choices in men, while grape wine was the second choice in women. Strong distilled spirit, beer and weaker distilled spirit were the main types of beverages consumed in China in the past year, accounting for 35.6, 24.4, 20.5 and 80.1%, respectively, of the overall alcohol consumption.

View this table:
Table 4.

Types of beverages consumed

Men (10 468)Women (4264)Total (14 732)
Strong distilled spirit414739.678718.5493433.5
Less strong distilled spirit288227.53849.0326622.2
Grape wine5044.894422.114489.8
Rice wine10089.659614.0160410.9
Paddy wine2812.71122.63932.7
Yellow wine390.4580.4970.4
Counterfeited beverages and others6366.12656.29016.2

Expenditure in drinking

The average cost of drinking by the 1-year drinkers in the past week was 16.0 (SD 38.6) yuan (1US$ = 8.2¥), which is 3.7% of the average weekly income. Men (mean 16.6¥, SD 39.0) spent more on alcohol beverages than did women (mean 11.6¥, SD 35.0), the difference was statistically significant (t = 4.127, P < 0.0001). As mentioned above, the amount of alcoholic beverages in pure alcohol consumed by male respondents aged 15 years or older was 13.4 times that of females (men 7.55 l; women 0.56 l), but the expenditure on purchase of alcoholic beverages was only 50% higher for men than for women, which resulted from the fact it is women who more often buy alcohol for the whole family.

Health status related to drinking

All respondents across the five areas self-assessed their health status using a structured questionnaire, in which the respondents had five choices: (1) very good; (2) good; (3) fair; (4) bad; (5) very bad. Physical and psychological diseases suffered in the 12 months preceding the interview were also recorded. The respondent rate was 96.1% (24 020/24 992). The results showed that the a significant difference existed between the drinkers and the non-drinkers in terms of self-rated health status, 16.8% of the drinkers and 22.6% of the non-drinkers considered that they were in very good health, and 3.0 % of the drinkers and 7.8 % of the non-drinkers thought that they were in bad or very bad health (χ2 = 297.493, d.f. = 4, P < 0.0001) (see Table 5). Considering influence of age and sex on health status, comparison of self-rating health status between 1-year drinkers and abstainers was stratified by sex and age. The results of self-rating level of health status among sex and age groups were inconsistent; so the sex and age might be confounding factors for health status (see Table 6).

View this table:
Table 5.

Self-rating health status of drinkers and abstainers (%)

Health statusDrinkers (n = 14 434)Abstainers (n = 9586)Total (n = 24 020)
Very good16.822.619.1
Very bad0.10.60.3
View this table:
Table 6.

Comparation of self-rating health status between drinkers and abstainers by age and sex

≤40 years>40 years≤40 years>40 years
Self-rating(n = 4921)(n = 1994)(n = 5291)(n = 1236)(n = 2330)(n = 3400)(n = 1853)(n = 2930)
Very good1419720466112468105674277
Very bad06123302616
Ridit value0.5000.4620.5000.5860.5650.5000.4940.500
95%CI for Ridit0.492∼0.5080.455∼0.4700.492∼0.5080.577∼0.5960.558∼0.5720.490∼0.5090.486∼0.5020.489∼0.511

In order to eliminate the effect of confounding factors, stepwise multivariable regression analysis was used to estimate the contribution of a number of health status variables, in which the dependent variable was the self-rated health status, and the independent variables were gender, occupation, ethnicity, marriage, education, family income, drinking, smoking, annual alcohol consumption and time since the age of first use of alcohol. The entry probability level was 0.05, and the removed probability level was 0.1. Seven variables, including age, marital status, education, gender, time since first alcohol use, drinking and smoking were entered into the regression equation in turn. The results of the regression analysis (see Table 7) showed that age was the most important influential factor and drinking was not a main factor related to health status in the sample.

View this table:
Table 7.

Stepwise multivariable regression in the whole sample

Marital status0.4620.0230.12019.8120.0000.416∼0.507
The time since first use of alcohol0.0050.0010.0808.5940.0000.004∼0.006
  • Time since first use of alcohol, education and family income per person and per month are scale variables; the others are ordinal or nominal, which were coded or recoded as the following. Self-rating of health status wasregarded as a scale variable, coded as follows: 1, very good; 2, good; 3, fair; 4, bad; 5, very bad. Gender: 0, female; 1, male. Age group: 0, ⩽40 years; 1, >40 years. Occupation: 0, general occupation (others); 1, specific occupation (heavy physical labourer; farmer; coal miner; worker in brewery; police). Marital status: 0, stable (single; married); 1, unstable (divorced; separated; widowed; cohabiting). Ethnic: 0, ethnic minority; 1, Han group. Drinking: 0, no; 1, yes. Smoking: 0, no; 1, yes. Annual alcohol consumption: 0, abstainer; 1, moderate drinker; 2, heavy drinker.

Morbidities of drinking-related physical diseases stratified by gender and age

Disc/back pain and gastritis/ulcer were more common in the whole sample, of which 1-year morbidities were 11.1 and 7.9% respectively (Table 8). Considering the effect of gender and age on health status, comparison of the morbidities of drinking-related physical diseases between 1-year drinkers and abstainers stratified by sex and age in the past 12 months were carried out (Table 9). The 1-year morbidity of gastritis/ulcer, migraine, disc/back pain and insomnia was significantly higher in both male drinkers and female drinkers than in male non-drinkers and female non-drinkers, while that of heart disease, cerebral infarction/cerebral haemorrhage, hearing/vision problems and diabetes was significantly lower both in male drinkers and female drinkers than in male non-drinkers and female non-drinkers. These results show that significant differences of morbidities of drinking-related physical diseases exist.

View this table:
Table 8.

The 1-year morbidity of alcohol-related physical and psychological diseases

Males (14 743)Females (10 249)Total (24 992)
Heart disease*4132.84554.48683.5
Cerebral infarction/cerebral haemorrhage**1130.81651.62781.1
Migraine/other headache**5423.72992.98413.4
Disc/back pains**179412.29929.7278611.1
Hearing/vision problems**3732.53723.67453.0
  • ** P < 0.01.

View this table:
Table 9.

Alcohol-related physical and psychological conditions in both drinkers and non-drinkers by sex and age (%)

Males (n = 13 992)Females (n = 11 000)Total (n = 24 992)
ConditionDrinkers (3514)Abstainers (4265)U-valueDrinkers (6735)Abstainers (14 743)U-valueDrinkers (10 249)AbstainersU-value
Heart disease2.54.04.6033.64.72.7792.84.46.811
Cerebral infarction/cerebral haemorrhage0.92.67.6210.41.13.9460.81.65.890
Migraine/other headache2.61.53.7456.33.76.2733.72.93.446
Disc/back pains11.46.38.67514.111.44.17912.29.76.170
Hearing/vision problems2.73.83.3202.23.53.8972.53.65.054
  • U-value = <1.960; P > 0.05; 1.960 < U-value < 2.576; P < 0.05; U-value > 2.276; P < 0.01.

Morbidities of drinking-related physical diseases in the past 1-year by annual alcohol consumption

The results of single variable analysis and multivariable analysis indicated that annual alcohol consumption was a powerful risk factor for overall health status among all the respondents. To describe the morbidities of the health problems related to alcohol use in the general population at different levels of annual alcohol consumption, the 24 992 respondents were divided into eight groups based on the amount of absolute alcohol consumed yearly (Table 10). The results showed that the gastritis/ulcer and sleep problem rates increased with increasing alcohol intake, which had a volume-dependent relationship, and those of heart disease (including coronary heart disease and heart attack) and cerebral infarction/cerebral haemorrhage were significantly higher in non-drinkers and heavy drinkers than in moderate drinkers, which had a V-shaped curve relationship with the annual alcohol intake. Although the positive correlation mentioned above did exist, the results of the chi-squared test of linear trend for ordinal categorical variables on gastritis/ ulcer showed that the relationship between the morbidities of gastritis/ulcer and annual alcohol consumption was non-linear, while no significant correlation between other alcohol-related physical diseases and annual alcohol consumption could be found.

View this table:
Table 10.

Morbidity of disease (%) by annual alcohol consumption (l/year)

DiseaseAbstainer≤0.50.5∼2.0∼4.0∼8.0∼16.0∼>32.0χ2 regressionχ2 partialχ2
Gastritis or ulcer5.**23.39398.45**
Heart diseases4.**1.1069.57**
Cerebral infarction/haemorrhage1.**0.1742.49**
Back pains/disc problems9.711.810.29.912.413.016.814.295.96**4.9591.01**
Hearing/vision problems3.**0.4230.94**
Sleep problems2.**0.8742.42**
  • The chi-squared test of linear trend for ordinal categorical variable was used.

  • ** P < 0.01.

Prevalence rates of alcohol-related psychiatric disorders

A total of 1874 problem drinkers were screened out of 24 992 community household residents aged 15 years or older in five areas in China by APST, and 1284 individuals were diagnosed as having mental disorders based on DSM-III-R criteria. The percentage of 52.1% of acute intoxication caused by alcohol in the problem drinkers was significantly higher than that (8.5%) in the non-problem drinkers (χ2 = 5644.204, P < 0.0001). Table 11 shows the major diagnoses of alcohol-related mental disorders. The prevalence rates of dependence, abuse, dementia, amnesia and personality disorder were given a point prevalence, that of acute intoxication had a 3-month prevalence and that of uncomplicated withdrawal and withdrawal delirium had a 1-year prevalence. The overall point-prevalence rate of mental disorders (if a respondent was diagnosed as having more than one alcohol-related mental disorders, he or she would still be calculated as one case) caused by alcohol was 5.1% (9.0% in men, 0.2% in women), the prevalence rate of alcohol dependence was 3.8% (6.6% in men, 0.2% in women) and the 3-month rate of acute alcohol intoxication was 8.3% (14.2% in men, 0.7% in women). In general, male drinkers had more problems than did their female counterparts.

View this table:
Table 11.

Alcohol-related disorders (DSM-III-R) by sex

Males (13 992)Females (11 000)Total (24 992)
Acute intoxication199214.237820.74520748.299
Uncomplicated withdrawal3082.20120.0183101.240
Withdrawal delirium70.05070.028
Personality disorder600.42910.009610.240


For economic and cultural reasons, the total amount of alcohol production and consumption and the prevalence of alcohol-related physical and mental diseases was relatively low before the end of the 1970s. However, with the rapid development of the economy, urbanization and westernization, alcohol production and consumption and numbers of admitted patients with alcohol-related physical and mental diseases have increased steadily over the past 25 years (Hao et al., 1995). Based on WHO data in Global Trends on Alcohol (World Health Organization, 1999), the per capita alcohol consumption in pure alcohol for adults in China in 1970 was 1.03 l, and rose to 5.17 l in 1996. A nation-wide epidemiological survey (Hao et al., 1998a) of the general population in six areas of China in 1993 indicated that the annual alcohol consumption among respondents aged 15 years or above was 3.6 l. Retrospective studies (Chen and Huang, 1995; Zhou et al., 1999) showed that the proportion of patients admitted with mental disorders related to alcohol (in terms of total psychiatric inpatients) increased almost 10-fold during 1965–1997 in Shanghai Mental Health Center, Guangzhou Psychiatric Hospital, and Nanjing Brain Hospital. In recent years, significant changes in ideology and lifestyle among Chinese people have taken place, and we wanted to know what has occurred in relation to alcohol use and alcohol-related problems.

In this study, 24 992 respondents in five areas of China were interviewed by using a questionnaire on drinking behaviour and health status related to drinking provided by WHO. The results showed the 1-year drinking rate in the whole sample was 59.0%, which was almost equivalent to the survey done by authors in same areas in 1993 (Hao et al., 1998a). The 1-year drinking rate in men was 74.9%, a decrease of 10% compared with results obtained in 1993. The drinking rate in women was 38.8%, an increase of almost 10% compared with the 1993 results. The ratios of male-to-female drinking rates and annual alcohol consumption in pure alcohol were 1.9:1.0 and 13.4:1.0, respectively. The figures verified the prediction that the numbers of women drinking would increase in the 21st century, as in many other developed countries (Hao et al., 1995). The women's liberation movement, changes of women's role in society and increasing of numbers of professional women in China are factors for explaining the reason for the increase in the female drinking rate. However, the ratio male-to-female drinking rates was still broader than that in industrial countries (Martin and Hubbard, 2000). Since the Family Planning Policy was adopted in China, the number of single-child families has increased, children from these families have viewpoints (including those concerning alcohol drinking) that differ from those who were born before the Cultural Revolution (1966–1976). As China enters into the World Trade Organization (WTO) and there is even more rapid development in China's economy, the authors predict that the drinking rate in women will continue to increase, and that the ratio of male-to-female drinking will narrow in the coming decade.

Per capita alcohol consumption in pure alcohol for adults is an essential predictor of alcohol-related problems. Based on WHO data, the growth rate of per capita alcohol consumption was 402% from 1970 to 1996 (World Health Organization, 1999). In this study, the average annual alcohol consumption for individuals aged 15 years or above was 4.47 l; an increase of 0.87 l has taken place over the past 8 years, as indicated by the results of the 1993 survey (Hao et al., 1998a).

In our survey, heavy drinkers were defined as those who used 50 ml or more pure alcohol per day. The results showed that 6.7% (1674/24 992) of the whole sample were heavy drinkers, who consumed 55.3% of the total alcohol consumption. It suggests that the heavy drinkers were the risk group for alcohol-related problems in the selected areas. The figure for average annual alcohol consumption was still low compared with that of the developed countries (World Health Organization, 1999), which was about 10.0 l yearly. For example, the recorded per capita consumption of pure alcohol per adult 15 years of age and over in 1996 was 11.90 l in Austria, 11.67 l in Germany, 11.27 l in Switzerland, 9.62 l in Italy, 9.55 l in Australia, 9.41 l in the UK and 8.90 l in the US. The current global trends on alcohol use were that per capital alcohol consumption in developed countries was decreasing sharply, and increasing steadily in developing countries. However, it is difficult to predict what the peak level of alcohol consumption is and when it will be reached in China. Thus, it is necessary to monitor the changes in the future in China.

The survey on unrecorded alcohol beverages could provide data with which to estimate the alcohol consumption in regions or nations. In the countries of the former Soviet Union and in many developing countries, alcohol production for home use or for the informed sector is extremely important, being as high as 80% of the total alcohol available for consumption. Reliable data exist regarding consumption of these forms of alcohol in more than 20 countries (World Health Organization, 1999). In China, people usually get alcoholic beverages from shop, restaurants or bars. In this study, counterfeit alcoholic beverages and home- and privately-brewed alcoholic beverages are defined as unrecorded. The results showed that 7.1% of respondents reported that they had used between one and three types of unrecorded alcoholic beverage in the 3 months prior to the interview, the amount of unrecorded alcoholic beverage (in pure alcohol) accounted for 14.9% of overall alcohol consumption in the five areas studied. However, the amount of unrecorded alcoholic beverage consumed varied across the five survey areas. For example, the rates of unrecorded alcoholic beverage use in HY and CD were 24.1 and 9.7%, respectively; the amounts of these beverages as a proportion of the overall alcohol consumption were 29.9 and 25.5%, respectively. However, the number of unrecorded alcoholic beverage users and the volume of beverage consumed was insignificant in other three survey areas. The differences in unrecorded alcoholic beverage consumption among the survey areas are related to specific drinking customs in HY and CD, where the population, especially those in rural areas, traditionally use home-brewed alcohol. We can conclude that unrecorded alcoholic beverages play an important role in the alcohol consumption in certain areas in Southern China. However, the survey is not representative of the total national population, because China is a large country with various drinking and brewing customs. It is necessary to carry out further surveys in more areas to reveal the true unrecorded alcoholic beverage consumption in China.

Alcohol use is related to a wide range of physical, mental and social harms. Most health professionals now agree that practically no organ in the body is immune to alcohol-related harm (Bower, 1992). A number of conditions have been identified that are wholly caused by alcohol use, and other conditions have been identified that are partly caused by alcohol use. Over the last two decades, many researches have found a decrease in all-cause mortality among certain light-to-moderate drinkers of alcoholic beverages compared to non-drinkers and heavy drinkers; light-to-moderate drinking of alcohol apparently being protective against heart disease.

The relationship between health level and alcohol consumption of the population as a whole is not yet clear. Our survey was designed to determine health status related to drinking and incidences of physical and psychological conditions in drinkers and non-drinkers in five areas of China and to examine the causality relationship between alcohol drinking and the occurrence of these conditions. The single variable analysis in this study showed that health status in drinkers was different from that in non-drinkers, and that age and gender were confounding factors to health status. The results of stepwise multivariable regression on self-rated health status revealed that some sociodemographic characteristics, including age, marital status, education and gender, were less important than were time since first use of alcohol and smoking. The results suggest that alcohol consumption has a slight association to health and that the health status of the general population was determined mainly by many risk factors, though drinking and smoking were primary factors.

Alcohol-related mental disorders were considered to be caused wholly by alcohol drinking. In this study, 5.1% of the general population living in five areas of China in the past year met the criteria for a diagnosis of alcohol-related mental disorder in DSM-III-R, which was lower than that of the national probability sample in the US in 1994 (World Health Organization, 1999). The point prevalence rate of alcohol dependence was 3.8%, with an increase of 0.4% compared with the survey conducted in the same areas in 1993 (Hao, 1998b), and the 3-month prevalence rate of acute intoxication was 8.3%, which was higher than that of the survey in 1993. Although the rate of alcohol-related mental problems in China rose during 1993–2001, it is relatively low compared with that of developed countries.

This study also indicates that the incidence of certain physical diseases, such as gastritis/ulcer and insomnia, increased with increased alcohol consumption. The influence of alcohol on the heart is a separate issue. Numerous studies have suggested that light-to-moderate alcohol consumption reduces the risk of heart disease (Renaud et al., 1993; National Institute on Alcohol Abuse and Alcoholism, 1999; Murray et al., 2002). The result of our survey was that morbidity from heart disease (including coronary heart disease and heart attack), cerebral infarction and cerebral haemorrhage was higher both in non-drinkers and in heavy drinkers than in light-to-moderate drinkers, suggesting the protective effect of alcohol against cardiovascular disease existed in the studied population, and was consistent with that of an epidemiological survey in the US (Klatsky, 1994). However, the results of the chi-squared test of linear trend for ordinal categorical variables suggested the 1-year morbidities of alcohol-related physical diseases except gastritis/ulcer had no association with alcohol consumption.

Traditionally, Chinese people believe that moderate drinking has a good effect on health and Traditional Chinese Medicine theory says that ‘alcohol is the leader of all kinds of medicine’. There are some Chinese alcoholic beverage medicines on the market for the treatment of health problems such as back and leg pain caused by rheumatism and sexual problems (e.g. impotence). In a survey sample of 2064 individuals aged 18 years or older performed in Xi'an rural areas in 1998, 6.1% drank alcoholic beverages for health reasons (Tao et al., 2002). However, it is not clear whether drinkers in this survey were more likely to suffer from disc problems and back pains than non-drinkers. Further investigation is required for clarification.

In general, the association between health and alcohol drinking is not simple. Alcohol consumption plays a role in the development of so-called alcohol-related physical diseases. There is no doubt that other factors, such as lifestyle and environment, also impact on health status. The authors therefore conclude that the health status of the individuals aged 15 years or older in the selected areas of China studied resulted from the integrative effects of many risk factors.


This study was supported by the World Health Organization special fund: WHO Project Pilot Survey on Unrecorded Consumption of Alcohol in China in 2001.


View Abstract