Mental Health and Substance Abuse Division of the Health and Development Group and the South African Cochrane Centre, Medical Research Council (Cape Town), PO Box 19070, Tygerberg, 7505, South Africa
Received 21 August 2000; in revised form 14 November 2000; accepted 18 December 2000
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Self-report of drinking is of doubtful accuracy in surveys as heavy drinkers tend to under-report (Conigrave et al., 1995; Wetterling et al., 1998
). In addition, a comprehensive drinking history usually requires a trained interviewer and can be time-consuming, further limiting its usefulness. The development and evaluation of brief screening instruments, such as the four-item CAGE questionnaire, has helped in this area (Mayfield et al., 1974
). The CAGE has good internal reliability and has been used widely across the world (Dawe and Mattick, 1997
). However, the validity of the CAGE in different cultural and language settings has been questioned (Indran, 1992
). Many researchers fail to recognize that the CAGE reflects lifetime drinking problems, rather than recent alcohol consumption, and is more sensitive to alcohol dependence than to less severe drinking disorders (Fiellin et al., 2000
). This has important implications for studies that compare it with recent quantityfrequency consumption reports and biological markers.
Accurate and reliable biological markers to identify both those who are alcohol dependent and those who are at risk for harmful or hazardous drinking in a community are highly desirable (Aithal et al., 1998). Mean corpuscular volume (MCV) and gamma-glutamyl transferase (GGT) have been used extensively to measure long-term alcohol use, but elevated levels do not always indicate alcohol use (Sillanaukee, 1996
). Over the past decade, carbohydrate-deficient transferrin (CDT) has been proposed as the most reliable marker of chronic alcohol consumption (Stibler, 1991
; Mikkelsen et al., 1998
; De Feo et al., 1999
). Studies have shown that CDT is a better marker for alcohol dependence than for actual consumption levels (Mikkelsen et al., 1998
; Mundle et al., 1999
). There is some evidence that CDT is not a useful marker in females (Gronbaek et al., 1995
; Huseby et al., 1997
), although investigators of these studies acknowledged that the low numbers of participating women could have affected the accuracy of their results. Mundle et al. (1999) found that CDT was significantly more sensitive in subjects who had a duration of abstinence of less than 4 days before blood testing, than those with a longer duration of abstinence. Recent research suggests that low iron status or high iron demand may elevate CDT levels (Van Pelt and Azimi, 1998
; De Feo et al., 1999
). The influence of age on the test is currently equivocal (Huseby et al., 1997
; Agelink et al., 1998
). To date, most studies using CDT have been conducted in clinical settings, with lower sensitivities reported in general population settings (Gronbaek et al., 1995
; Huseby et al., 1997
). Findings conducted in the developed world may not be generalizable to rural communities living in the developing world and little is known about possible racial variation of CDT levels (Conigrave et al., 1995
; Wickramasinghe et al., 1995
).
Our aim in this paper is three-fold: (1) to profile the drinking behaviour of a rural Lesotho community; (2) to compare the usefulness of the following measures of alcohol misuse in this community: self-report of quantity and frequency, the CAGE questionnaire and CDT levels; (3) to describe community awareness of, and attitude towards, treatment services.
Background to the Lesotho Highlands Water Project
Lesotho is a small mountainous country, entirely surrounded by South Africa. From 1884 until independence in 1966 Lesotho was a British protectorate (Mphi, 1994). It is populated by just over 2 million Basotho people, many of whom maintain a traditional agricultural lifestyle in the mountains. The Lesotho Highlands Water Project (LHWP) is a massive hydro-electric scheme established to generate electric power for Lesotho and for the sale of water to South Africa. It is jointly funded by South Africa, Lesotho, the European Investment Bank and the World Bank, and has involved the building of dams, bridges, roads and tunnels. This has required the relocation of several communities due to the inundation of their villages and surrounding areas by water.
During the construction of a dam in the Mohale area of Lesotho, the Lesotho Highland Development Authority commissioned a health impact assessment to provide baseline information on the living conditions and health of the communities to be affected by the construction programme, both for the purposes of planning for future health and welfare needs of the communities and for monitoring the health impacts of the construction programme. The Medical Research Council (MRC) of South Africa and its partners were contracted to undertake the baseline assessment of the health of the residents living in and around the environs of the proposed dam.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The field work was contracted to a private research organization, Sechaba Consultants. Two research teams, each comprising a team leader, a public health nurse, a research assistant and a community health worker where possible, conducted face-to-face interviews with members of each household sampled. All the field workers had previous data collection experience and had received training in conducting interviews from Sechaba Consultants and the MRC. They were all Basotho.
Consent for household participation in the study was requested from the household head or principal caregiver. After the household head or caregiver had given permission for the study to proceed, formal written consent was obtained from each adult (aged over 16 years) who consented to be interviewed. Therefore, in some households more than one adult participated. The face-to-face interview covered a broad range of health issues and included questions about alcohol use and misuse.
The alcohol section of the interview was developed by two of the authors (C.P. and N.M.), translated into Sesotho and back-translated into English. Twenty questions focused specifically on alcohol and included whether the person drank alcohol, the nature of his or her drinking (e.g. alone or as part of a celebration), the last time alcohol was drunk, the time of day alcohol is typically drunk, the number of days per week that the participant typically drinks, the drinking history during the past 7 days, the type and amount of alcohol drunk on a typical drinking day, and the length of time spent drinking on a typical drinking day. As it was anticipated that many participants would describe drinking traditional home-brewed beer which is packaged in a variety of containers of different volumes, provision was made for participants to specify the exact containers used.
The four questions of the CAGE questionnaire were included. Additional structured questions asked participants which treatment services they would access should they have an alcohol problem (a menu of available services was provided), and if so, how accessible these services were; whether they would be able to afford treatment, and whether they believed treatment would be successful.
Definition of hazardous drinking
Using the quantity and frequency information obtained during the interview, we defined a person with hazardous drinking as all those consuming more than 350 g (males) or 225 g (females) of alcohol/week, in accordance with previous studies (Bohn et al., 1995; Wetterling et al., 1998
). Participants who indicated that they engaged in bouts of heavy drinking 12 days a month or more during the past 12 months were also included in the definition of hazardous drinkers. This definition allowed for the inclusion of people who were likely to be alcohol dependent and those who engaged in early-stage harmful or hazardous drinking (Bohn et al., 1995
). It did not differentiate between them.
Procedures for CDT collection and analysis
Each adult who consented had three 5 ml vials of blood taken for measurement of syphilis, hepatitis B and HIV status, and nutritional indicators which included CDT. The tubes of clotted blood were stored at 4°C and transported daily to the field station where the tubes were centrifuged at 2500 r.p.m. for 10 min and the serum decanted into labelled 5 ml plastic tubes and stored at 20°C. The samples were transported twice weekly to Maseru, the capital of Lesotho, before being transported to the South African Institute for Medical Research in Johannesburg for analysis. Kabi Pharmacia sponsored the kit, CDTect®. The cut-off points recommended by the manufacturer were used: for males >20 U/l; females >26 U/l.
Statistical analysis
Statistical analysis was performed using an SPSS program package. Associations between categorical variables were examined using 2-tests. Dimensional variables were analysed using t-test (two-tailed) when appropriate. Alpha was set at 0.05.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Descriptive details of the participants can be seen in Table 1. Eighty per cent (279/348) were women and 76% (266/348) were aged from 16 to 40 years. The majority (80%) were married. The most frequently described occupation was household work (61%), followed by brewing beer (8%) and agriculture (7.5%).
|
Of those who drank alcohol, almost half (41/87) drank twice a month or less often, 29% (25/87) drank on 12 days a week and 24% (21/87) drank on 3 days a week or more often, with 10% (9/87) drinking every day. Sixty-five per cent of the sample reported drinking traditional home-brewed beer exclusively, 15% drinking Western beer exclusively and 20% drinking both Western and traditional beer (data available from 72 cases). One person reported drinking spirits and one person cider as a second drink of choice after Western beer.
The daily volumes of traditional beer that were consumed on a typical drinking day ranged from 340 to 10 000 ml with a mean (SD) of 2634 (± 1737 ml). This was not significantly different for men and women (t = 1.72; df = 64; P = 0.09). Daily volumes of Western beer consumed ranged from 750 ml to 9000 ml with a mean of 3377 (± 2310 ml). There was no significant sex difference (t = 1.90; df = 56; P = 0.06).
Hazardous drinking
According to the definition of a hazardous drinker described in the Methods section, 25 males and 25 females were classified as hazardous drinkers (36% of all men and 9% of all women). Men were significantly more likely to be classified as hazardous drinkers than women (2 = 32.45; df = 1; P = 0.000). The hazardous drinkers were significantly older (37.2 ± 7.9 years) than the rest of the sample (31.5 ± 8.8 years) (t = 4.17; df = 341; P = 0.000). There was no significant difference in marital status between the hazardous drinkers and the rest of the sample (
2 = 0.45; df = 1; P = 0.50). Four of the 27 participants who described brewing beer as their main occupation were classified as hazardous drinkers.
CAGE questionnaire
Fifty-five per cent of the 78 drinkers on whom CAGE data were available scored above the commonly recommended cut-off value for the CAGE (score 2). Of those who were classified as hazardous drinkers, 67% (33/49) scored above the cut-off score for the CAGE. Of those drinkers who were not classified as hazardous drinkers, 58% (17/29) also scored above the cut-off score. The sensitivity, specificity and positive predictive values for the CAGE in the sample of drinkers were calculated, with hazardous drinking being used as the standard (Table 2
).
|
We assumed that all those who reported not drinking and who had CDT levels assessed would report amount and frequency scores equal to zero and therefore not be classified as hazardous drinkers. We then calculated the sensitivity, specificity and positive predictive value of CDT on the sample of participants who had CDT levels assessed (n = 181). These can be seen in Table 3, where hazardous drinking is the standard.
|
Half of the participants (179/347) reported that reaching treatment facilities would be difficult or very difficult for them. Thirty-six per cent (127/346) of the participants would require longer than 4 h to reach their destination, with four of these requiring more than 24 h travelling time. Forty-five per cent of all participants required 2 h or less to reach their preferred treatment service. There was no significant difference between hazardous drinkers and the rest of the sample in reporting access to treatment.
Approximately two-thirds (226/344) of all participants stated that they would be able to afford treatment for an alcohol problem, 15% would not be able to afford it and 20% were uncertain. The majority (77%; 266/344) believed that the treatment services would be successful in their treatment, 3% did not, and 19% were uncertain. Hazardous drinkers did not differ significantly from the rest of the sample in reporting affordability of treatment or treatment success.
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The sample was biased towards female participants. This was to be expected, as men in Lesotho tend to migrate to South Africa for long periods of time in search of better paid jobs, leaving the women to shoulder the major family and economic responsibilities at home (Mphi, 1994). Although this limits our interpretation of the findings for males, it does provide a comprehensive picture of the experience of women living in rural communities. Most of the participants were younger than 40 years, reflecting the typically higher mortality and fertility of a developing country population. It is notable that after household work, the most common occupation described is the brewing of beer. Until white farmers and settlers arrived in this part of Southern Africa in the early 19th century (Meursing and Morojele, 1989
) low-alcohol sorghum beer was the only alcoholic beverage available to Basotho people. Consumption of this brew was restricted to adults, governed by strict behaviour norms, and drunk only on special occasions. According to local culture, women are discouraged from drinking. However, 19% of the women in our study admitted to drinking alcohol. More consistent with traditional norms was our finding that 58% of the women last drank at a feast or celebration, compared with only 20% of the male drinkers.
The finding that 75% of the male drinkers reported last drinking on no particular occasion confirms the rapid shift away from traditional drinking patterns described by Meursing and Morojele (1989). They ascribe this change largely to aggressive and widespread advertising of alcohol and an expanding range of available alcohol products. However, traditional beer was the drink of preference for most participants in our sample, reflecting the rural setting and high costs and limited availability of Western alcohol products. A greater median volume of traditional beer was consumed than Western beer. This difference in volumes consumed can be explained by the difference in alcohol content, with traditional beer having ~3% alcohol content and Western beer having 5% alcohol content (Wolmarans et al., 1993). As the alcohol content of traditional beer varies from brew to brew, and brewer to brewer, quantifying the actual alcohol content per volume can only be approximate. We deliberately chose a conservative definition to quantify hazardous drinking to minimize over-estimating hazardous drinking due to this variability.
Prevalence of hazardous drinkers among our sample was 9% for women and 36% for men. This difference in rates between men and women is consistent with other Southern African studies of risky drinking, which show alcohol consumption rates to be 1520% higher for men than women (Parry and Bennetts, 1998). Currently, there are few comparative data available for patterns of drinking among adults in rural communities, as most studies have been conducted in urban or peri-urban areas and many focus on young people of school-going age. Despite the lack of comparative data, we believe the prevalence patterns found in our study are alarmingly high. This high rate may be due to stressors impacting on the community, as a result of imminent relocation and disruption of traditional lifestyles. Hazardous drinkers were more likely to be male and older than other participants. This is again consistent with what would be expected in a community where younger, able-bodied men leave to seek work elsewhere.
The sensitivity of the CAGE to detect hazardous drinking was poor in men, but high in women. However, this was reversed for specificity. As drinking is culturally unacceptable for women, the high proportion of false-positives in women may be because they are more likely to under-report their consumption and/or because they are more likely to answer yes to the CAGE question: Have you ever felt guilty about your drinking? The positive predictive value (PPV) of the CAGE was lower for women (62%) than for men (75%). With hazardous consumption as the gold-standard, the CAGE is not appropriate for screening women in this setting.
The sensitivity and specificity recorded for CDT must be interpreted with caution, due to the low numbers of males in the sample. However, the high specificities for both men (100%; 19/19) and women (77%; 110/142) is consistent with other studies (Aithal et al., 1998; Wetterling et al., 1998
). The low PPV of 9% (3/35) for women and 14% (5/37) for men and women combined suggests that CDT is not effective as a predictor of hazardous drinking in this population. However, the cut-off for drinking, especially for women's drinking, was lower than the usual cut-off in published CDT studies.
The fact that 20% (28/141) of those who stated that they never drank alcohol at all had elevated CDT levels is of interest. All of these were women. Although CDT has been shown to be a poor measure of alcohol consumption in women (Gronbaek et al., 1995; Huseby et al., 1997
), the following reasons may be especially relevant in our sample. Firstly, depleted iron stores may raise CDT levels (Van Pelt and Azimi, 1998
; De Feo et al., 1999
). In deprived rural communities with poor nutritional status, such as the communities in this study, many women of child-bearing age may be iron-deficient with resultant raised CDT levels. Unfortunately, we cannot give CDT as a percentage of total transferrin, which might have clarified this point, because total transferrin was not measured. Secondly, as this is the first study to measure CDT levels in a black African population, the cut-off points for CDT may be inaccurate and further research is required to determine appropriate levels.
We are encouraged by the finding that there was a relatively high awareness of available treatment services specific to alcohol problems. Even more encouraging is the finding that most participants believed they would not only be able to afford treatment, but that treatment would be successful. Although these findings cast the Lesotho health system in a positive light, it is a matter of concern that 36% of participants require longer than 4 h to reach these treatment services. This is largely due to the surrounding rugged terrain and lack of transport services in the area. Based on the fieldworkers' knowledge and experience, it is our opinion that it would take most participants a day to reach the facilities and that treatment at the designated alcohol treatment facilities would be too expensive for most. The construction of roads as a result of the LHWP will potentially improve the villagers' access to treatment services.
Throughout Africa and the developing world, the traditional lifestyles of many communities are threatened by construction programmes similar to the LHWP. Our study achieved its goal of providing a baseline profile of the drinking behaviour of a rural Lesotho community facing such forced lifestyle changes. Alcohol use patterns can provide an indication of the underlying health of a community. It will be important to monitor the drinking patterns of adults in this community over time, to determine if changes occur in response to relocation and the disruption of traditional lifestyles.
![]() |
ACKNOWLEDGEMENTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
FOOTNOTES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Aithal, G. P., Thornes, H., Dwarakanath, A. D. and Tanner, A. R. (1998) Measurement of carbohydrate-deficient transferrin (CDT) in a general medical clinic: is this test useful in assessing alcohol consumption? Alcohol and Alcoholism 33, 304309.[Abstract]
Bohn, M. J., Babor, T. F. and Kranzler, H. R. (1995) The Alcohol Use Disorders Identification Test (AUDIT): Validation of a screening instrument for use in medical settings. Journal of Studies in Alcohol 56, 423432.[ISI][Medline]
Conigrave, K. M., Saunders, J. B. and Whitfield, J. B. (1995) Diagnostic tests for alcohol consumption. Alcohol and Alcoholism 30, 1326.[Abstract]
Dawe, S. and Mattick, R. P. (1997) Review of Diagnostic Screening Instruments for Alcohol and Other Drug Use and Other Psychiatric Disorders. Australian Government Publishing Service, Canberra.
De Feo, T. M., Duca, L., Mattioli, M., Cappellini, M. D., Sampietro, M., Cesana, B. M. and Fiorelli, G. (1999) Carbohydrate-deficient transferrin, a sensitive marker of chronic alcohol abuse, is highly influenced by body iron. Hepatology 29, 658663.[ISI][Medline]
Fiellin, D. A., Reid, M. C. and O'Connor, P. G. (2000) Screening for alcohol problems. Archives of Internal Medicine 160, 19771989.
Gronbaek, M., Henriksen, J. H. and Becker, U. (1995) Carbohydrate-deficient transferrin a valid marker of alcoholism in population studies? Results for the Copenhagen City Heart Study. Alcoholism: Clinical and Experimental Research 19, 457461.[ISI][Medline]
Huseby, N.-E., Nilssen, O. and Kanitz, R.-D. (1997) Evaluation of two biological markers combined as a parameter of alcohol dependency. Alcohol and Alcoholism 32, 731737.[Abstract]
Indran, S. K. (1992) Quantity frequency (consumption index) versus CAGE in the detection of alcoholism. Australian and New Zealand Journal of Psychiatry 26, 493501.[ISI][Medline]
Mayfield, D., McLeod, G. and Hall, P. (1974) The CAGE questionnaire: validation of a new alcoholism screening instrument. American Journal of Psychiatry 131, 11211123.[ISI][Medline]
Meursing, K. and Morojele, N. (1989) Use of alcohol among high school students in Lesotho. British Journal of Addiction 84, 13371342.[ISI][Medline]
Mikkelsen, I.-M., Kanitz, R.-D., Nilssen, O. and Huseby, N.-E. (1998) Carbohydrate-deficient transferrin: marker of actual alcohol consumption or chronic alcohol misuse? Alcohol and Alcoholism 33, 646650.[Abstract]
Mphi, M. (1994) Female alcoholism problems in Lesotho. Addiction 89, 945949.[ISI][Medline]
Mundle, G., Ackermann, K., Munkes, J., Steinle, D. and Mann, K. (1999) Influence of age, alcohol consumption and abstinence on the sensitivity of carbohydrate-deficient transferrin, gamma-glutamyl transferase and mean corpuscular volume. Alcohol and Alcoholism 34, 760766.
Parry, C. D. H. and Bennetts, A. L. (1998) Alcohol Policy and Public Health in South Africa. Oxford University Press, Cape Town.
Sillanaukee, P. (1996) Laboratory markers of alcohol abuse. Alcohol and Alcoholism 31, 613616.[Abstract]
Stibler, H. (1991) Carbohydrate-deficient transferrin in serum: a new marker of potentially harmful alcohol consumption reviewed. Clinical Chemistry 37, 20292037.
Van Pelt, J. and Azimi, H. (1998) False-positive CDTect values in patients with low ferritin levels. Clinical Chemistry 44, 22192220.
Wetterling, T., Kanitz, R.-D., Rumpf, H.-J., Hapke, U. and Fischer, D. (1998) Comparison of CAGE and MAST with the alcohol markers CDT, GGT, ALAT, ASAT and MCV. Alcohol and Alcoholism 33, 424430.[Abstract]
Wickramasinghe, S. N., Corridan, B., Izaguirre, J., Hasan, R. and Marjot, D. H. (1995) Ethnic differences in the biological consequences of alcohol abuse: a comparison between South Asian and European males. Alcohol and Alcoholism 30, 675680.[Abstract]
Wolmarans, P., Langenhoven, M. and Faber, M. (1993) Food, Facts and Figures. Oxford University Press, Cape Town.