TITLE-6617887
Addiction (1993) 88, 1627-1636 RESEARCH REPORT The Structure and correlates of alcohol dependence: WHO collaborative project on the early detection of persons with harmful alcohol consumption--III WAYNE HALL,'-" JOHN B. SAUNDERS,' THOMAS F. BABOR/ OLAF G. AASLAND," ARVID AMUNDSEN,' RAY HODGSON* & MARCUS GRANT' 'Centre for Drug and Alcohol Studies, Royal Prince Alfred Hospital, and Departments af Medicine and Psychiatry, University of Sydney, Syd?iey, Australia, ^National Drug and Alcohol Research Centre, University of New South Wales, P. O. Box 1, Kensington, New South Wales, Australia, ^Department of Psychiatry, University of Connecticut, School of Medicine, Farmington, Connecticut, USA, ^Ministry of Health and Social Services, Oslo, Norway, ^National Institute for Alcohol and Drug Research, Oslo, Norway, ^District Department of Clinical Psychology, Whitchurch Hospital, Whitchurch, Cardiff, Wales, and University of Wales & ^Programme on Substance Abuse, World Health Organization, Geneva, Switzerland Abstract The cross-cultural validity of the Alcohol Dependence Syndrome was tested on 13 symptoms of alcohol dependence which were assessed as part of a WHO collaborative study of the early detection of harmful drinking. The subjects were drinking patients in health care settings in Australia, Bulgaria, Kenya, Mexico, Norway, and the US. Principal Components Analyses were performed on the symptoms in each centre, and the degree of agreement between the results was assessed by calculating coefficients of congruence between the item loadings on the first principal component. In aU six centres the first Principal Component accounted for at least half of the total variance and all symptoms had positive loadings greater than 0.40 on the first Principal Component. The coefficients of congruence were all 0.98 or more, and the 13 symptoms had internal consistency coefficients af 0.94 or more. An alcohol dependence score defined by the sum of positive responses to the 13 alcohol dependence symptoms was positively correlated with self-reported alcohol consumption, alcohol-relaud problems, serum gamma glutamyltransferase and a clinical examination assessment of alcoholism in all six samples. Introduction originally devised by Edwards & Gross (1976) The Alcohol Dependence Syndrome which was has subsequently been incorporated in modified -- form in DSM-III-R and ICD-IO. It identifies a Correspondence to: Wayne Hall, National Dmg ^ d AJcohol ^^ ^. symptoms, and behaviours as Research Centre, tJniversity of New South Wales, P.O. Box I, T f i.- i_ ? i i . i Kensington, New South Wales 2033, Australia. indicators of a s y n d r o m e w h i c h is likely t o 1627 1628 Wceyne Hall et al. develop in persons with a history of chronic heavy alcohol consumption, and which has been conceptualized as a psychobiological condition that drives drinking (e.g. Babor, 1992). As originally conceived by Edwards & Gross, the Alcohol dependence Syndrome comprised some or all of the following symptoms: a narrowing of drinking repenoire, an increased salience of alco- hol, a subjective compulsion to drink, increased tolerance to alcohol, withdrawal symptoms, drinking to reheve withdrawal, and rapid rein- statement of the syndrome after a period of abstinence. The concept of the Alcobo! Dependence Syn- drome developed partly in response to criticisms of the hypothesis that alcoholism was a specific disease entity which was qualitatively distin- guishable from heavy drinking Qellinek, 1960). Many critics of Jellinek's disease model argued that the evidence from population and clinical studies indicated tbat there was a continuous distribution of alcohol dependence symptoms at the extreme of whicb clinically diagnosed "alcoholics" were to be found. The Alcohol Dependence Syndrome retained the hypothesis that there was a grouping of behaviours which cohered suffidently to constitute a syndrome of alcohol dependence, while recognizing that alcohol dependence could be present to varying degrees (Edwards, 1986). Among the evidence offered in support of the syndrome were the results of factor analyses of self-report measures of behavioural symptoms within eacb of tbe conceptual domains of the syndrome. Although different studies measia-ed these behaviours in different ways, and very few measured all seven conceptual domains (see Caetano, 1990 for an exception), their results generally supported the validity of the Alcohol Dependence Syndrome (e.g. Stockwell et al., 1979; Polich, Armour & Braker, 1980; Skinner, 1981; Skinner & Allen, 1982; Babor, Lauerman & Cooney, 1987; Davidson, Bunting & Raistruck, 1989). The majority of such studies have typically reponed that a single factor (usu- ally defined by principal components analysis) explained a substantial part of tbe common vari- ance in self-reported dependence symptoms in clinical samples, and that the majority of symp- toms had substantial positive loadings on that fector. Chick (1980) failed to obtain a single dimension from a Principal Components Analy- sis of dependence symptoms for reasons that are tmclear, although even in this study there was a reasonable degree of coherence among a subset of the dependence symptoms. The validity of the Alcohol Dependence Syndrome bas not been universally accepted (e.g. Room, 1980, 1983; Shaw, 1980; Heather & Robertson, 1990). Sociological critics con- tend that it is a culturally specific expression of alcohol-related problems that arises in Anglo- Saxon societies in which the disease model of the Alcohohcs Anonymous movement has been influential. On this alternative view, the alcohol dependence syndrome is a ctilturally based explanation of the persistence of heavy drinking in the face of adverse consequences, the acceptance of which by many problem drinkers in treatment samples affects tbeir responses to self-report symptom inventories. The apparent coherence of the symptoms of tbe Alcohol Dependence Syndrome, on this hypothesis, is the result of response bias. This hypothesis is difficvJt to refute since most of the evidence in flavour of the syndrome has come from studies of clinical samples in Britain, Canada, and the US. Only one study by Babor and his colleagues has looked at the cross-cul- tural generalisabihty of the Alcohol Dependence Syndrome in cUnical samples in the US and France with results that supported its cross- cultural validity (Babor, Lauerman & Cooney, 1987). The present paper tested this sceptical view of the Alcohol Dependence Syndrome on symp- toms of alcohol dependence reported by drinking patients in primary health care samples in six countries as part of the WHO Collaborative Pro- ject on Early Detection of Harmful Alcohol Consumption (Saunders et al., 1990, 1991). In this study a standardized set of instruments was used to collect data on alcohol consumption, symptoms of dependence, and alcohol-related problems in each of six countries: Australia, Bul- garia, Kenya, Mexico, Norway and the US. These samples represented cultures that differed substantially in the proportion of their popula- tions who consumed alcohol, in their drinking practices, and in the range and nature of treat- ment facilities for drinking problems. Only a minority of the drinkers in each sample were seeking treatment for drinking problems. Ilie analysis asked two questions of the data. First, was lie structure of alcohol dependence symptoms similar in the six samples? Second, Structure and correlates of alcohol dependence 1629 were there similar relationships between alcohol dependence, alcohol consumption, and alcohol- related problems in each of tbe six centres? "Hie sceptical sociocultural view predicts that the structure of dependence symptoms and the relationships between dependence, consumption and problems would vary across the six centres. The highest degree of similarity in structure and relationships would be between the two English-speaking centres which repre- sent cultures in which the research that led to the development of the concept of the Alcohol Dependence Syndrome was undertaken (Aus- tralia and the US). Proponents ofthe hypothesis that the ADS is a universal consequence of chronic heavy drinking would predict a substan- tial degree of similarity in tbe structure of dependence, and in the relationships between alcohol dependence and consumption and prob- lems. In these respects the data provides a test of the cultural relativity of the Alcohol Dependence Syndrome. Method Study Centre samples The six study centres were: the Centre for Alco- hol Studies, Royal Prince Alfred Hospital, Sydney, Australia; the Department of Alco- holism, Institute of Neurology and Psychiatry, Sofia, Bulgaria; the Department of Psychiatry, University of Nairobi and Kenyatta National Hospital, Nairobi, Kenya; the Instituto Mexi- cano de Psychiatria, Mexico City, Mexico; the National Directorate for the Prevention of Alco- hol and Drug Problems, Oslo, Norway; and the Department of Psychiatry, University of Con- necticut, School of Medicine, Farmington, Connecticut, the US. In all centres subjects were recruited from primary care and hospital settings and selected so as to comprise three groups: (1) "drinking patients" who were currently seeking treatment for medical problems other than an alcohol problem, and who had not previously sought treatment for an alcohol problem; (2) clinically diagnosed alcoholics from specialist treatment centres; and (3) at least 60 non-drinkers. Sampling was over the period of a year to ensure that a representative sample of patients was obtained which was not affected by seasonal factors. Among drinking patients the following sampling quotas were used: at least 30 men and women were to be obtained in each of three age-strata, namely, 18-30 years, 31-iO years, 41-55 years. The quota for the alcohoUc and non-drinker samples were 60 each althougb the quota for alcoholic patients was only achieved in Norway and the US. Only the data on drinking patients and alcoholics are presented in this paper. Instruments Infonnation was collected by a standardized 150 item interview schedule which covered the fol- lowing conceptual domains: sociodemographic characteristics; physical health; presenting condi- tion (lCD-9 coding); current psychological and physical symptoms; past medical history; sub- stance use and diet; alcohol consumption; alcohol dependence; psychological reactions to alcohol; alcohol-related problems (psychological, social, physical); family history of alcohohsm, alcohol abuse and liver cirrhosis; self-perception of an alcohol problem; clinical examination findings, and results of laboratory tests. Only a subset of these data is reported in this paper, namely, alcohol dependence, alcohol consump- tion, alcohol-related problems, serum gamma glutamyltransferase (GGT), and the Le Go Grid (Le Go, 1976), a series of clinical ratings of such signs of chronic heavy drinking as hand tremor, liver enlargement, conjunctival injection, and al- cohol-related skin disorders. Details on tbe remaining questionnaires are supplied in Saun- ders et al. (1990). Alcohol consumption Alcohol consumption was assessed over the previous month by the tri-levei method in which each respondent was asked first to define "light", "medium" and "heavy" consumption, and second to indicate on how many of the past 30 days they were abstinent, or engaged in each of these drinking panems. Responses to these items were used to construct an estimate of the aggregate consumption of alcohol (in grams of ethanol) during the previous thirty days. Respondents were also asked to indicate the firequency with which they consumed more tban 6 and more than 12 drinks in the previous month. The validity of self-reported consump- tion was assessed by examining tbe correlations between log consumption, and log GGT and Le Go score. 1630 Wcome HaU et al. Alcohol dependence Thirteen items were designed to tap six of the seven conceptual elements of the ADS, exclud- ing re-instatement of the syndrome after a period of abstinence which was excluded as it had been deleted from DSM-III-R (see Appendix for indi- vidual items). Although these questions were compiled before publication of DSM-III-R (APA, 1987) and ICD-10 (WHO, 1992) they were based upon the conceptual elements incor- porated therein. Six items assessed the experience of impaired control or compulsion to drink. These were whether tbe person had ever: "found it difficult to get the thought of alcohol out of their mind"; "experienced being unable to stop drinking once you had started"; "found it difficult to stop drinking before becoming completely intoxi- cated"; "gulped drinks in order to speed up the effect of alcohol"; "stayed dnmk for several days at a time"; and "tried to reduce your alcohol consumption and failed". The salience of alcobol was assessed by two items which asked whether the person had ever: "skipped meals because you were drinking"; and "failed to do what was normally expected from you because of drinking". Tolerance was as- sessed by two questions asking whether the person had ever "been in a situation where you drank more than your friends", and "needed more alcohol tban you previously did in order to get tbe desired effect". Withdrawal was assessed by the single item "have you had your hands shake after a heavy drinking session" while relief avoidance was assessed by two items which in- quired wbether the drinker had ever "needed a drink to get yourself going the morning after a heavy drinking session" and "needed to drink alcohol at times ofthe day when you normally do not drink". Responses were recorded on a five point scale (0-4) on the basis of their frequency of occurrence, namely, "never", "less than monthly", "monthly", "weekly", "daily or almost daily". Alcohol-related problems Twenty-four items assessed the prevalence of having experienced personal, social and health problems commonly reported among clinical samples of alcoholics. These included questions as to: whether anyone had been harmed by tbe person's drinking; wbether the drinker had suf- fered a serious injuiy or a fracture as a result of their drinking; whether a fomily member had ever, or in the last year, suggested that the drinker cut down their alcohol consumption, or expressed concem about their level of drinking; whether tbey had ever, or in the past year, had experienced a problem with their job as a result of their alcohol use; whether they had experi- enced any legal problems or arrests as a consequence of their alcohol use. Data analysis The structure of the alcohol dependence items was explored by a Principal Components Analy- sis ofthe drinking subjects' responses to the 13 symptoms of the ADS in eacb of the six partici- pating centres. Cronbach's alpha was calculated on the 13 items for each centre to assess the internal consistency of the items. The degree of agreement between the factor loadings for the items in each centre was assessed by: the coefficient of congruence (Levine, 1988), an in- dex on which maximum congruence is indicated by a value of 1.00, and the closer the index is to this value tbe higher the degree of congruence between the samples; and the Root Mean Square difference between item loadings on the first Principal Component, with congruence being indicated by a small root mean square difference (Levine, 1988). A number of sensitivity analyses were under- taken to ensure that the results of the main Principal Components Analyses were robust. These included: separate analyses for males and females in each of the samples as a check on the legitimacy of combining males and females in the main analysis; Principal Factor Analyses in each of the six samples as a check on the sensitivity of the results to the method used to extract the factors; and Principal Components Analyses on dichotomised dependence symptoms as a check on tbe sensitivity of the results to departures from multivariate normality in the distribution of dependence symptoms. A number of analyses were undertaken to provide a limited check on the validity of self-re- ported alcohol consumption and dependence symptoms. The validity of self-reported alcohol consiimption was assessed by calculating Pear- son product moment correlation coefficients between alcohol consumption, the frequency Table Structure and correlates of alcohol dependence 1631 Demographic characteristics and patterns of alcohol consumption among the drinking patients in the samples from each of the six study centres n Male (%) Mean age Mean years education Median (g) consumption Drink 6 + weekly (%) Problem (%) Dependence G G T Aus 194 69.1 35.3 n.4 616.0 18.6 2 6 3 5.9 69.3 Bui 108 87.0 37.6 11.2 635.0 16.7 6.5 3.0 19.9 Ken 105 78.1 35.4 8.4 799.0 48.6 46.7 19.5 40.4 Nor 167 57.5 34.6 11.9 574.0 18.0 38.3 14.0 65.2 Mex 256 53.9 36.6 11.1 548.0 10.5 28.5 7.9 61.3 US 205 52.7 35.5 13.8 630.0 21.0 49.3 11.1 50.9 Key: Aus, Australia; BuJ, Bulgaria; Ken, Kenya; Mex, Mexico; Nor, Norway; US, United States. with which 12 or more drinks were consumed, and GGT and Le Go score. The validity of self-reported dependence symptoms was indirectly assessed by examining the relationships between alcohol dependence, alcohol consumption and alcohol-related problems. Although alcohol dependence, con- sumption and alcohol-related problems are conceptually independent valid measures of each should be positively correlated. Heavy alcohol consumption is a necessary (although probably not a sufficient) condition for developing an alcohol dependence syndrome, and the heavier a person's alcohol consumption is, the more likely they are to experience symptoms of depend- ence. Hence, there should be a positive correla- tion between dependence symptoms and self-reported alcohol-consumption. There should also be a positive relationship between alcohol dependence and alcohol-related prob- lems. A person can experience symptoms of alcohol dependence without experiencing alco- hol-related problems, and vice versa. Nonetheless, dependence and problems should be positively correlated because the probability of experiencing each increases with increasing levels of consumption. In all these correlational analyses the natural logarithms of alcohol dependence, self-reported alcohol consumption, and alcohol-related prob- lems were used instead of the raw scores as the latter were skewed. The sensitivity of these cor- relations to non-normality of score distributions and outliers were assessed by calculating both Pearson Produa Moment and Spearman Rank correlation coefficients, and by calculating both with and without outliers excluded from the analyses. Unless otherwise stated, the results reported are product moment correlations with all data included. Results Sample characteristics The sex and mean ages of the samples of drink- ing patients in each of the six study centres are shown in Table 1. The proportion of males and females in the sample varied from equality in the Mexico, Norway and the US to a predominance of males in the samples in Australia, Bulgaria and Kenya. The average age was in the middle thirties for all samples and the average years of education varied between 8.4 years in Kenya and 13.8 years in the US. There were differences in the median con- sumption for a typical month among drinkers in the six centres. The patterns of consumption were most alike in Australia (616 g), the US (630 g), and Bulgaria (635 g). The median con- sumption values for Norway (548 g) and Mexico (574 g) were lower, while that for Kenya was higher (799 g) than those for Australia, Bulgaria and the US. Analyses of Covariance on log monthly consumption (with sex and age as co- variates), indicated that the highest average consumption was in Kenya and the US with minimal differences between the other four centres. The prevalence of drinking to intoxication, as reflected in the proportion of drinkers who regu- larly drank more than 6 drinks on an occasion, was highest in Kenya where of 72.4% of drinkers reported doing so. Among the remaining five centres, the proportion varied between 23.4% in Mexico and 42.3% in Australia. The percentage of drinkers reporting that they currently had a problem with alcohol varied between a low of 1632 Wayne HaU et al. Table 2. Factor loadings on first Prmdpal Component iEN, P. (Eds) Cultural Stud- ies of Drinking Problems {Helsinki, The Social Research Institute of Alcohol Studies). BABOR, T . F . , STEPHENS, R . S . & MARLATT, G . A. (1987) Verbal report methods in clinical research on alcoholism: response hias and its minimization. Jour- nal of Studies in Alcohol, 48, pp. 410-424. CAETANO, R . (1990) The factor structiire congruence of the D S M - n i - R and ICD-10 concepts of alcohol dependence. Alcohol and Alcoholism, 2 5 , pp. 3 0 3 - 318. CHICK, J. (1980) Is there a unidimensional Alcohol Dependence Syndrome? British Joumal of Addiction, 75, pp. 265-280. DAVIDSON, R . , BUNTING, B . & RAISTRUCK, R . (1989) The homogeneity of the Alcohol Dependence Syn- drome: a factorial analysis of the SADQ questionnaire, British Joumal of Addiction, 84, pp. 907-915. EDWARDS, G . & GROSS, M . M . (1976) Alcohol depen- dence: provisional description of a clinical syndrome, British Medical Journal, 1, pp. 1058- 1061. EDWARDS, G . (1986) T h e alcohol dependence syn- drome: a concept as stimulus to enquiry, British Joumal of Addiction, 8 1 , pp. 171-183. HEATHER, B . B . & ROBERTSON, 1. (1990) Problem Drinking (Oxford, Oxford University Press). JELUNEK, E . M . (I960) 77K Disease Concept of Alco- holism (New Jersey, Hillhouse Press). L E G O , P. M. (1976) Le Depistage Precoce et Systema- tique dur Buveur Excessif (Paris, Department d'alcoologie therapeutique de Rion laboratories). LEVINE, M . S . (1988) Canonical Anafysis and Factor Comparison (Beverley HiUs, Sage Publications). MIDANIK, L . (1988) Validity of self-reported alcohol use: a literature review and assessment, British Jour- nal of Addiction, 83, pp. 1019-1029. P O U C H , J. M., ARMOR, D . J. & BRAIKER, H . B . (1980) 1636 Wayne HaU et al. The Course of Alcoholism: Four Ytars After Treatment (Santa Monica, CA, Rand Corporation). ROOM, R. (1980) Treatment seeking populatiotis and lai^er realities, in: EDWARDS, G. & GRANT, M . (Eds) Alcoholism Treatment in Transition (London, Croom Helm). ROOM, R. (1980) New curves in the course: a com- ment on Polich, Armor & Braiker "The Course of Alcoholism", British Joumal of Addiction, 75, pp. 351-360. ROOM, R. (1983) Sociology and the disease concept of alcoholism. Research Advances in Alcohol and Drug Problems, 17, pp. 47-91. SAUNDERS, J. B., AASLAND, O . G . , AMUNDSEN, A. & GRANT, M . (1990) \PHO Collaborative Projea on Eat^ Detection of Persons with Harrnful Alcohol Consumption. I. Alcohol consumption and related prob- lems among primary health care patients (Geneva, WHO). SAUNDERS, J. B., AASLAND, O . G . & GRANT, M . (1991) A UDIT--the World Health Organisation screening in- strument for harmful and hazardous alcohol consumption. A Report from the WHO Project on Identification and Treatment of Persons with Harmful Alcohol Consumption (Geneva, WHO). SHAW, S. (1980) A critique of the concept of the alcohol dependence syndrome, British Joumal of Ad- diction, 74, pp. 339-348. SKINKER, H . A. (I98I) Primary syndromes of alcohol abuse: their measurement and correlates, British Joumal of Addiction, 76, pp. 63-76. SKINNER, H . A. & ALLEN, B . A. (1982) Alcohol dependence syndrome: measurement and valida- tion, Joumal of Abnormal Psychology, 91, pp. 199-- 209. STOCKWHLL, T . , HODGSON, R., EDWARDS, G., TAYLOR, C. & RANKIN, H . (1979) The development of a questionnaire to measure alcohol dependence, British Joumal of Addiction, 74, pp. 79-87. WORLD HEALTH ORGANIZATION (1992) The ICD-10 Classification of Mental and Behavioural Disorders: Clinical Descriptions and Diagnostic Guidelines (Geneva, WHO). Appendix The 13 dependence items used in the study "Now I am going to ask you some questions about the way you drink. Please use the card to describe how often you have experienced each of the foEowing dur- ing the past 12 months: CARD 0: never during the last year 1: less than monthly 2: monthly 3: weekly 4: daily or almost daily 1. Found it dif&ctilt to get the thought of alcohol out of your mind. 2. Skipped meals because you were drinking. 3. Experienced that you were unable to stop drinking once you started. 4. Found it difficult to stop drinking before you became completely intoxicated. 5. Needed a first drink to get yourself going the morning after a heavy drinking session. 6. Been in a situation where you drank more than your friends. 7. Been gulping drinks in order to speed up the effect of alcohol. 8. Failed to do what was normally expected from you because of drinking. 9. Stayed drunk for several days at a time. 10. Needed more alcohol than you previously did in order to get the desired effect. 11. Tried to reduce your alcohol consumption and failed. 12. Needed to drink alcohol at times of the day when you normally do not drink. 13. Had your hands shake a lot in the morning after drinking."
