Taking Account of Cultural and Societal Influences on Substance Use Diagnoses and Criteria

This paper is concerned with potential variations between cultures in the meaning and meaningfulness of five different diagnostic categories in the substance use disorders: dependence, abuse, harmful use, intoxication and withdrawal. We are thus not concerned with differences between cultures in population rates of the diagnostic categories and of their criteria, but rather with prior questions of the meaningfulness and meaning of the criteria and diagnoses in different cultures.

There are two main traditions by which the issue of such cross-cultural variations has been addressed, primarily but not solely with reference to alcohol. One tradition starts from a position of universalism and philosophical realism, presuming that there is a single underlying dependence disorder applicable, for instance, to all humankind. The tradition may recognize problems in applying a diagnosis or instrument in a particular culture, but the solution to the problems lies in finding new operationalizations—the universal applicability of the underlying concept is not questioned. This has been the mainline position of American psychiatric epidemiology since the “St Louis revolution”, and it is probably difficult for many of us to set it aside and consider alternative perspectives.

The alternative tradition is more nominalist and particularistic, viewing substance use diagnoses as culturally influenced and allowing for the possibility that the cultural influence in framing the criteria or diagnoses can be strong enough that they are inherently different in different cultures. This tradition applies to the field of diagnostic concepts and instruments ethnographic perspectives on cultural variations in the meaning of substance use and intoxication as developed for alcohol, for instance, by MacAndrew & Edgerton (1).

In his late writing about alcoholism, Jellinek was the first modern proponent of this way of thinking about substance use diagnoses. Jellinek was a thoroughgoing nominalist about what counts as a disease or disorder—more thoroughgoing than I would choose to be, for instance, when he said that “it comes to this, that a disease is what the medical profession recognizes as such” (2). His late definition expanded the frame of “alcoholism” so that it lost most of its specific meaning: “any use of alcoholic beverages that causes damage to the individual or society or both” (2), and his Greek-letter types of alcoholism within this overarching frame essentially reflected the different “species” he had heard described as alcoholism by doctors from different countries at World Health Organization (WHO) meetings: gamma alcoholism was the “Anglo-Saxon” (i.e. American) species, delta alcoholism the French and epsilon the Finnish (3).

In recent decades, there has been considerable research on the cross-cultural applicability of substance use disorder criteria and diagnoses, particularly for alcohol. Many of these studies have been framed in the paradigm of the realist and universalist tradition. For example, the book by Helzer & Canino (4) on studies conducted with the Diagnostic Interview Schedule (DIS) instrument relies primarily on the fact of common methodology and that usable data could be produced as its warrants of comparability across a variety of studies (5), but does include some side-comments about issues of applicability. Another study, a side-product of the WHO collaborative project which produced the Alcohol Use Disorders Identification Test (AUDIT), performed principal-component factor analyses in each of six diverse cultures of 13 alcohol dependence items covering aspects of impaired control, salience of drinking, tolerance and withdrawal (6). The analysis found a strong general factor in each factor analysis, with very high Cronbach's alphas, and interpreted the results as support for “the hypothesis that the Alcohol Dependence syndrome has considerable cross-cultural generalisability, regardless of treatment ideology, culturally learned drinking patterns or societal response to drinking problems”. However, the paper also found that the dependence score formed from the items correlated quite highly with frequency of drinking 12 + drinks on an occasion (0.67–0.86) and with a logged score of alcohol-related personal, social and health problems (0.65–0.89). These findings are further evidence of cross-cultural commonality, but raise the question: commonality in what terms? Is it specifically the Alcohol Dependence syndrome which serves as the engine of the cross-cultural commonalities, or might it as well be the symptomatology of frequent intoxication or the experience of alcohol-related problems?

Another study was focused more specifically on assessing the validity of Diagnostic and Statistical Manual (DSM) and International Classification of Diseases (ICD) diagnoses cross-culturally, with test-retest and cross-instrument comparisons. However, the primary weight in the main round of published analyses from this study (7,8) was on analyses combining the different sites. The main site-specific comparative analysis in this series (9) found substantial variation across sites in test-retest reliability, with Sydney, Australia showing the highest reliability on seven of nine current alcohol dependence items (Sydney range: 0.73–0.90). Bangalore, India showing the lowest on all nine items (Bangalore range: 0.29–0.53) and the results in Jebel, Romania closer to those for Sydney than those for Bangalore. Sydney and Jebel both also showed relatively high reliability on two alcohol abuse items, while again Bangalore showed lower reliability. Examining the patterns of discrepancies, the authors concluded that the Bangalore respondents appeared to have “difficulty understanding the constructs underlying the questions”.

The alternative tradition relies on a broader range of types of evidence, reaching outside the bounds of DSM or ICD instruments. Some of the evidence comes from quantitative studies in the realist/universalist tradition which, like Chatterji et al. above, have sometimes produced findings which are problematic for the paradigm. Thus Klausner & Foulks (10) were stimulated by the outrage of their subjects at the practical impact of their study (including an abrupt fall in the market value of the community's municipal bonds) to return to their data for a kind of after-the-fact protocol analysis of what their Inuit respondents could have meant in their answers to the items of the Michigan Alcoholism Screening Test (MAST) score. Somewhat ruefully, the authors concluded that many of the MAST items might well have a different meaning in an Arctic indigenous culture than, for instance, in urban Michigan. Applying the Munich Alcoholism Test (MALT) developed in Germany to samples in Spain and Ecuador, Gorenc et al. (11) found that five of the 31 items were “relatively free of cultural differences” by their criteria, but added that none of the five items as used in Ecuador passed the filters used to screen out items when the test was developed in Germany. Drawing on historical and anthropological studies, the present author argued 20 years ago (12) that perhaps alcohol dependence should be viewed as a “culture-bound syndrome”, with a specific history and cultural inception which starts with the early American temperance movement (13).

The most extensive study so far drawing on this more phenomenological tradition was the nine-site WHO study of the cross-cultural applicability of the substance use disorders (14–16). With limited resources and time for the data collection, the study's ambitious programme of key informant interviews, focus groups and reference case interviews was only partly carried out, and the analyses rely principally on the key informant interviews with knowledgeable local professionals and laypeople—20 at each site concerning alcohol concepts and terms and 20 concerning the drug with the highest apparent rate of harmful use at the site. The study identified problems of cross-cultural applicability at the level of instrument items, at the level of criteria, and at the level of concepts and diagnoses. Some problems at the item level were easily solved: phrases such as “driving an automobile or operating a machine” obviously require adaptation. Others were less tractable: it was noted that “the diagnostic criteria and their operationalizations assume a self-consciousness about feelings, knowledge and consciousness which is foreign to the folk traditions of some cultures” (15). Thus there was no accurate translation in one or another society for words such as “feel” and “anxiety”. Items and criteria “often also have built-in attributional, causal and other relational assumptions which are not customary in some languages and cultures”. Thus formulations such as “trouble because of drinking”, “after you had realized it had caused you” and “where it increased your chances of getting hurt”, and items mentioning intentions presume “both self-consciousness and a style of causal attribution which is unrecognizable in some cultures” (15). The problems posed by such items are not a mere problem of translation. Built into the DSM and ICD criteria are formulations which reach across and connect different domains of meaning. Thus, for instance, “the substance use continued despite knowledge of having a…problem that is likely to have been caused or exacerbated by the substance” is a criterion which requires connecting together acknowledgements of use, of a problem and of cognition about a causal relation between the two. The attribution of causality to alcohol and drug use, in particular, varies across cultures, and for that matter has varied in different periods within cultures (16, 17). The study identified several different types of difficulty with the criteria for dependence. In some cultures there was no term for the criterion; in others the meaning of two criteria overlapped; while in others the criterion was not considered of diagnostic significance (15).

In the context of categorical diagnoses and criteria, the issue of the threshold at which an item or criterion is considered positive becomes particularly important (16). The WHO cross-cultural applicability study found many instances of different thresholds being applied. Where use of the substance is particularly suspect, the thresholds may be set very low. Thus in Bangalore, several reference cases qualified as positive on three or four alcohol dependence criteria on the basis of drinking a maximum of three drinks up to three times a month (15). Conversely, in Athens and Santander, Spain, where regular drinking is normalized, thresholds for what is problematized were set much higher (16). We may expect the same kinds of cross-cultural variation for other drugs according to the social acceptability and familiarity of the substance. Whether and how a threshold is defined is important in distinctions between normal drinking or drug use and harmful use or intoxication (14), and between hangover and withdrawal (15), as well for the criteria of dependence.

The issue of divergent thresholds brings to the fore the fact that clinical diagnoses in the alcohol and drug area, more than most other diagnoses, usually carry a weight of moral judgement, whatever the clinician's intentions. Of course, it is not that all use is always negatively evaluated. In most human societies, one or another psychoactive substance is a valued commodity for human ingestion. Human use values for psychoactive substances are varied (18)—to ensure wakefulness, to promote sleep, to bring euphoria, to deaden pain, to pursue a transcendent experience, to quench thirst, as a nutrient, as a medium of commensality and sociability, as a signal of exclusion, and so on. The same substance often has apparently contradictory use values, sometimes simultaneously.

On the other hand, use of psychoactive substances beyond some socially defined limit (or, in some cases, at all) is commonly moralized and stigmatized. One has only to mention terms such as “drug fiend”, “demon rum” and the “scourge” or “menace” of drugs to recognize the extent to which drug use is often stigmatized. In many societies a common means of derogating opponents is to label them as drunks or drug users (19); hence, as a policy of political prophylaxis, the ban on alcohol at the December 2004 encampments of the “orange revolution” in Kiev (20). Another WHO study, of the cross-cultural applicability of disability concepts and measures in 14 societies, found some variation between societies in the ranking by informants of “alcoholism” and “drug addiction” in terms of degree of stigma (21) (Table 1). However, the overall picture was that both conditions were ranked as among the most stigmatized of 18 conditions, roughly on a par with being “dirty and unkempt” and having a “criminal record for burglary”. The moralization of drinking or drug use, beyond thresholds that vary between cultures, seems to be one commonality we can find between many modern societies.

To a certain extent, the stigmatization is built into the diagnostic terminology of DSM-IV. This is most obvious in the use of the term “abuse”. Accordingly, in 1993 the Board of the American Society of Addiction Medicine, “while recognizing that ‘abuse’ is part of present diagnostic terminology”, recommended “that an alternative term be found for this purpose because of the pejorative connotations of the word “abuse”' (22). It can be argued that the term “dependence” also came with some built-in stigma; when it was adopted in the 1960s and 1970s by World Health Organization committees it already carried a baggage of discreditable previous meanings in the United States, as in “dependent personality” or “welfare dependency” (23).

Whatever diagnostic terminology is employed, the diagnoses are used primarily in clinical settings late in a process which begins in everyday life and interactions. Typically, for a majority of the cases coming into an alcohol- or drug-specific treatment service, someone—a spouse, another family member, a judge, a social worker—has made a judgement that there is a problem needing clinical attention. In fact, of those entering alcohol treatment in a California county, over 40% had received an ultimatum from someone to enter treatment, in 24% of the cases from a family member (24). In such circumstances, a de-facto part of the diagnostic decision-tree is the threshold at which family, friends or officials in the society notice a behavior and decide that it should be brought to professional attention. Often attached to these processes of “noticing”, as Table 1 implies, is a great deal of stigmatization.

One decision for DSM is whether and to what extent a medical diagnostic system should build these essentially social judgements into the diagnosis and criteria, and to what extent it should seek to build diagnoses and criteria which are independent of them. This is the primary issue on which DSM-IV and ICD-10 parted company, with ICD-10's “harmful use” in principle excluding negative social consequences or reactions of others to drug use as evidence of harmful use (25, p. 74–75], while DSM-IV's “abuse” was primarily built around them. I have noted that this divergence “reflects a longstanding difference between British and American psychiatry, with the British taking the view that social reactions and consequences do not belong in definitions of diseases and disorders” (3). Behind this difference, I believe, lie not only differences about the inclusion of stigmatizing terms in the diagnostic system, but also the very different institutional frames of British and American psychiatry. In the context of the National Health Service, British psychiatry has been in a good position to define for itself the limits of its reach, with little to lose from turning away cases that fall outside those limits. In the absence of a national health system, the social environment of American psychiatry has been more entrepreneurial and less inclined to decide collectively that cases lie outside its competence. “A health system like the American, characterized by fee-for-service and managed care”, has encouraged inclusiveness in the criteria and thresholds, to make it “unlikely that a clinician will have to turn away anyone appearing for treatment on the grounds that they do not qualify for the diagnosis” (3).

To the extent this judgement is right, it underlines that cultural differences in the nature of alcohol and drug problems which are presented to the health system reflect not only cultural differences in norms and behaviors around substance use, but also societal and cultural differences in how alcohol and drug problems are defined and handled. That is, the difference between the British and American views is not so much a reflection of differences in the nature of alcohol and drug use (in a global perspective, these differences are not very great) as it is of differences in the way that problems from the use are handled.

DSM started as a diagnostic manual for the United States, but it has obviously taken on a much broader significance. The efforts to conform the substance use diagnoses and criteria in DSM-IV and ICD-10 (3) exemplify the fact that DSM has a global reach. In this context, it seems important to take into account that problem definitions and social handling systems for alcohol and drug problems differ considerably in different societies. For instance, to include “abuse” within the competence of psychiatrists and other clinicians may make sense in the context of the US system, with its strong interlinkage with the criminal justice system, but may make less sense elsewhere.

To take one example, in Sweden, where the general lay term for those with alcohol or drug problems translates as “misusers”, the primary institutional frame for alcohol and drug treatment (accounting for two-thirds of it) has long been the social welfare system, with the health-based system taking on specific tasks such as detoxification and opiate maintenance (26). The Swedish system thus does not need a medical diagnosis of “abuse” to function. It is not that Swedish doctors ignore alcohol and drug problems: in fact, there has been a long tradition of concern by Swedish doctors about alcohol problems, but the fairly consistent theme for a century has been that, while there are medical aspects such as cirrhosis, the problem is primarily social in nature (27).

For the following discussion. I take as one of my cues the comments by Marc Schuckit at a symposium on the validity of DSM-IV dependence, including the possibility of moving toward a dimensional approach (28). Another cue is the fact that, by international treaty, national diagnostic systems must be based on the ICD, which means that a DSM classification has to be fitted within the frame of ICD diagnoses. The other cues for the discussion come from the findings of the studies of cross-cultural comparability:

It is recognized that reference to feeling and affect states cannot be avoided for one of the diagnoses, dependence, discussed below.

My suggestions for directions of work are specified in terms of five current diagnoses (three shared by ICD-10 and DSM-IV, and harmful use from ICD-10 and abuse from DSM-IV).

While there is much in the existing literature to draw upon, a substantial programme of research and developmental work would be required to move in the directions recommended here. Reanalyses of existing data sets, both qualitative and quantitative, can contribute to the refinement of conceptualizations and measures. But if diagnoses, criteria and instruments are to be truly cross-culturally applicable, there is a need for cross-cultural testing in the developmental phase, with a substantial programme of work using such methods as key informant interviews and reference case or protocol analysis studies. Quantitative studies of test-retest reliability and convergent validation, for instance, should follow, but the purpose of these studies is more negative than positive: to establish that the measures meet acceptable standards across a range of societies, rather than establishing validity in any absolute sense. The possibility should be held open that some diagnoses or criteria should be specified as having a bounded applicability: to apply in a specified range of societies, but not necessarily outside them.