Skip to main content
PMC home page
PLOS Medicine logoLink to PLOS Medicine
. 2008 Jul 1;5(7):e141. doi: 10.1371/journal.pmed.0050141

Toward a Global View of Alcohol, Tobacco, Cannabis, and Cocaine Use: Findings from the WHO World Mental Health Surveys

Louisa Degenhardt 1,*, Wai-Tat Chiu 2, Nancy Sampson 2, Ronald C Kessler 2, James C Anthony 3, Matthias Angermeyer 4, Ronny Bruffaerts 5, Giovanni de Girolamo 6, Oye Gureje 7, Yueqin Huang 8, Aimee Karam 9, Stanislav Kostyuchenko 10, Jean Pierre Lepine 11, Maria Elena Medina Mora 12, Yehuda Neumark 13, J Hans Ormel 14, Alejandra Pinto-Meza 15, José Posada-Villa 16, Dan J Stein 17, Tadashi Takeshima 18, J Elisabeth Wells 19
Editor: Phillipa Hay20
PMCID: PMC2443200  PMID: 18597549

Abstract

Background

Alcohol, tobacco, and illegal drug use cause considerable morbidity and mortality, but good cross-national epidemiological data are limited. This paper describes such data from the first 17 countries participating in the World Health Organization's (WHO's) World Mental Health (WMH) Survey Initiative.

Methods and Findings

Household surveys with a combined sample size of 85,052 were carried out in the Americas (Colombia, Mexico, United States), Europe (Belgium, France, Germany, Italy, Netherlands, Spain, Ukraine), Middle East and Africa (Israel, Lebanon, Nigeria, South Africa), Asia (Japan, People's Republic of China), and Oceania (New Zealand). The WHO Composite International Diagnostic Interview (CIDI) was used to assess the prevalence and correlates of a wide variety of mental and substance disorders. This paper focuses on lifetime use and age of initiation of tobacco, alcohol, cannabis, and cocaine. Alcohol had been used by most in the Americas, Europe, Japan, and New Zealand, with smaller proportions in the Middle East, Africa, and China. Cannabis use in the US and New Zealand (both 42%) was far higher than in any other country. The US was also an outlier in cocaine use (16%). Males were more likely than females to have used drugs; and a sex–cohort interaction was observed, whereby not only were younger cohorts more likely to use all drugs, but the male–female gap was closing in more recent cohorts. The period of risk for drug initiation also appears to be lengthening longer into adulthood among more recent cohorts. Associations with sociodemographic variables were consistent across countries, as were the curves of incidence of lifetime use.

Conclusions

Globally, drug use is not distributed evenly and is not simply related to drug policy, since countries with stringent user-level illegal drug policies did not have lower levels of use than countries with liberal ones. Sex differences were consistently documented, but are decreasing in more recent cohorts, who also have higher levels of illegal drug use and extensions in the period of risk for initiation.

Louisa Degenhardt and colleagues report an international survey of 17 countries that finds clear differences in drug use across different regions of the world.

Editors' Summary

Background.

Understanding how much disability and death a particular disease causes (known as the “burden of disease”) is important. Knowing the burden of a disease in a country contributes to the development of healthier nations by directing strategies and policies against the disease. Researchers' understanding of the burden of diseases across different countries was piecemeal until the 1990 launch of a special World Health Organization (WHO) project, the Global Burden of Disease Project. In 2002, on the basis of updated information from this ongoing project, the WHO estimated that 91 million people were affected by alcohol use disorders and 15 million by drug use disorders.

Why Was This Study Done?

It is widely accepted that alcohol, tobacco, and illegal drug use are linked with a considerable amount of illness, disability, and death. However, there are few high-quality data quantifying the amount across different countries, especially in less-developed countries. The researchers therefore set out to collect basic patterns of alcohol, tobacco, cannabis, and cocaine use in different countries. They documented lifetime use of these substances in each county, focusing on young adults. They also wanted to examine the age of onset of use and whether the type of drugs used was affected by one's social and economic status.

What Did the Researchers Do and Find?

Data on drug use were available from 54,069 survey participants in 17 countries. The 17 countries were determined by the availability of collaborators and on funding for the survey. Trained lay interviewers carried out face-to-face interviews (except in France where the interviews were done over the telephone) using a standardized, structured diagnostic interview for psychiatric conditions. Participants were asked if they had ever used (a) alcohol, (b) tobacco (cigarettes, cigars or pipes), (c) cannabis (marijuana, hashish), or (d) cocaine. If they had used any of these drugs, they were asked about the age they started using each type of drug. The age of first tobacco smoking was not assessed in New Zealand, Japan, France, Germany, Belgium, The Netherlands, Italy, or Spain. The interviewers also recorded the participants' sex, age, years of education, marital status, employment, and household income.

The researchers found that in the Americas, Europe, Japan, and New Zealand, alcohol had been used by the vast majority of survey participants, compared to smaller proportions in the Middle East, Africa, and China. The global distribution of drug use is unevenly distributed with the US having the highest levels of both legal and illegal drug use among all countries surveyed. There are differences in both legal and illegal drug use among different socioeconomic groups. For example, males were more likely than females to have used all drug types; younger adults were more likely than older adults to have used all drugs examined; and higher income was related to drug use of all kinds. Marital status was found to be linked only to illegal drug use—the use of cocaine and cannabis is more likely in people who have never been married or were previously married. Drug use does not appear to be related to drug policy, as countries with more stringent policies (e.g., the US) did not have lower levels of illegal drug use than countries with more liberal policies (e.g., The Netherlands).

What Do These Findings Mean?

These findings present comprehensive and useful data on the patterns of drug use from national samples representing all regions of the world. The data will add to the understanding of the global burden of disease and should be useful to government and health organizations in developing policies to combat these problems. The study does have its limitations—for example, it surveyed only 17 of the world's countries, within these countries there were different rates of participation, and it is unclear whether people accurately report their drug use when interviewed. Nevertheless, the study did find clear differences in drug use across different regions of the world, with the US having among the highest levels of legal and illegal drug use of all the countries surveyed.

Additional Information.

Please access these Web sites via the online version of this summary at http://dx.doi.org/10.1371/journal.pmed.0050141.

Introduction

Alcohol, tobacco, and illegal drug use are held responsible for considerable mortality and morbidity [1], but in the most recent World Health Organization (WHO) Global Burden of Disease estimates, the authors unanimously asserted that better epidemiological data on use were needed, particularly in less established market economies [2–4]. This paper presents data on lifetime alcohol, tobacco, cannabis, and cocaine use from rigorously conducted field surveys using a common research approach in the first 17 countries to participate in the WHO's World Mental Health (WMH) Survey Initiative [5,6]. A number of less established market economies are included in this set of countries.

Cross-national research on diseases and their putative risk factors has long been acknowledged as extremely difficult [7,8]. The creation of an international classification of causes of death advanced this line of inquiry [9], and was followed by the developments of international classifications and measurement approaches for chronic conditions, including mental disorders [10]. Historically, cross-national comparisons for alcohol and tobacco were undertaken using correlation studies of nation-level consumption (e.g., taxation data) plotted in relation to pertinent causes of death (e.g., liver cirrhosis, lung cancer). There is continuing uncertainty, however, about the comparability of death certification practices across countries, and for illegal drugs particularly, official statistics are considered unreliable in many countries [11].

Following refinements in survey research, well-specified and standardised methods were developed for population surveys on alcohol use [12–14], illegal drug use [15], and tobacco use [16]. Cross-national elaboration of these protocols is difficult [14,17], though, and has consequently been limited. More qualitative “rapid assessment” methods, widely adopted in the emerging market economies and valuable for within-country planning purposes, are difficult to use for the purposes of cross-national comparisons [18].

Cross-national comparisons within regions have been conducted in Europe [19–22] and in Panama, Central America, and the Dominican Republic (PACARDO) [23]. Efforts to collate existing survey data on alcohol and tobacco use (and abstention) have been undertaken through the WHO's Global Status Report on Alcohol [24] and the Tobacco or Health: Global Status Report [25]. Recently, the Global Youth Tobacco Survey (GYTS) [26] and a cross-national survey known as the Gender, Alcohol, and Culture: an International Study (GENACIS) [20], included population surveys of tobacco or alcohol use in selected populations in multiple countries. None of these, though, examined alcohol, tobacco, and illegal drug use concomitantly, in all regions, and across all ages.

The current study presents data on basic epidemiological patterns of alcohol, tobacco, cannabis, and cocaine use in 17 countries participating to date in the WMH surveys [5,6]. The WMH countries include countries in which cannabis use has been strongly prohibited (the US) as well as countries in which a harm reduction policy has long been in place (the Netherlands). The survey also includes cocaine source (Colombia) and consumer (US, Europe) countries.

This paper has the following objectives: (a) document the cumulative (lifetime) use of alcohol, tobacco, cannabis, and cocaine in each country, with some focus on young adults; (b) consider sociodemographic correlates of these types of drug use; and (c) examine the age of onset (AOO) distribution of such drug use.

Method

Participants

Eighteen surveys were carried out in 17 countries in the Americas (Colombia, Mexico, US), Europe (Belgium, France, Germany, Italy, Netherlands, Spain, Ukraine), the Middle East and Africa (Israel, Lebanon, Nigeria, South Africa), Asia (Japan, and separate surveys in Beijing and Shanghai in the People's Republic of China), and Oceania (New Zealand). This set of countries was determined by availability of collaborators in the country who were able to obtain funding for the survey and complete the World Mental Health Surveys (WMHS) protocol. Details of each sample are presented in Table 1 (see also http://www.hcp.med.harvard.edu/wmh/) [27].

Table 1.

Characteristics of Studies Included in the World Mental Health Survey

graphic file with name pmed.0050141.t001.jpg

Open in a new tab

All surveys were based on multistage probability samples. All interviews were carried out face-to-face by trained lay interviewers. The six Western European surveys were carried out jointly [28,29]. Consistent use of a standardized interview translation protocol, training procedures, and field quality control monitoring were used to minimize between-country variation in data quality [30].

Sample sizes ranged from 2,372 (Netherlands) to 12,992 (New Zealand), with a total of 85,052 participating adults. Response rates range from a high value of 88% (Colombia) to a low value of 46% (France), with a weighted average response rate of 70%. As described in detail elsewhere [27], internal subsampling was used to reduce respondent burden by dividing the interview into two parts. All participants completed Part I, which included core standardised items to assess conditions of central interest, including whether the participant was a current or former tobacco smoker. Part II included standardised items about correlates and disorders of secondary interest. Part II was administered to an enriched subsample (n = 43,249) that included 100% of those who met criteria for any Part I disorder and a random subsample of approximately 25% of other Part I respondents. As alcohol and illicit drug use were assessed in Part II, the Part II sample is considered in this report. Part II cases were weighted by the inverse of their probability of selection in order to adjust for differential sampling. There was also poststratification adjustment to bring the sample distributions into balance with population sociodemographic and geographic distributions [6]. The complexity of these adjustments differed across countries depending on the amount of population data available for poststratification. More details about sampling and weighting procedures are discussed elsewhere [27].

Measures

All WMH surveys used the WHO Composite International Diagnostic Interview, Version 3.0 (Composite International Diagnostic Interview [CIDI] 3.0), a fully structured diagnostic interview for psychiatric conditions [5,6]. Within this assessment, participants were asked if they had ever used (a) alcohol, (b) tobacco (cigarettes, cigars, or pipes), (c) cannabis (marijuana, hashish), and (d) cocaine. Those who had used these drugs were asked about the AOO of use of each drug class, except in New Zealand, Japan, France, Germany, Belgium, the Netherlands, Italy, and Spain, where age of first tobacco smoking was not assessed.

Sex and age of all participants was recorded. The following covariates defined as at year of interview were also studied: (a) completed years of education (grouped as: no education, some primary, primary finished, some secondary, secondary finished, some college, and college or more); (b) relationship status (never married, previously married, married-cohabitating); (c) employment (homemaker, retired, other [including unemployed], working/student); and (d) family income (low, low-average, high-average, and high, defined via assessment of total household income, with in-country medians calculated). In the case of household income, the standard international labour economics method [31] was used to define low-income respondents as those whose after-tax household income per family member was less than half the median within their country. Low-average income was defined as up to the median income per family member in the country. High-average income was defined as income per family member above the median up to three times the median, while high income was defined as income per family member above three times the population median.

Retrospective AOO reports were based on a question series designed to avoid the implausible response patterns obtained in using the standard Composite International Diagnostic Interview [CIDI] AOO question [32]. Experimental research shows this question sequence yields responses with much more plausible AOO distributions than the standard Composite International Diagnostic Interview [CIDI] AOO question [33]. Although AOO questions were asked both about important symptoms (e.g., first panic attack) and full syndromes, the ages used here are for first use of each drug.

Training and Field Procedures

The central WMH staff trained bilingual supervisors in each country. Consistent interviewer training documents and procedures were used across surveys. The WHO translation protocol was used to translate instruments and training materials. Standardized descriptions of the goals and procedures of the study, data uses and protection, and participants' rights were provided in written and verbal form to all respondents before verbal informed consent was secured. Quality control protocols were standardized across countries to check on interviewer accuracy and to specify data cleaning and coding procedures. The institutional review board of the organisation that coordinated the survey in each country approved and monitored compliance with procedures for obtaining informed consent and protecting participants. A more detailed discussion of these procedures is presented elsewhere [27,30].

Data Analysis

The cumulative incidence of use was estimated in the conventional fashion, as the proportion of respondents who ever had a given disorder up to their age at interview. AOO distributions and projected lifetime risk at given ages were estimated using the two-part actuarial method implemented in SAS 8.2 [34]. The actuarial method differs from the more familiar Kaplan-Meier [35] method in using a more accurate way of estimating the timing of onsets within a given year [36], although both methods assume constant conditional risk of onset at a given year of life across cohorts. Sociodemographic predictors of lifetime risk were examined using discrete-time survival analysis with person-year as the unit of analysis [37]. The predictors considered here were sex, cohort, and a sex-by-cohort interaction. Cohort was defined by age at interview. We also examined a categorical version of the cohort variable, which distinguished respondents who were in the age ranges 18–29, 30–44, 45–59, and ≥60 y at interview. The sex-by-cohort interaction was examined to determine if the well-known gender difference in drug use has become smaller in recent years in some or all countries studied. Sociodemographic correlates of lifetime use at the time of interview were examined using logistic regression analysis [38]. Sociodemographic variables (e.g., family income, relationship status) were coded as of the time of interview, not as of the time of first drug use, which means that the associations examined might reflect influences of previous drug use on sociodemographic characteristics. It is consequently illegitimate to interpret the associations in temporal terms. Instead, the associations provide only cross-sectional descriptive information. Standard errors of cumulative incidence estimates, survival coefficients, and logistic regression coefficients were estimated using the Taylor series linearisation method [39] implemented in the SUDAAN software system [40,41]. Survival coefficients, logistic regression coefficients, and their 95% confidence intervals were exponentiated and are reported in the form of odds ratios (ORs) for ease of interpretation. Multivariate significance tests were made with Wald χ2 tests using design-based coefficient variance-covariance matrices. Significance tests were consistently evaluated at the 0.05 level with two-sided tests.

Results

Cumulative Lifetime Incidence

Clear differences can be seen in the cumulative (lifetime) incidence of drug use across countries (Table 2). Lifetime alcohol use was reported by the vast majority of respondents in the Americas, Europe, Japan, and New Zealand, while considerably smaller proportions of respondents ever used alcohol in the Middle East, Africa, and China. Lifetime tobacco use was most common in the US (74%), Lebanon (67%), Mexico (60%), and in some European countries (Netherlands, 58%; Ukraine, 60%), with by far the lowest proportions in the African countries (South Africa, 32%; Nigeria, 17%). The proportions of respondents who ever used cannabis were highest in the US (42%) and New Zealand (42%), whereas lifetime cannabis use was virtually nonexistent in the Asian countries (Table 2). The US was an outlier in lifetime cocaine use, with 16% of respondents reporting that they had tried cocaine at least once compared to 4.0%–4.3% in Colombia, Mexico, Spain, and New Zealand, and extremely low proportions in countries in the Middle East, Africa, and Asia.

Table 2.

Estimated Cumulative Incidence of Drug Use

graphic file with name pmed.0050141.t002.jpg

Open in a new tab

AOO Distributions

Figure 1 presents country-specific data on the AOO distributions of drug use among those reporting use of each drug type. As is clear from these graphs, there was remarkable similarity in the AOO distributions for specific types of drugs across countries. The median AOO was between 16–19 y for alcohol for all countries except South Africa (20 y), and for tobacco in all countries except Nigeria (21 y) and China (20 y). The median AOO of illegal drug use was slightly older in all countries. For cannabis, median AOO was between 18–19 y, with the exception of Nigeria and Israel (both 22 y) and Lebanon (21 y). Cocaine use typically began at a slightly older age, with median AOO between 21–24 y for all countries where sufficient data were available to make an estimate.

Figure 1. AOO of Drug Use among Those Reporting Any Use by Country.

Figure 1

Open in a new tab

Note: Where lines are not presented for an individual country, either there was no assessment of the AOO of that drug, or fewer than 30 persons reported having used the drug (see Tables 2 and 5).

Equally remarkable as the consistent median AOO across countries is the age range of risk. Fully half the people who had ever used alcohol began using in the 7-y age range between 14–21 y. The interquartile range (IQR) (i.e., 25th–75th percentiles) of AOO distributions were typically 15–21 y for tobacco, 16–22 y for cannabis, and 19–28 y for cocaine.

Cross-National Differences in Lifetime Incidence in Recent Cohorts

Table 3 characterises the drug use history of young adults (22–29 y) in each country by age. Alcohol use by age 15 y was far more common in European countries than in the Middle East and Africa. By age 21 y, the vast majority of young adults in European countries (76%–99%), Japan (92%), New Zealand (94%), and the Americas (78%–93%) had begun using alcohol; estimates were lower in the Middle East and Africa (40%–63%). In the Netherlands, Belgium, France, Germany, and New Zealand, >60% of young people had started to drink by age 15 y. With three exceptions (South Africa, Lebanon, and Nigeria), this threshold value of 60% was crossed by age 21 y in all countries studied, with especially large proportions starting to drink between ages 15 and 21 y in the Ukraine and Japan.

Table 3.

Estimated Cumulative Incidence of Drug Use by Age 15 and 21 y among 22–29 y Olds

graphic file with name pmed.0050141.t003.jpg

Open in a new tab

Data on age at first tobacco smoking were available for fewer countries; nonetheless, among those aged 22–29 y, an estimated 46% of young adults in the Ukraine had started to smoke by age 15 y, and 72% by age 21 y; similar estimates were obtained in the US (44% and 72%, respectively). Nigeria had the smallest estimated cumulative incidence proportion for tobacco smoking by age 15 y (7%), and the following intermediate estimates were observed: Israel (9%), South Africa (11%), Colombia (12%), Peoples' Republic of China (15%), Lebanon (18%), Mexico (21%). The rank ordering of countries with respect to use by age 21 y was almost identical (Table 3).

Differences in illegal drug use were more marked among young adults: by age 15 y, those in New Zealand (27%) and the US (20%) had the highest levels of cannabis use, with almost no use in Asia, Middle East, or Africa among this cohort. Few young adults in the Netherlands had used cannabis by age 15 y (7%; Table 3). The majority of young adults in New Zealand (62%) and the US (54%) had used by age 21 y, compared to 35% of those in the Netherlands.

Among this youngest cohort, cocaine use was extremely rare in all countries at age 15 y. By age 21 y, young adults in the US had by far the highest cumulative incidence of cocaine use (16%; Table 3). In Colombia (the only cocaine-producing country in this group) the estimate was 3%; and for the Netherlands, 1%.

Sex and Cohort Differences in Lifetime Risk

Table 4 presents the results of discrete-time survival analyses examining the association between cohort and first onset of use of each drug type. The model adjusts for age differences in the cohorts by examining differential risk at each year of life assuming linear associations between cohort and risk. The associations are remarkably consistent across countries and drug types, in that every one of the ORs is <1.0, indicating that risk at any given age is consistently higher in more recent cohorts than in older cohorts. The ORs have been normed to reflect relative-odds per 10 y of life. An OR of 0.8, for example, indicates that the odds of first use in any given year of life is, on average, only 80% as high among respondents who were age A at the time of interview compared to respondents who were age A − 10 y at interview.

Table 4.

Estimated Association between Cohort and Lifetime Risk of Each Drug, by Country

graphic file with name pmed.0050141.t004.jpg

Open in a new tab

These coefficients assume, possibly incorrectly, that the same relative-odds hold throughout the life course and linearly across the age range of respondents at the time of interview, but simplifying assumptions of these sorts are needed to grasp such a wide range of associations all at once. These assumptions are investigated next, but within the context of these limitations, the data are clear in showing that the dominant tendency in the data across countries, and across drug types, is for risk to have increased over historical time, defined by the life courses of the respondents in these surveys.

The easiest way to investigate the linearity assumption in Table 4 is to examine AOO distributions across cohort, so as to determine the extent to which between-cohort divergence can be seen consistently over the life course. Such data are presented in Figure 2, where the AOO distributions of drug use are reconstructed based on retrospective AOO reports using the actuarial method pooled across countries.

Figure 2. Age-Specific Cumulative Incidence of Drug Use by Birth Cohort.

Figure 2

Open in a new tab

Pooled (weighted) data from the WMHSs (n = 54,068).

Focusing first on the AOO curves for alcohol use, we can see a clear nonlinear association across cohorts, with respondents in the oldest cohort having a substantially lower AOO curve than respondents in the more recent cohorts. Although there is some evidence of higher risk of beginning use in the late teens among the youngest compared to the intermediate-aged cohorts, this difference is much less dramatic than the evidence of lower risk in the oldest cohort. It is also noteworthy that the intercohort variation, although discernible prior to the late teens, is relatively small up to this point in the life course, by which time roughly 60% of respondents in each age cohort had started to drink. It was largely in later-onset use rather than early-onset use that the intercohort variation emerged most clearly, with initiation of alcohol use continuing later on into young adulthood for those in the younger cohorts.

For tobacco, the interval for risk of starting use was similar, but the cumulative level was lower. There were no intercohort differences until around age 21 y, by which age around 40% in each cohort had used. Of notable significance, as for alcohol, were the changes across cohorts after this age. The highest overall level of use, reflecting continued initiation into the mid-20s, was in the second oldest cohort; lower cumulative levels (reflecting less initiation during the same period of life) were observed among those in the two youngest cohorts.

The interval of risk for starting cannabis use began in late adolescence and continued far longer in life for all cohorts, but very large intercohort differences were observed here, with the oldest cohort dramatically lower than all others, and the two youngest cohorts much higher than the second oldest cohort. Clear and important age-specific differences were also evident across cohorts: the two youngest cohorts were very similar until the age of around 18 y in their cumulative incidence of drug use—but initiation of cannabis use continued to occur at a higher level following this age. A similar and more marked difference between the second and third cohorts was also evident, with initiation of cannabis use highly unlikely to occur for the second oldest cohort after the age of around 25 y. Clearly, then, two trends are evident: early onset use is greater for the three youngest cohorts than for the oldest, but much more marked is the fact that with each successive cohort, there is a prolongation of the period of risk for initiation of cannabis use far beyond adolescence.

The estimated cumulative incidence of cocaine use to age 60 y was much lower than for the other drugs, making it difficult to derive a useful “interval of risk,” but the intercohort differences appeared more linear—as can be seen in Figure 2, the distance between curves for adjacent cohorts was fairly consistent. Also clear is the fact that for most cohorts, initiation of cocaine use continued into the third decade of life, flattening after age 30 y. It is unclear whether such a trend would be observed when the youngest cohort reached their thirties (given they have not yet passed that age).

Are Women Catching Up to Men?

Table 5 presents country-specific estimates of the sex-age interaction predicting incident drug use, again derived from a discrete-time survival model. Sex was coded 1 for female and 0 for male in this model. Evidence of women becoming more similar to men in more recent cohorts would be indicated by a sex-by-cohort interaction OR that was <1.0: this would mean that the relative-odds of use among women compared to men were lower in older cohorts than younger cohorts.

Table 5.

Estimated Sex–Age Interaction Predicting Incident Use of Each Drug, by Country

graphic file with name pmed.0050141.t005.jpg

Open in a new tab

A negative interaction of exactly this sort was found fairly consistently across countries. All 17 interactions for alcohol use were <1.0 (14 of them significant at the 0.05 level). Seven of nine interactions were <1.0 in predicting tobacco use (five of them significantly so). The comparable counts were 13 of 15 interactions (five significant) in predicting use of cannabis, and six of 11 (two significant) in predicting cocaine use. It is noteworthy that the interactions are less consistent for illegal drugs, although the pattern is in the same direction as for the legal drugs.

Table 6 examines the possibility that given the sex-age interactions presented in Table 5 above, that younger age groups may no longer have sex differences in risk of incident use (given that the differences were particularly marked among older age groups). Two columns are shown for each drug type: one for the entire sample, and one for those aged 18–29 y, each being the ratio of cumulative incidence estimates according to sex (in any given year of life), derived from discrete-time survival models. As can be seen, for the overall sex association, there was an almost universal association across countries and drug types, whereby women were less likely than men in any given year of life to initiate drug use of all kinds. Among those aged 18–29 y, however, less consistent sex effects were found. In the European countries, there was no effect of sex on the likelihood of initiating alcohol use at any given age (with the exception of Italy and Ukraine, where females were still slightly less likely). This pattern was also found for cannabis and to a lesser extent cocaine initiation in these countries. In Japan, China, and New Zealand, no sex effect existed for alcohol initiation among this youngest age group. In the remaining countries, there remained a reduction of risk of incident drug use (of all types) among females relative to males (Table 6).

Table 6.

Estimated Association between Sex and Incident of Use of Each Drug among the Total Sample, and among Young Adults (18–29 y), by Country

graphic file with name pmed.0050141.t006.jpg

Open in a new tab

Sociodemographic Correlates of Lifetime Use

Finally, Table 7 presents pooled analyses of associations between drug use and six core demographic variables, with covariate adjustments via multiple logistic regression terms for each country (results from this pooled analysis are consistent with associations observed in country-level analyses; details of country-specific data are available upon request). Some demographic variables were consistently related to drug use of all kinds (Table 7). In particular, males were more likely than females to have used all drug types and younger adults were more likely than older adults to have used all drug types. As noted above, these two associations were consistent across drug types and countries. Income was positively related to lifetime use of both legal and illegal drugs. Marital status was also related to tobacco, cannabis, and cocaine use, but not alcohol use, (with the never married and previously married having higher odds of lifetime use than the currently married). Although education was related to drug use, the relationship was not consistent across drug types. Education was positively related to lifetime alcohol use, but negatively related to lifetime tobacco use, while education was unrelated to lifetime illegal drug use.

Table 7.

Bivariate and Adjusted Associations between Selected Variables and Drug Use. Pooled Data from the World Mental Health Surveys

graphic file with name pmed.0050141.t007.jpg

Open in a new tab

Discussion

Globally, drug use is not distributed evenly. In general, the US had among the highest levels of use of all drugs. Much lower levels were observed in lower income countries in Africa and the Middle East, and lower levels of use were reported in the Asian locales covered.

These variations cannot be regarded as static: there was greater drug involvement among younger than older adults in all countries, suggesting that drug use has and may continue to change over historical time. Interestingly, there was also evidence to suggest that male-female differences in risk of initiating drug use may be changing in more recent birth cohorts. This change was a consistent finding across countries, suggesting that a general shift may be occurring with respect to the traditional sex differences so often documented with drug use.

Consistent trends were also documented with respect to the period of risk for initiation of drug use. In most countries, the period of risk for initiation of use was heavily concentrated in the period from the mid to late teenage years; there was a slightly older and more extended period of risk for illegal drugs compared to legal drugs. Analyses of possible intercohort differences in risk of initiation suggested not only that the levels of illegal drug use were higher, but also that in more recent cohorts, the period of risk was extending further into adulthood. This extension of the period of risk has implications for drug use prevention efforts, which often focus upon adolescents and do not actively target young adults. Clearly, for illegal drugs, there continues to be a window of risk of initiation of illegal drug use that persists well beyond that of most commonly targeted ages.

Legal and illegal use of drugs was most strongly associated with age, sex, and income. Higher income was associated with a greater likelihood of drug use for all drug types examined, which is perhaps not surprising given that drug use requires disposable income. Relationship status was linked to illegal (but not legal) drug use: both cocaine and cannabis use were more likely among persons who had never been married or previously been married. These associations remained statistically robust after adjustment for age, sex, and the other variables considered here. These associations are consistent with previous research in the developed countries, which has linked illegal drug use with an individual's marital status.

The use of drugs seems to be a feature of more affluent countries. The US, which has been driving much of the world's drug research and drug policy agenda, stands out with higher levels of use of alcohol, cocaine, and cannabis, despite punitive illegal drug policies, as well as (in many US states), a higher minimum legal alcohol drinking age than many comparable developed countries. The Netherlands, with a less criminally punitive approach to cannabis use than the US, has experienced lower levels of use, particularly among younger adults. Clearly, by itself, a punitive policy towards possession and use accounts for limited variation in nation-level rates of illegal drug use.

Limitations

This study has limitations. We were limited to inclusion in the study of the countries that had the resources and interest in being involved in this exercise. Considerable effort was expended to ensure that countries from every region worldwide were represented in the consortium, but we did not succeed in getting as much coverage as we would have liked; for example, French-speaking West Africa is not represented in the WMHS. This has meant that the current data do not represent every world region sufficiently.

One limitation that certainly may have affected this study's estimates is the level of survey participation and nonresponse, which varied across countries. This is a likely source of underestimation of illegal drug involvement (and perhaps alcohol and tobacco use), but not necessarily a source of bias with respect to estimated associations with other variables [42]. In the WMHS initiative, we compensated for survey nonresponse via poststratification adjustments, but this approach is limited if nonresponse is associated with drug use in other ways.

A considerable strength of the WMHS initiative is that the population survey research approach generally has been held constant. Each respondent has been sampled via advanced population survey methods, has been presented with the same type of survey introduction, and has completed a highly structured and standardised field survey interview assessment using the same questions in each country, in accord with standardised translations. One exception to this was France, whose sampling frame only included households that had a telephone, although interviewing itself was conducted face-to-face.

Nonetheless, in a cross-national study such as this one, there might be differential social stigma and legal practices in each country that might affect self-reported drug use. Attempts were made to ensure that truthful, honest answers were provided by participants in these surveys in four major ways. First, pilot testing in each country was carried out to determine the best way to describe study purposes and auspices in order to maximize willingness to respond honestly and accurately. Second, in countries that do not have a tradition of public opinion research, and where the notions of anonymity and confidentiality are unfamiliar, we contacted community leaders in sample sites to explain the study, obtain formal endorsement, and have the leaders announce the study to community members and encourage participation. The announcements were most typically made by religious leaders as part of their weekly sermons, although there are other cases, such as the formal community leaders in each neighbourhood in Beijing and Shanghai, where secular community leaders, who were given presents by the study organizers, made formal announcements and encouraged members of their neighbourhood to participate in the survey. Third, interviewers were centrally trained in the use of nondirective probing, a method designed to encourage thoughtful honest responding. Finally, especially sensitive questions were asked in a self-report format rather than an interviewer-report format, although this could be done only for respondents who could read. These methods were not completely effective in removing cross-national differences in willingness to report though, so it is important to recognise the possible existence of remaining differences of this sort in interpreting cross-national differences in results.

In future cross-national surveys, bioassays of drug use might be included. In the interim, we must use population survey data such as these to complement macro-level taxation records that summarise only legal alcohol and tobacco use across countries. These data are reasonable for developed countries, but are likely to significantly underestimate consumption in some developing countries. The findings of this study are consistent, however, with data collected in some countries from other epidemiological studies, which have been consistent with the findings here. The large cross-national differences documented here are consistent with approximations of drug use levels given in the United Nations World Drug Report [43] and with country-specific research—for example, researchers have documented high levels of cannabis use and early onset alcohol use in New Zealand [44], early onset alcohol use in Europe, and cocaine use in the US [45].

It is important to note that we kept full age ranges because we did not wish to truncate age in 90% of countries in order to accommodate the few that had more restricted age ranges, but we also wished to retain those few surveys in the total. The presentation of estimates specific to the 18–29-y age range is unaffected by these differences however, and the same patterns of cross-country results was observed in that instance. We also do not have an “urbanicity” variable in most countries that could be considered comparable, so thus far in the WMHS papers we have not studied urbanicity in cross-national work.

Any cross-sectional survey research has limitations [46]. Some of the observed cohort differences might be traced to higher mortality among individuals in the older cohorts who began drug use at an early age. Nonetheless, we believe that differential mortality is unlikely to explain the rather large differences in cumulative incidence for illegal drug use across age groups observed: in the case of cannabis, with substantial age-related variations observed, there is no convincing evidence of substantial premature mortality [47]. Conversely, the evidence of tobacco-related premature mortality is substantial, but tobacco use showed the least prominent age-associated variation.

This study found very strong cohort differences in illegal drug use in particular. Although this may reflect actual cohort differences in drug use, they may also reflect response biases. Retrospective reporting of age of first drug use may be subject to error, given that respondents are being asked about events that, for older persons, may have occurred decades ago. Longitudinal studies of adolescents have found that estimates of the age of first use do tend to increase upon repeat assessment (i.e. as people age), but the rank ordering for different drugs remains the same [48–50].

It is unlikely that response or other biases completely account for the strong trends observed here. In this study, there were contrasting cohort trends across different drug types, suggesting that a uniform bias or pattern of “forgetting” did not apply. Similar birth cohort trends in age of initiation of illegal drug use have been observed in other epidemiological studies in the US [51,52] and Australia [53], some of which used data collected across time (rather than relying solely on retrospective reports; e.g., see [51]).

In this paper, we have examined the cumulative incidence (sometimes referred to as “lifetime prevalence”) of use—this includes both experimental and heavier use. We also focused upon cumulative incident use rather than past year prevalence of use. Detailed examinations of prevalent (past year) use, and of transitions to dependent use of these drugs, were beyond the scope of the present paper, but are of obvious interest and importance. These are the subjects of future work currently being undertaken.

Conclusions

This study presents novel data on the epidemiology drug use from representative, cross-national samples representing all regions of the world. Clear differences in drug use existed across the regions of the world, with the US estimated to have among the highest levels of both legal and illegal drug use among all countries surveyed. These differences may be closing in more recent birth cohorts, with higher levels of drug use seen among young adults across countries. Drug use is related to income, but does not appear to be simply related to drug policy, since countries with more stringent policies towards illegal drug use did not have lower levels of such drug use than countries with more liberal policies.

Acknowledgments

The surveys discussed in this article were carried out in conjunction with the World Health Organization World Mental Health (WMH) Survey Initiative. We thank the WMH staff for assistance with instrumentation, fieldwork, and data analysis. A complete list of WMH publications can be found at http://www.hcp.med.harvard.edu/wmh/.

Abbreviations

AOO

age of onset

OR

odds ratio

WHO

World Health Organization

WMH

World Mental Health

WMHS

World Mental Health Surveys

Footnotes

Author contributions. RCK, JCA, GdG, AK, JPL, JHO, and DJS designed the experiments/the study. LD, W-TC, NS, RCK, MA, RB, GdG, OG, AK, SK, JPL, MEMM, YN, JHO, JP-V, DJS, and JEW collected data or did experiments for the study. W-TC, RCK, JCA, GdG, YH, AK, and MEMM analysed the data. RCK, MA, RB, GdG, AK, JPL, and JHO enrolled patients. LD wrote the first draft of the paper. LD, W-TC, RCK, JCA, MA, RB, GdG, OG, YH, AK, SK, JPL, MEMM, YN, JHO, AP-M, JP-V, and JEW contributed to writing the paper.

Funding: This work has been supported by multiple awards. The work of LD and JCA has been supported by the National Institute on Drug Abuse (K05DA015799; R01DA016558); and the work by LD supported by the Australian Government Department of Health and Ageing, and an Australian National Health and Medical Research Council (NHMRC) Senior Research Fellowship. These activities were supported by the United States National Institute of Mental Health (R01MH070884), the John D. and Catherine T. MacArthur Foundation, the Pfizer Foundation, the US Public Health Service (R13-MH066849, R01-MH069864, and R01 DA016558), the Fogarty International Center (FIRCA R01-TW006481), the Pan American Health Organization, the Eli Lilly & Company Foundation, Ortho-McNeil Pharmaceutical, GlaxoSmithKline, and Bristol-Myers Squibb. The Chinese World Mental Health Survey Initiative is supported by the Pfizer Foundation. The Colombian National Study of Mental Health (NSMH) is supported by the Ministry of Social Protection. The ESEMeD project is funded by the European Commission (Contracts QLG5-1999-01042; SANCO 2004123), the Piedmont Region (Italy), Fondo de Investigaciœn Sanitaria, Instituto de Salud Carlos III, Spain (FIS 00/0028), Ministerio de Ciencia y Tecnología, Spain (SAF 2000-158-CE), Departament de Salut, Generalitat de Catalunya, Spain, Instituto de Salud Carlos III (CIBER CB06/02/0046, RETICS RD06/0011 REM-TAP), and other local agencies and by an unrestricted educational grant from GlaxoSmithKline. The Israel National Health Survey is funded by the Ministry of Health with support from the Israel National Institute for Health Policy and Health Services Research and the National Insurance Institute of Israel. The World Mental Health Japan (WMHJ) Survey is supported by the Grant for Research on Psychiatric and Neurological Diseases and Mental Health (H13-SHOGAI-023, H14-TOKUBETSU-026, H16-KOKORO-013) from the Japan Ministry of Health, Labour and Welfare. The Lebanese National Mental Health Survey (LEBANON) is supported by the Lebanese Ministry of Public Health, the WHO (Lebanon), Fogarty International, Act for Lebanon, anonymous private donations to IDRAAC, Lebanon, and unrestricted grants from Janssen Cilag, Eli Lilly, GlaxoSmithKline, Roche, and Novartis.The Mexican National Comorbidity Survey (MNCS) is supported by The National Institute of Psychiatry Ramon de la Fuente (INPRFMDIES 4280) and by the National Council on Science and Technology (CONACyT-G30544- H), with supplemental support from the PanAmerican Health Organization (PAHO). Te Rau Hinengaro: The New Zealand Mental Health Survey (NZMHS) is supported by the New Zealand Ministry of Health, Alcohol Advisory Council, and the Health Research Council. The Nigerian Survey of Mental Health and Wellbeing (NSMHW) is supported by the WHO (Geneva), the WHO (Nigeria), and the Federal Ministry of Health, Abuja, Nigeria. The South Africa Stress and Health Study (SASH) is supported by the US National Institute of Mental Health (R01-MH059575) and National Institute of Drug Abuse with supplemental funding from the South African Department of Health and the University of Michigan. The Ukraine Comorbid Mental Disorders during Periods of Social Disruption (CMDPSD) study is funded by the US National Institute of Mental Health (RO1-MH61905). The US National Comorbidity Survey Replication (NCS-R) is supported by the National Institute of Mental Health (NIMH; U01-MH60220) with supplemental support from the National Institute of Drug Abuse (NIDA), the Substance Abuse and Mental Health Services Administration (SAMHSA), the Robert Wood Johnson Foundation (RWJF; Grant 044708), and the John W. Alden Trust. The funders had no role in the design of this study, data collection and analysis, the decision to publish this manuscript, nor in preparation of the manuscript.

Competing Interests: The authors have declared that no competing interests exist.

References

  1. Ezzati M, Lopez A, Rodgers A, Vander Hoorn S, Murray C, et al. Selected major risk factors and global and regional burden of disease. Lancet. 2002;360:1347–1360. doi: 10.1016/S0140-6736(02)11403-6. [DOI] [PubMed] [Google Scholar]
  2. Rehm J, Room R, Monteiro M, Gmel G, Graham K, et al. Chapter 12. Alcohol use. In: Ezzati M, Lopez AD, Rodgers A, Murray R, editors. Comparative quantification of health risks: global and regional burden of disease attributable to selected major risk factors. 2nd edition. Geneva: World Health Organization; 2004. pp. 959–1108. [Google Scholar]
  3. Degenhardt L, Hall W, Lynskey M, Warner-Smith M. Chapter 13. Illicit drug use. In: Ezzati M, Lopez AD, Rodgers A, Murray R, editors. Comparative quantification of health risks: global and regional burden of disease attributable to selected major risk factors. 2nd edition. Geneva: World Health Organization; 2004. pp. 1109–1176. [Google Scholar]
  4. Ezzati M, Lopez AD. Chapter 11. Smoking and oral tobacco use. In: Ezzati M, Lopez AD, Rodgers A, Murray R, editors. Comparative quantification of health risks: global and regional burden of disease attributable to selected major risk Factors. 2nd edition. Geneva: World Health Organization; 2004. pp. 883–957. [Google Scholar]
  5. Kessler RC, Abelson J, Demler O, Escobar JI, Gibbon M, et al. Clinical calibration of DSM-IV diagnoses in the World Mental Health (WMH) version of the World Health Organization (WHO) Composite International Diagnostic Interview (WMHCIDI) Int J Methods Psychiatr Res. 2004;13:122–139. doi: 10.1002/mpr.169. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Kessler RC, Ustun TB. The World Mental Health (WMH) Survey Initiative Version of the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI) Int J Methods Psychiatr Res. 2004;13:93–121. doi: 10.1002/mpr.168. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Farr W. Vital statistics: memorial volume of selections from the reports and writings, 1885. Bulletin of the World Health Organization. 1885;78:88–95. [PMC free article] [PubMed] [Google Scholar]
  8. Dudfield R. A critical exmianation of the methods of recording and publishing statistical data bearing on public health; with suggestions for the improvement of such methods. Journal of the Royal Statistical Society. 1905;68:1–49. [Google Scholar]
  9. World Health Organization. International Classification of Diseases. 3rd revision. Geneva: World Health Organization; 1920. [Google Scholar]
  10. World Health Organization. Report of the International Pilot Study of Schizophrenia. Geneva: World Health Organization; 1973. [Google Scholar]
  11. Manski C, Pepper JV, Petrie CV. In: Informing America's policy on illegal drugs: what we don't know keeps hurting us. Council NR, editor. Washington (D.C.): National Academy Press; 2001. [Google Scholar]
  12. Clark WB. Operational definition of drinking problems and associated prevalence rates. Q J Stud Alcohol. 1966;27:648–668. [Google Scholar]
  13. Cahalan D, Room R. Monograph number 7. New Brunswick (New Jersey): Rutgers Center of Alcohol Studies; 1974. Problem drinking among American men. [Google Scholar]
  14. Room R. Cross-cultural research in alcohol studies: research traditions and analytical issues. In: Harford T, Towle L, editors. Cultural influences and drinking patterns: a focus on Hispanic and Japanese populations NIAAA Research Monograph number 19. Rickville (Maryland): National Institute of Alcohol Abuse and Alcoholism; 1988. pp. 9–40. [Google Scholar]
  15. Robins LN. Deviant children grown up: a sociological and psychiatric study of sociopathic personality. Baltimore (Maryland): Williams & Wilkins; 1966. [Google Scholar]
  16. Nicolaides-Bouman A, Wald N, Forey B, Lee P. International smoking statistics. Oxford: Oxford University Press; 1993. [Google Scholar]
  17. Room R, Janca A, Bennett L, Schmidt L, Sartorius N. WHO cross-cultural applicability research on diagnosis and assessment of substance use disorders: an overview of methods and selected results. Addiction. 1996;91:199–220. doi: 10.1046/j.1360-0443.1996.9121993.x. [DOI] [PubMed] [Google Scholar]
  18. Fitch C, Stimson GV, Rhodes T, Poznyak V. Rapid assessment: an international review of diffusion, practice and outcomes in the substance use field. Soc Sci Med. 2004;59:1819–1830. doi: 10.1016/j.socscimed.2004.02.028. [DOI] [PubMed] [Google Scholar]
  19. Ahlström S, Bloomfield K, Knibbe R. Gender differences in drinking patterns in nine European countries: descriptive findings. Subst Abus. 2001;22:69–85. doi: 10.1080/08897070109511446. [DOI] [PubMed] [Google Scholar]
  20. Obot IS, Room R, editors. Alcohol, gender and drinking problems: perspectives from low and middle income countries. Geneva: World Health Organization; 2003. [Google Scholar]
  21. Leifman H. A comparative analysis of drinking patterns in six EU countries in the year 2000. Contemp Drug Probl. 2002;29:501–548. [Google Scholar]
  22. Bloomfield K, Gmel G, Neve R, Mustonen H. Investigating gender convergence in alcohol consumption in Finland, Germany, the Netherlands and Switzerland: a repeated survey analysis. Subst Abus. 2001;22:39–53. doi: 10.1080/08897070109511444. [DOI] [PubMed] [Google Scholar]
  23. Dormitzer CM, Gonzalez GB, Penna M, Bejarano J, Obando P, et al. The PACARDO research project: youthful drug involvement in Central America and the Dominican Republic. Rev Panam Salud Publica. 2004;15:400–416. doi: 10.1590/s1020-49892004000600006. [DOI] [PubMed] [Google Scholar]
  24. World Health Organization. Global status report on alcohol. Geneva: World Health Organization; 2004. [Google Scholar]
  25. World Health Organization. Tobacco or health: a global status report. Geneva: World Health Organization; 1997. [Google Scholar]
  26. Warren C, Jones R, Eriksen M, Asma S, for the Global Tobacco Surveillance System (GTSS) collaborative group Patterns of global tobacco use in young people and implications for future chronic disease burden in adults. Lancet. 2006;367:749–753. doi: 10.1016/S0140-6736(06)68192-0. [DOI] [PubMed] [Google Scholar]
  27. Heeringa SG, Wells JE, Hubbard F, Mneimneh Z, Chiu WT, et al. Sample designs and sampling procedures. In: Kessler RC, Üstün TB, editors. The World Mental Health Survey Initiative volume one: patterns of mental illness in the WMH Surveys. New York: Cambridge University Press; 2008. In press. [Google Scholar]
  28. Alonso J, Angermeyer MC, Bernert S, Bruffaerts R, Brugha TS, et al. Sampling and methods of the European Study of the Epidemiology of Mental Disorders (ESEMeD) project. Acta Psychiatrica Scandinavica. 2004;420:8–20. doi: 10.1111/j.1600-0047.2004.00326. [DOI] [PubMed] [Google Scholar]
  29. Alonso J, Ferrer M, Romera B, Vilagut G, Angermeyer M, et al. The European Study of the Epidemiology of Mental Disorders (ESEMeD/MHEDEA 2000) project: rationale and methods. Int J Methods Psychiatr Res. 2002;11:55–67. doi: 10.1002/mpr.123. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Harkness J, Pennell B-E, Villar A, Gebler N, Aguilar-Gaxiola S, et al. Translation Procedures and Translation Assessment in the World Mental Health Survey Initiative. Kessler RCÜ, T.B., editor. The World Mental Health Survey Initiative volume one: patterns of mental illness in the WMH Surveys. New York: Cambridge University Press. 2008. In press.
  31. Solow R. Inequality and poverty. New York: WW Norton and Co; 1967. [Google Scholar]
  32. Simon GE, VonKorff M. Recall of psychiatric history in cross-sectional surveys: implications for epidemiologic research. Epidemiol Rev. 1995;17:221–227. doi: 10.1093/oxfordjournals.epirev.a036180. [DOI] [PubMed] [Google Scholar]
  33. Knauper B, Cannell CF, Schwarz N, Bruce ML, Kessler RC. Improving the accuracy of major depression age of onset reports in the US National Comorbidity Survey. Int J Methods Psychiatr Res. 1999;8:39–48. [Google Scholar]
  34. SAS Institute. SAS/STAT software: changes and enhancements, Release 8.2. Cary (North Carolina): SAS Publishing; 2001. [Google Scholar]
  35. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457–481. [Google Scholar]
  36. Halli SS, Rao KV, Halli SS. Advanced techniques of population snalysis. New York: Plenum; 1992. [Google Scholar]
  37. Efron B. Logistic regression, survival analysis, and the Kaplan-Meier curve. J Am Stat Assoc. 1988;83:414–425. [Google Scholar]
  38. Hosmer D W, Lemeshow S. Applied logistic regression. 2nd edition. New York: Wiley & Sons; 2000. [Google Scholar]
  39. Wolter KM. Introduction to variance estimation. New York: Springer-Verlag; 1985. [Google Scholar]
  40. Research Triangle Institute. SUDAAN Version 8.0.1. 7.5.3 edition. Research Triangle Park (North Carolina): Research Triangle Institute; 2002. [Google Scholar]
  41. Research Triangle Institute. SUDAAN: Professional Software for Survey Data Analysis. 8.01 edition. Research Triangle Park (North Carolina): Research Triangle Institute; 2002. [Google Scholar]
  42. Kessler RC, Little RJ, Groves RM. Advances in strategies for minimizing and adjusting for survey nonresponse. Epidemiol Rev. 1995;17:192–204. doi: 10.1093/oxfordjournals.epirev.a036176. [DOI] [PubMed] [Google Scholar]
  43. United Nations Office on Drugs and Crime. World drug report 2005. Geneva: United Nations; 2005. [Google Scholar]
  44. Fergusson DM, Horwood LJ. Early onset cannabis use and psychosocial adjustment in young adults. Addiction. 1997;92:279–296. [PubMed] [Google Scholar]
  45. Substance Abuse and Mental Health Services Administration Office of Applied Studies. Result from the 2002 National Survey on Drug Use and Health: national findings. Rockville (Maryland): Office of Applied Studies, National Household Survey on Drug Abuse Series H-22, Department of Health and Human Services Publication number SMA 03–3836; 2003. [Google Scholar]
  46. Wu LT, Korper S, Marsden M, Lewis C, Bray R. Use of incidence and prevalence in the substance use literature: A review. Rockville (Maryland): Substance Abuse and Mental Health Services Administration, Office of Applied Studies; 2003. [Google Scholar]
  47. Hall W, Degenhardt L, Lynskey M. National Campaign Against Drug Abuse Monograph number 44. Canberra: Australian Publishing Service; 2001. The health and psychological consequences of cannabis use. [Google Scholar]
  48. Engels R, Knibbe R, Drop M. Inconsistencies in adolescents' self-reports of initiation of alcohol and tobacco use. Addict Behav. 1997;22:613–623. doi: 10.1016/s0306-4603(96)00067-6. [DOI] [PubMed] [Google Scholar]
  49. Labouvie E, Bates M, Pandina R. Age of first use: its reliability and predictive utility. J Stud Alcohol. 1997;58:638–643. doi: 10.15288/jsa.1997.58.638. [DOI] [PubMed] [Google Scholar]
  50. Henry B, Moffitt T, Caspi A, Langley J, Silva P. On the “remembrance of things past”: a longitudinal evaluation of the retrospective method. Psychol Assess. 1994;6:92–101. [Google Scholar]
  51. Kerr W, Greenfield T, Bond J, Ye Y, Rehm J. Age-period-cohort influences on trends in past year marjiuana use in the US from 1984, 1990, 1995 and 2000 National Alcohol Surveys. Drug Alcohol Depend. 2008;86:132–138. doi: 10.1016/j.drugalcdep.2006.05.022. [DOI] [PubMed] [Google Scholar]
  52. Johnson RA, Gerstein D. Initiation of use of alcohol, cigarettes, marijuana, cocaine and other substances in US birth cohorts since 1919. Am J Public Health. 1998;88:27–33. doi: 10.2105/ajph.88.1.27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Degenhardt L, Lynskey M, Hall W. Cohort trends in the age of initiation of drug use in Australia. Aust N Z J Public Health. 2000;24:421–426. doi: 10.1111/j.1467-842x.2000.tb01605.x. [DOI] [PubMed] [Google Scholar]

Articles from PLoS Medicine are provided here courtesy of PLOS

RESOURCES