Prevalence and Correlates of Substance Use Among South African Primary Care Clinic Patients

Abstract

We aimed to assess prevalence and correlates of hazardous use of tobacco, alcohol and other drugs in a primary care population in Cape Town, South Africa. Stratified random sampling was used to select 14 of the 49 clinics in the public health sector in Cape Town, and every “nth” patient, with those ages 18–25 oversampled (N = 2,618). Data were collected from December 2003 through 2004, using the World Health Organization Alcohol, Smoking, and Substance Involvement Screening Test. Hazardous use of tobacco was most common, followed by alcohol and then other drugs. Hazardous tobacco use was associated with the 18–25 years age group, no religious involvement, high school completion, and higher stress. Hazardous alcohol use was associated with male gender, younger men, no religious involvement, employment, some high school education, and higher stress. Hazardous use of other drugs was associated with Colored (mixed) race (particularly among men), no religious involvement, employment, and stress. For all substances, women, particularly Black women, had the lowest rates of hazardous use. Although the study is cross-sectional, it does identify groups that may be at high risk of substance misuse and for whom intervention is urgent. Because prevalence of substance use is high in this population, routine screening should be introduced in primary care clinics.

The use of tobacco, alcohol, and other drugs is among the major contributors to the international burden of disease and injury, and this is particularly true of tobacco and alcohol in African countries (World Health Organization [WHO], 2002). In South Africa, tobacco and alcohol contribute significantly to rates of death and injury (Parry, 2005; Sitas et al., 2004). Cannabis and methaqualone are the most frequently identified illicit drugs in drug-related arrests, psychiatric diagnoses, and trauma patients (Parry et al., 2002), but treatment demand for cocaine, heroin, and methamphetamine use is increasing (Parry et al.; Parry, Myers, and Plüddemann, 2004).

Primary health care services provide an advantageous location for intervening in substance use (Whitlock, Polen, Green, Orleans, and Klein, 2004; Fleming et al., 2002). In South Africa, the public health sector is the main provider of primary care (Health Systems Trust, 2004), yet no data are available on prevalence of substance misuse among South African patients.

Risk factors are those associated with increased likelihood of negative health behaviors or outcomes, whereas protective factors are those that are associated with increased likelihood of positive health behaviors or outcomes, or with decreased likelihood of negative health behaviors or outcomes (Kirby, 2001). Although the association with disease occurrence may not be causal, such correlates are useful for identifying high-risk groups or factors that may be usefully targeted in intervention strategies. Risk and protective factors influencing substance use have not been much studied in the developing world (De Lima, Dunn, Novo, Tomasi, and Reisser, 2003).

Personal characteristics identified as risk factors may include age, race, and gender. In South Africa, the racial categories defined under apartheid—Black, Colored (mixed race), Indian, and White—have a long association in South Africa with access both to alcohol and to treatment (Mager, 2004). Different race groups may demonstrate different prevalence or responses to risk factors by gender (Flisher, Parry, Evans, Muller, and Lombard, 2003). Young adults who use substances may be more at risk for other health risks such as sexual risk behaviors, violence, and suicide (Flisher, Ziervogel, Chalton, Leger, and Robertson, 1996).

Protective factors include being married (Judd et al., 2002; Teesson, Dietrich, Degenhardt, Lynskey, and Beard, 2002), having children, and having religious involvement (Chen et al., 2004; Galen and Rogers, 2004), as well as higher education (Judd et al.; Teesson, Dietrich, Degenhardt, Lynskey, and Beard), employment (Judd et al.; Teesson, Dietrich, Degenhardt, Lynskey, and Beard), higher socioeconomic status (Flisher, Parry, Evans, Muller, and Lombard, 2003; Furr-Holden and Anthony, 2003; Judd et al.), manageable stressors (Goeders, 2004; Judd et al.), and not feeling depressed or anxious (Brady and Sinha, 2005). Both stressors and depression or anxiety are also associated with greater use of primary care services (Zantinge, Verhaak, and Bensing, 2005; Brantley et al., 2005).

In this paper we examine prevalence of hazardous use of alcohol, tobacco, and other drugs among demographic subgroups, in a large, representative sample of patients using the primary care service of the public health system in Cape Town, South Africa.

Methods

Sample

The study employed a multistage cluster, stratified sampling design. Consistent with other South African research, we stratified the 49 clinics providing primary care in Cape Town by race as defined under apartheid, because of the continuing association with health disparities and substance use (Mager, 2004; McIntyre et al., 2000). The population served by the public health sector is chiefly Black and Colored, and so we stratified clinics into those serving populations 80% or more Colored; 80% or more Black; and a more diverse population. The “Black” and “diverse” strata were approximately equal in size, and the “Colored” stratum 1.5 times the size of these. We randomly selected 14 clinics (proportional to the annual number of visits): 6 from the larger Colored stratum, and 4 from each of the proportionately smaller others.

On data collection days, we constructed a log of the age, race, and gender of all patients who came for a clinic visit. From this log, we randomly selected patients, except that we sampled every patient ages 18–24: This age group comes to clinics less frequently and is particularly at risk for substance use and HIV risk (Ward et al., 2005). The patient log data were later used to construct weights to estimate population-level statistics.

Measures

We developed the questionnaire in English, then translated it into Afrikaans and isiXhosa, and checked this by back-translation. The questionnaire was developed during a pilot study (Ward et al., 2005) in which items were modified for cultural relevance.

Risk and Protective Factors

Demographic data collected included age, race, gender, marital status, education, employment, and number of children. For socioeconomic status, we used items from the South African census measuring relative deprivation in urban areas (McIntyre et al., 2000): access to piped water, access to electricity at home, living in formal housing rather than in a shack or traditional dwelling, and employment status of the head of household. We later formed a scale representing socioeconomic status from the first three variables (Cronbach’s alpha = 0.68).

Patients were asked with which religion they affiliated and how often they took part in a religious activity; answers of “once or twice a month” or “weekly” were used to indicate involvement, whereas answers of “never” or “seldom” to indicate noninvolvement. To screen for anxiety and depression, we asked two questions from the Patient Health Questionnaire (Spitzer, Kroenke, and Williams, 1999). To assess stress, we asked which stressors had been experienced in the last year, from the International Classification of Primary Care, Second Edition (ICPC-2, 1998). The ICPC-2 lists 23 stressors that may be reasons for encounters, to which we added one: “Have you or anyone else in your household had an unplanned pregnancy?” Rates of births to teenagers are very high in South Africa (Bradshaw, Pettifor, MacPhail, and Dorrington, 2004).

Substance Use

We used the ASSIST (Alcohol, Smoking, and Substance Involvement Screening Test; WHO ASSIST Working Group, 2002) to assess prevalence of problematic substance use.

Specific scores were calculated for each substance where use was reported in the prior 3 months. These can be categorized as low- (including zero), medium- and high-risk use (except for tobacco, which can only be categorized high or low risk). Medium risk indicates problematic use, whereas high risk indicates high probability of dependence (Henry-Edwards, Humeniuk, and Ali, 2003). We dichotomized the risk category at the threshold of hazardous risk so that medium and high risk were coded “1” and low and no risk was coded “0.”

Procedures

Interviewers recruited patients as they waited for their medical visit. Patients were interviewed in private rooms by a trained research assistant and matched for gender and language. After the interview, respondents were given a list of referral resources, and those who had reported risky behavior were encouraged to seek help.

Data Analysis

The interview data were weighted to account for clustering, the over-quota of 18- to 24-year-olds, differential nonresponse rates (by gender, age, and race) within clinics, and the size of clinics proportional to the full population served by Cape Town’s Community Health Centres. Weights ranged from 0.02 to 12.1 (median = 0.34; interquartile range = 0.14–0.72).

A total of 6,135 patients were selected for the sample, and 2,618 were interviewed. There were several reasons why selected patients were not interviewed. First, 293 patients were missed because the interviewer with the necessary gender or language match was absent on that day; 136 patients did not speak any of the three languages included in the study; 58 patients were too ill to be interviewed; 30 were judged by the interviewer to be too cognitively impaired to consent; 8 patients were accompanied by a child and were not comfortable to be interviewed in the child’s presence; and 142 refused. By far the largest group (n = 2,866) was those patients who, when sought by the interviewers, had already seen a practitioner and left the clinic. These nonresponse rates appear to be primarily related to practical arrangements within the clinics, and it is unlikely that there is systematic bias in terms of the variables of interest.

Ten patients (eight Black women and two Black men) refused to continue the interview after it was initiated, and therefore had missing data for substance use questions and were excluded from analysis. Listwise deletion was used to manage other missing variables. Data from White and Indian respondents were excluded because of their very low representation in the sample. In each case, age group, gender, and socially defined race group were included first and, as the main variables of interest, were retained in each model. Each model started with a one-variable model, then a two-variable model, and so on. Variables found to be significant were also tested for interactions with other significant variables. Significant variables and interaction terms were included sequentially until a final “best” model was derived.

Results

Sample

Details of demographic characteristics of the sample are provided in Table 1. Respondents reported a mean of 4.9 stressors in the last year (range: 1–24), and most frequently mentioned were problems with money (53.5%), having sufficient food (29.9%), employment (42.6%), and health (31.5%).

Table 1.

Demographic characteristics of the sample

Demographic characteristic	N1	Demographic subcategory	n (unweighted)	Percentage (unweighted)	n (weighted)	Percentage (weighted)	95% CI
Age-group	2,618	Age 18–24	1072	41.0	225	8.6	[5.2,13.8]
Gender	2,618	Female	1490	56.9	1670	36.2	[34.3,38.2]
Race	2,618	Black	1122	42.9	1559	59.5	[31.5,82.5]
		Colored	1411	53.9	1024	39.1	[17.0,66.8]
		White	69	2.6	34	1.31	[0.2,10.2]
		Indian	16	0.6	1	0.06	[0.02,0.1]
Marital status	2,615	Married	935	35.8	1154	44.5	[40.3,48.9]
Has children?	2,611	Has children	1717	65.8	2143	82.7	[77.8,86.7]
Employment status	2,611	Employed	998	38.2	766	29.5	[19.4,42.2]
Education status	2,611	None/elementary only	665	25.5	966	37.3	[29.4,45.9]
		Some high school	1,543	59.1	1,374	53.0	[46.4,59.6]
		Graduated high school	403	15.4	251	9.7	[7.5,12.4]
SES	2,610	0	82	3.1	71	2.7	[1.4,5.3]
		1	190	7.3	217	8.4	[5.67,12.2]
		2	349	13.4	294	11.3	[7.8,16.2]
		3	1,989	76.2	2,010	77.6	[69.2,84.2]
Head of household’s employment status	2,608	Employed	1,449	55.6	1,144	44.4	[31.3,58.0]
Religious involvement status	2,618	Involved	1,471	56.2	1,272	48.6	[34.00,63.4]
Stressors	2,618	0–5	1,578	60.3	1,615	61.7	[57.2,66.0]
		6–10	666	25.4	677	25.9	[21.7,30.6]
		>10	374	14.3	326	12.5	[10.0,15.4]
Depression and anxiety status	2,591	Depressed or anxious	1,734	66.9	1,645	64.6	[57.5,71.1]

¹Ns vary from 2,603 to 2,618 because of missing data.

Prevalence and Correlates of Substance Use

Prevalence rates of hazardous (moderate- and high-risk) tobacco, alcohol, and other drug use are presented by demographic subgroups in Table 2.

Table 2.

Weighted prevalence of hazardous substance use in the last 3 months by subgroup (N = 2,608)

Subgroups	Tobacco		Alcohol		Other drugs
	Prevalence	95% CI	Prevalence	95% CI	Prevalence	95% CI
Age-group
18–24	32.1	[24.0,41.4]	16.7	[13.9,19.9]	8.4	[6.1,11.4]
Over 24	27.7	[20.5,36.3]	12.2	[10.5,14.2]	2.9	[2.1,4.1]
Gender
Male	43.1	[37.9,48.3]	26.5	[23.4,29.9]	7.4	[5.3,10.2]
Female	19.5	[10.7,32.9]	4.6	[3.2,6.5]	1.1	[0.5,2.4]
Race3
Black	19.7	[15.8,24.4]	12.3	[10.6,14.2]	2.5	[2.0,3.1]
Colored	40.6	[36.5,44.8]	12.9	[9.6,17.1]	4.9	[2.8,8.4]
Marital status
Unmarried	27.2	[19.6,36.3]	13.4	[10.9,16.3]	4.3	[3.0,6.1]
Married	29.3	[22.6,37.0]	11.6	[10.3,13.2]	2.3	[1.5,3.5]
Has children
No	28.8	[22.6,36.0]	24.5	[16.3,35.1]	8.1	[5.2,12.4]
Yes	28.0	[20.5,36.9]	10.1	[7.1,14.2]	2.4	[1.3, 4.4]
Employment status
Unemployed	25.7	[16.7,37.3]	9.8	[8.0,12.0]	2.6	[1.8,3.6]
Employed	33.9	[28.0,40.4]	19.2	[14.5,25.0]	5.4	[4.0,7.3]
Education status
None/elementary only	26.7	[18.3,37.3]	9.13	[6.4,12.9]	2.3	[0.8,6.8]
Some high school	31.3	[25.1,38.4]	15.3	[13.1,17.9]	4.4	[3.0,6.3]
Graduated high school or more	16.0	[9.6,25.4]	11.2	[7.2,17.0]	2.4	[1.4,4.1]
SES
0	40.0	[9.8,80.4]	38.1	[17.3,64.4]	7.7	[1.7,29.0]
1	14.8	[8.3,25.2]	5.4	[2.5,11.3]	0.8	[0.2, 2.3]
2	25.8	[17.5,36.1]	20.8	[10.2,37.8]	6.3	[3.1,12.6]
3	29.5	[22.7,37.4]	11.3	[8.8,14.4]	3.1	[2.2,4.5]
Head of household employed
No	25.0	[14.4,39.7]	8.7	[5.8,12.9]	1.7	[0.6, 5.0]
Yes	31.6	[27.0,36.6]	17.6	[13.2,23.1]	5.5	[3.6,8.4]
Religious involvement
No	30.3	[21.4,41.1]	17.9	[14.6,21.7]	4.8	[3.5,6.4]
Yes	25.8	[20.0,32.5]	7.1	[5.3,9.5]	2.0	[1.1,3.5]
Stressors
0–5	26.1	[20.5,32.6]	11.2	[9.6,13.1]	2.7	[1.9,4.0]
6–10	32.5	[22.1,45.0]	13.6	[9.8,18.6]	4.4	[3.0,6.4]
>10	28.9	[20.2,39.4]	17.7	[8.8,32.3]	5.0	[2.2,11.0]
Depression or Anxiety
No	24.4	[21.6,27.5]	11.2	[7.6,16.1]	1.4	[0.5,3.9]
Yes	30.9	[21.5,42.1]	13.7	[10.8,17.3]	4.6	[3.5,6.0]

¹Results for White and Indian participants were omitted from the race analysis, because the numbers of participants were too small to estimate prevalence accurately. All participants are included in the gender and age analyses.

In all subgroups, hazardous use of tobacco was most common, followed by alcohol and then other drugs. Among the other drugs, cannabis had highest prevalence: 2.04% (95% CI = 1.12%, 3.66%). For alcohol and other drugs (excluding tobacco) together, prevalence among those ages 18–24 was 22.1% (95% CI = 17.6%, 26.5%); whereas prevalence among those ages 25 and more was 13.6% (95% CI = 10.0, 16.4%).

An initial model for tobacco use revealed that an apparent race effect among women was due to very low prevalence among Black women: For Black women compared with other groups, the odds ratio is 0.04 (95% CI = 0.01, 0.22). Because of this, we analyzed the data separately for Black women and for other groups. Table 3 shows the odds ratio for tobacco use for the two groups. Hazardous use was associated with the 18- to 24-year-old age group among Colored people and Black men, but not among Black women. In both groups, hazardous use was associated with higher stress, and respondents who reported religious involvement and those with some high school education were less likely to meet criteria for hazardous use of tobacco.

Table 3.

Odds ratios and 95% confidence intervals for factors associated with hazardous tobacco use

	Black men and Colored men and women (n = 1,742)1	Black women (n = 559)
Age group
25+	1.00	1.00
18–24	1.57* (1.11,2.23)	2.21 (0.37,13.12)
Religious involvement
No	1.00	1.00
Yes	0.59*(0.37,0.94)	0.12+ (0.06,0.24)
Education
Graduated high school or higher	1.00
Some high school	0.55+ (0.38,0.81)	0.05* (0.00,0.58)
Stressors2	1.03*(1.00,1.05)	1.43+ (1.21,1.69)
Mental health
No depression or anxiety	1.00
Depression or anxiety	1.50 (0.96,2.34)	0.23 (0.02,2.81)

¹Because of the very low prevalence of tobacco use among Black women, data were analyzed separately for Black women and for the other groups.

^*p < .05.

⁺p < .01.

²Treated as a continuous variable.

Table 4 shows the results of the model examining hazardous alcohol use. Men were more likely to have hazardous alcohol use relative to women, and younger men reported lower rates relative to men aged 25 and older. Hazardous alcohol use was also associated with the respondents’ employment, with the respondent having some high school education (compared with less or more education), and with higher levels of stress. Religious involvement appeared to be a protective factor, and having children approached significance as protective. Aside from the interaction between race and gender, no other interactions terms were significant.

Table 4.

Odds ratios and 95% confidence intervals for factors associated with hazardous alcohol use (N = 2,311)

Risk/protective factor	Odds ratio (95% confidence interval)
Gender
Female, 25+	1.00
Male, 25+	7.85+ (4.58,13.47)
Female, 18–24	1.00
Male, 18–24	1.84* (1.11,3.07)
Age-group
25+, female	1.00
18–24, female	1.52 (0.67,3.44)
25+, male	1.00
18–24, male	0.36* (0.14,0.89)
Religious involvement
No	1.00
Yes	0.43+ (0.29,0.62)
Has children
No	1.00
Yes	0.27 (0.07,1.01)
Head of household employed
No	1.00
Yes	1.84 (0.92,3.66)
Respondent employed
No	1.00
Yes	1.46* (1.06,2.00)
Education
Elementary school or less	1.00
Some high school	2.09+ (1.46,2.99)
Graduated high school or more	0.89 (0.57,1.40)
Stressors1	1.13+ (1.08,1.19)

⁺p < .01.

^*p < .05.

¹Treated as a continuous variable.

Hazardous use of other drugs (Table 5) had very low prevalence among Black women—only 3 Black women reported any use of other drugs. Interaction terms were tested but were not significant; thus the effects of the risk factors were similar across gender and race, even though drug use prevalence differed. Both Colored men and women were significantly more likely to use other drugs at hazardous levels compared with Black men, and Colored men were most at risk. There was no age effect. Both employment and higher stress were significantly associated with increased use of other drugs. Religious involvement was identified as a protective factor.

Table 5.

Odds ratios and 95% confidence intervals for factors associated with hazardous use of other drugs (N = 2,312)

Risk/protective factor	Odds ratio (95% confidence interval)
Race and gender
Black men	1.00
Colored men	5.04+ (2.40,10.61)
Colored women	1.96* (1.08,3.56)
Age group
25+	1.00
18–24	1.88 (0.69,5.14)
Religious involvement
No	1.00
Yes	0.33+ (0.24,0.46)
Has children
No	1.00
Yes	0.25 (0.06,1.01)
Respondent employed
No	1.00
Yes	2.03+ (1.31,3.14)
Education
Elementary school or less	1.00
Some high school	1.68 (0.39,7.18)
Graduated high school or more	0.43 (0.16,1.19)
Stressors1	1.17* (1.04,1.31)

⁺p < .01.

^*p < .05.

¹Treated as a continuous variable.

Discussion

In terms of prevalence of substance misuse, tobacco was most likely to be used at hazardous levels, followed by alcohol and then other drugs. Younger age was associated with hazardous tobacco use, but only among men was it associated with hazardous alcohol use, and it was not significantly associated with use of other drugs. Women, particularly Black women, had the lowest rates of use of all substances, and Colored men and women had higher rates than Black men for use of drugs other than tobacco and alcohol.

An important finding is the very low prevalence of hazardous use of substances among Black women. This may be because women in Southern Africa face strong cultural proscriptions against substance use (Mphi, 1994). They may also have less disposable income, be less likely to report their substance use, or be laggards in the diffusion of innovations (Rogers, 1995). Further study should be conducted to identify protective factors that might keep rates low in this group and be integrated into interventions for other groups.

Consistent with international literature (WHO, 2002), we found that men and young people are most at risk for substance use. However, Black men in our sample were at much lower risk for use of drugs other than alcohol and tobacco than Colored men and women. One possible factor influencing this finding may be urbanization. Cape Town is a large urban area with a great deal of in-migration, particularly among Black people, from very rural areas (Statistics South Africa, n.d.). The Black population of the Western Cape thus tends to be less urbanized than other groups, and urbanization may increase risk for substance misuse (Flisher and Chalton, 2001; Judd et al., 2002). It is also possible that there was a bias against reporting substance use in this group. Further study should seek to identify whether urbanization and other factors might influence either substance use or its reporting.

Our finding that not finishing high school is associated with hazardous alcohol use strengthens the urgency for intervening with youth. Approximately 60% of South African youth who begin high school do not graduate (Department of Education, 2003), and other studies suggest an association between substance use and high school dropout (Flisher and Chalton, 1995; Krohn, Lizotte, and Perez, 1997). By contrast, those who had graduated from high school were at higher risk for reporting tobacco use. Since tobacco use is typically initiated during adolescence (U.S. Department of Health and Human Services, 1994); however, intervening at high school may also be important here.

Our finding that employment increased risk for hazardous use of alcohol and other drugs is at odds with the literature suggesting that unemployment is a risk factor for substance misuse (Judd et al., 2002; Teesson, Dietrich, Degenhardt, Lynskey, and Beard, 2202). But in this relatively impoverished population, it may well be that having some income makes it possible to buy alcohol or drugs that otherwise simply could not be accessed. Poverty is one of South Africa’s main development challenges: A quarter of South Africans have an income under $14 a month (Drimie and Mini, 2003); many of the clinics in which we worked were situated in the most deprived areas of the City of Cape Town (Noble et al., 2006). Unemployed people in this population may therefore be using their resources for sheer survival rather than for purchasing substances. This should be investigated in future studies, but our finding does both identify another high-risk group and suggest that intervention may be strategically delivered via the workplace.

Consistent with the cross-national literature (Bazargan, Sherkat, and Bazargan, 2004; Galen and Rogers, 2004; Matthew et al., 1998), religious involvement was associated with lower prevalence of hazardous use of all substances. Religion may operate as a protective factor by discouraging drinking and improving coping and social support (Koenig, 2001). Our other consistent finding was that higher stress was associated with increased risk for substance use and misuse: It is clear that interventions to reduce stressors and to improve coping, are particularly important.

Study’s Limitations

Our study has several limitations. First, we relied on self-report data; thus the prevalence rates are subject to biases such as problems recalling drug use and social desirability of responses (Johnson, 2005). However, the ASSIST has been found to be reliable and acceptable for screening use internationally (WHO ASSIST Working Group, 2002), and to have validity similar to other established self-report instruments (Newcombe, Humeniuk, and Ali, 2005). Second, the timing of the study means that the more recent epidemic in methamphetamine use (Parry, Myers, and Plüddemann, 2004) is not reflected in our data. We also missed a large number of patients. However, this was most likely caused by factors related to the clinics’ ability to process patients and not to the variables under study, and so is unlikely to have introduced any systematic bias. Finally, this is a cross-sectional study and so does not address causal relationships.

Despite these limitations, it is clear that there is a high prevalence of at least tobacco and alcohol use, among patients attending primary care clinics in the public health system in South Africa. Routine screening for substance use (particularly alcohol and tobacco) should be integrated into every clinic visit, and treatment made easily available through the public health system. Interventions for high-risk groups are urgently needed, and future studies should seek to elucidate the mechanisms by which particular risk and protective factors operate.

Glossary

Prevalence

The proportion of individuals in a population who have a disease, disorder or other characteristic at a specific instant in time

Multistage sampling design

This describes a situation where sampling takes place in a number of stages, involving different sampling strategies (e.g., stratified sampling followed by cluster sampling)

Cluster sampling design

This involves the selection of affiliated units, or clusters (e.g., patients attending specific health facilities) for study

Stratified sampling design

This describes the situation where there is division or stratification of a population into relatively homogenous groups called strata, and the selection of samples independently in each of the strata

Reason for encounter

The reason the patient gives for seeking care at the clinic

Biographies

Catherine L. Ward, Ph.D., is a Senior Research Specialist in the Child, Youth, Family and Social Development Research Programme at the Human Sciences Research Council, South Africa. She is a clinical-community psychologist, and her areas of research interest include exposure to violence and its consequences for development and well-being, including risk behaviors such as substance misuse.

Jennifer R. Mertens, Ph.D., is a Staff Scientist in the Drug and Alcohol Research Team in the Division of Research, Kaiser Permanente Medical Care Program in Oakland, California. Her research interests include medical consequences and correlates of substance use problems and substance use interventions in primary care.

Alan J. Flisher, Ph.D., F.C. Psych. (S.A.), is a Professor of Psychiatry and Mental Health at the University of Cape Town (UCT); Head of the Division of Child and Adolescent Psychiatry at UCT and Red Cross War Memorial Children’s Hospital; Director of the Adolescent Health Research Institute at UCT; and Professor II at the Research Centre for Health Promotion at the University of Bergen, Norway. His main research interests are adolescent mental health and risk behavior, and mental health services policy and planning.

Graham F. Bresick, M.B. Ch.B., M.P.H., is a senior lecturer in the School of Public Health and Family Medicine, Health Sciences Faculty, University of Cape Town, where he teaches family medicine and motivational interviewing to undergraduate and postgraduate medical students and convenes the final year MB. Ch.B family medicine clerkship in district-based health services. His current research interests include the role of family medicine principles and motivational interviewing in improving primary care service delivery.

Stacy A. Sterling, MSW, MPH, is a Senior Research Project Manager, with the Drug and Alcohol Research Team at the Division of Research, Northern California Kaiser Permanente Medical Care Program. Her work focuses on developing health policies and improving clinical practice to increase treatment access and improve outcomes for adolescents and women, as well as disseminate findings into clinical practice.

Francesca Little, PhD., is a Senior Lecturer in the Department of Statistical Sciences, University of Cape Town, South Africa. She specializes in Bio-Statistics and her main research interests include the application of statistical techniques to medical research. She is consultant statistician to the South-East African Combination Anti-malarial Therapy (SEACAT) evaluation, the South African Tuberculosis Vaccine Initiative, Institute of Infectious Diseases and Molecular Medicine, the Cape Town data centre for the International Epidemiological Databases to Evaluate Aids (IeDEA), and the NIH funded study on sexual risk and substance user behavior.

Constance M. Weisner, Dr. P.H., M.S.W, is an Investigator in the Division of Research, Kaiser Permanente Medical Care Program, Drug and Alcohol Research Team in Oakland, California, and a Professor in the Department of Psychiatry, University of California, San Francisco. Her research interests include alcohol and drug user treatment access, outcomes, and costs.