http://informahealthcare.com/ada ISSN: 0095-2990 (print), 1097-9891 (electronic) Am J Drug Alcohol Abuse, 2014; 40(3): 213–224 ! 2014 Informa Healthcare USA, Inc. DOI: 10.3109/00952990.2014.892950

Gender and racial/ethnic differences in patterns of adolescent alcohol use and associations with adolescent and adult illicit drug use Maria R. Khan, PhD1, Charles M. Cleland, PhD2, Joy D. Scheidell, MPH1, and Amanda T. Berger, PhD3 Department of Epidemiology, University of Florida, Gainesville FL, 2NYU College of Nursing, New York, NY, and 3Department of Fertility and Family Structure, Child Trends, Washington, DC, USA Abstract

Keywords

Objectives: The study objective was to use latent class analyses (LCAs) to identify gender- and racial/ethnic-specific groups of adolescent alcohol users and associations between alcohol use group and adolescent and adulthood illicit drug use in a nationally-representative US sample. Methods: We used Wave I (1994–1995, adolescence) of the National Longitudinal Study of Adolescent Health to conduct LCAs by gender and race/ethnicity and measure associations between class membership and Wave I and Wave III (2001–2002, young adulthood) drug use. Participants included white (n ¼ 9548), African American (n ¼ 4005) and Hispanic (n ¼ 3184) participants. LCAs were based on quantity and frequency of adolescent alcohol use; physiological and social consequences of use; and peer use. Results: Males and females were characterized by different alcohol use typologies and consequences. Males in the highest severity class (i.e. drank both heavily and frequently) experienced disproportionate risk of alcohol-related consequences compared with abstainers and other alcohol-using groups. Females who drank heavily when drinking even if only occasionally, experienced high risk of alcohol-related consequences. Substantial proportions of males reported diverse alcoholrelated problems, whereas females most commonly reported alcohol-related problems with dating and sexual experiences. Though levels of alcohol use and report of problems associated with use were higher among white versus minority populations, other racial/ethnic differences in patterns of alcohol use were minimal. Classification in any drinking class was a strong risk factor for adolescent and adulthood illicit drug use, with heavy drinkers at greatest risk of drug use. Conclusions: Gender-specific adolescent alcohol and substance use prevention programs are warranted.

Adolescents, alcohol, illicit drug use, latent class analysis, racial/ethnic, young adulthood

Introduction Alcohol is the most commonly used substance among adolescents in the United States (US) (1); 72.5% of high school students have tried alcohol, 41.8% currently use alcohol, and 24.2% report heavy or binge drinking (2). Adolescent alcohol use negatively influences cognitive, neurological, and psychosocial development (3–8), which may lead to inhibited judgment and increased impulsivity (9–11) and involvement in high-risk behaviors and environments (12,13). Further, adolescent alcohol use is associated with negative physical and psychosocial outcomes such as an increased risk for exhibiting violence and aggression (14,15), poor school performance (16), unintended injuries, homicide, and suicide (17). Adolescent alcohol use also is strongly associated with use of other illicit substances both during adolescence (18,19) and in adulthood (20–22).

Address correspondence to Maria R. Khan, PhD, Assistant Professor, Department of Epidemiology, University of Florida, 2004 Mowry Rd., PO Box 100231 Gainesville FL 326211, USA. Tel: +1 352 273 5366. Fax: +1 352 573 5365. E-mail: [email protected]

History Received 29 April 2013 Revised 5 February 2014 Accepted 5 February 2014 Published online 21 April 2014

A one-dimensional measure of alcohol use may not accurately identify problem groups. It is accepted that a more appropriate analytic approach involves identifying groups characterized by multiple dimensions of use (23). Latent class analysis (LCA), a technique that captures the heterogeneity in a population by identifying groups of individuals that are homogenous with respect to a set of observed variables (24), has been employed to identify groups of adolescent alcohol users (25–27). A number of prior LCA studies have identified the latent class structure of adolescent alcohol use in US subpopulations, including current drinkers, college students, and clinical samples, such as those in addiction treatment centers (18,27–29), but there is limited research on typologies of adolescent alcohol use using nationally-representative data (30). In addition, there is limited research assessing whether these typologies differ by race/ethnicity in general-population US samples, though evidence points to race differences in adolescent alcohol use (30,31). Dauber et al. (2009) examined alcohol use typologies among white and African American adolescent girls using the public use dataset of the National Longitudinal Study of Adolescent Health (Add Health), a data source which contains

20 14

Am J Drug Alcohol Abuse Downloaded from informahealthcare.com by University of Laval on 06/30/14 For personal use only.

1

Am J Drug Alcohol Abuse Downloaded from informahealthcare.com by University of Laval on 06/30/14 For personal use only.

214

M. R. Khan et al.

half the observations of the full Add Health (30). Their study highlighted differences in underlying latent class structure comparing white and African American girls and pointed to a need to conduct an alcohol use LCA in the full Add Health dataset and that considered gender differences and that examined the three major racial/ethnic groups in the US: white, African American, and Hispanic. To our knowledge, no prior study has summarized latent classes of alcohol users in a Hispanic sample, an important research gap given this sub-population represents the largest minority population in the US (32). The purpose of the current study was to use Waves I (1994–1995; adolescence) and III (2001–2002; young adulthood) of the complete sample of the National Longitudinal Study of Adolescent Health (Add Health) to conduct LCAs to identify adolescent alcohol use groups based on indicators of frequency, severity, and consequences of alcohol use, by race/ethnicity (white, Hispanic, and African American) and gender, and to measure associations between latent class membership and both adolescent and young adult illicit drug use. By identifying groups of adolescent alcohol users by race and gender and by examining associations between latent class membership and use of other drugs, it may be possible to identify the degree to which alcohol users – and which types of users – constitute priority populations for substance use prevention and treatment.

Methods Ethical approval for this research was obtained from the University of Florida Institutional Review Board. Data Add Health is a nationally-representative cohort study designed to investigate health factors from adolescence into adulthood (33–39). During Wave I of the survey (1994–1995), data were collected from adolescents (age range: 11–21), schools, and parents; adolescents provided information on characteristics including alcohol use. During Wave III (2001– 2002), participants were re-interviewed as adults, and behaviors including substance use were assessed. Using Wave I data, we conducted LCAs to identify adolescent alcohol use latent classes by gender and by race (white n ¼ 9548; Hispanic, n ¼ 3184; African American, n ¼ 4005). We then used Wave I latent class membership to assess associations with Wave I and III substance use. Measures To identify latent classes of adolescent alcohol users, the following indicators of adolescent alcohol use were used: Alcohol use Frequency of drinking (‘‘During the past 12 months, on how many days did you drink alcohol?’’) was rescaled to: 0 ¼ never; 1 ¼ once a month; 2 ¼ two to three times a month; 3 ¼ once a week or more. Frequency of being drunk (‘‘Over the past 12 months, on how many days have you gotten drunk or ‘very, very high’ on alcohol?’’) and Frequency of binge drinking (‘‘Over the past 12 months, on

Am J Drug Alcohol Abuse, 2014; 40(3): 213–224

how many days did you drink five or more drinks in a row?’’) were rescaled to: 0 ¼ never binging/getting drunk, 1 ¼ binging/getting drunk less than monthly, and 2 ¼ binging/getting drunk once a month or more. Quantity of drinking (‘‘Think of all the times you have had a drink during the past 12 months. How many drinks did you usually have each time?’’) was rescaled to: 0 ¼ no drinks, 1 ¼ one drink, 2 ¼ two to three drinks, and 3 ¼ four or more drinks per occasion. Physiological consequences of alcohol use Two indicators of physiological consequences of alcohol were used: Hangovers (‘‘Over the past 12 months, how many times were you hung over?’’) and Getting sick to stomach/throwing up (‘‘Over the past 12 months, how many times were you sick to your stomach or threw up after drinking’’). Each was rescaled: 0 ¼ never drank; 1 ¼ drank but the problem was not reported; and 2 ¼ drank and the problem was reported. Social consequences of alcohol use These indicators measured whether adolescents had social problems due to alcohol and included: Problems with parents (‘‘You got into trouble with your parents because you had been drinking’’); Problems with school (‘‘You’ve had problems at school or with school work because you had been drinking’’); Problems with friends (‘‘You had problems with your friends because you had been drinking’’); Problems with dating (‘‘You had problems with someone you were dating because you had been drinking’’); Getting into regrettable situations (‘‘You did something you later regretted because you had been drinking’’); Getting into regrettable sexual situations (‘‘Over the past 12 months, did you get into a sexual situation that you later regretted because you had been drinking’’); and Getting into physical fights (‘‘Over the past 12 months, did you get into a physical fight because you had been drinking’’). Each variable was rescaled: 0 ¼ never drank; 1 ¼ drank but the problem was not reported; and 2 ¼ drank and the problem was reported. Peer alcohol associations The peer alcohol indicator indicated whether the respondent drank with friends. Friends who drink (‘‘Of your 3 best friends, how many drink alcohol at least once a month?’’) was rescaled to: 0 ¼ never drank; 1 ¼ drank but no best friends drank; and 2 ¼ drank and one or more best friends drank monthly. Adolescent and young adult correlates of latent class membership Sociodemographic and behavioral characteristics We examined associations between latent class membership and the following: age; maternal education measured by Wave I self-report if the mother was interviewed, otherwise by adolescent’s report; Wave III low functional income status in the past year (mother or father on public assistance); and high levels of delinquency, defined as endorsement of six or seven of the following seven items: damage to others’ property, stealing something worth less than $50, going into a house or building to steal, use or threat of use of a weapon, selling

Adolescent alcohol classes and adulthood drugs

DOI: 10.3109/00952990.2014.892950

drugs, stealing something worth more than $50, taking part in a fight between groups of friends. Adolescent illicit drug use We examined associations between class membership and use of marijuana given its high prevalence of use among adolescents (Wave I). We also examined associations between class membership and use of marijuana and cocaine, common illicit substances in young adulthood (Wave III). Respondents were asked whether they had ever used each substance: 0 ¼ never used and 1 ¼ used.

Am J Drug Alcohol Abuse Downloaded from informahealthcare.com by University of Laval on 06/30/14 For personal use only.

Analyses Using Stata (version 13.1), we calculated descriptive statistics and compared gender and race subgroups on variables characterizing alcohol use and physical and social consequences. All analysis took into account the survey sampling design, including specification of sampling weights, strata, and primary sampling units. Using Mplus software for survey data (Version 7.11) (40), LCAs were conducted to identify classes of adolescent alcohol use based on the aforementioned indicators of alcohol use. Given the expected association with patterns of alcohol use and alcohol use consequences, age was included as covariate in all LCAs. Six gender and racial/ethnic subgroups were defined (white males and females, Hispanic males and females, and African American males and females). We estimated two, three, four, five, and six class solutions within each sub-population. The optimal number of classes for each group was identified using multiple fit indices including the adjusted Bayesian Information Criterion (BIC) and Akaike’s Information Criterion (AIC), for which low values are desired, and the Lo-Mendell-Rubin adjusted likelihood ratio test which compares the fit of a k-1 class model with the fit of a k class model, with the p-value indicating whether the improvement in fit due to adding another class is statistically significant; when the test is not statistically significant, the k1 class model is preferred over the k class model. We also considered class prevalence, preferring classes with 5% or more of the sample for improved reliability of estimates. For example, because in most sub-populations the six class solution yielded classes that were composed of 55% of the population, the six class models were not evaluated further. Finally, model fit also was determined by class interpretability and entropy, a measure of fit with higher values representing better classification. We included those who abstained from alcohol use during adolescence in the LCAs because we wished to conduct nationally-representative analyses; hence we sought to include all sub-groups in the US general population. Further, including abstainers allowed us to estimate how the odds of illicit drug use increased relative to minimal risk. Preliminary analyses indicated the remaining drinking classes were very similar with and without abstainers included in the models. In Mplus, we extended the LCAs to include drug use variables as additional indicators in order to measure adjusted associations (odds ratios [ORs] and 95% confidence intervals [CIs]) between latent class membership and use of illicit drugs in adolescence and young adulthood. Separate analyses

215

were conducted for adolescent (Wave I) and young adult (Wave III) drug use. Relationships between latent classes and alcohol indicators were constrained to be the same in models with drug use variables as they were in models with alcohol indicators only. All analyses controlled for poverty, high delinquency level, and maternal education. Additionally, in analyses examining associations between adolescent alcohol latent class and young adulthood marijuana and cocaine use, adolescent illicit drug use (marijuana, cocaine, inhalants, and injection drugs) was controlled. All LCAs in Mplus, including estimation of adjusted odds ratios relating alcohol latent classes to adolescent and young adulthood drug use, took into account the survey sampling design, including specification of sampling weights, strata, and primary sampling units.

Results Race differences in adolescent alcohol use indicators Whites and Hispanics were significantly more likely than African Americans to have used alcohol in the previous year (whites: 52%, Hispanics: 50%, African Americans: 35%) and to drink frequently, indicated by drinking two to three times a month (whites: 20%, Hispanic: 18%, African Americans: 12%) or drinking weekly (whites: 11%, Hispanic: 10%, African Americans: 8%) (Table 1). Whites and Hispanics were also more likely to drink a greater number of drinks per occasion, to get drunk, to experience physiological consequences, to regret an action taken due to drinking (whites: 16%, Hispanics: 14%, African Americans: 8%) and to experience social consequences of drinking including drinking-related problems with a parent (whites: 11%, Hispanics: 11%, African Americans: 5%), a dating partner (whites: 10%, Hispanics: 11%, African Americans: 5%), a sexual situation (whites: 7%, Hispanics: 7%, African Americans: 3%), and a friend (whites: 7%, Hispanics: 7%, African Americans: 3%). In addition, being a drinker and having a peer who drank at least once per month was more common among whites (44%) and Hispanics (39%) than African Americans (28%). Patterns of adolescent alcohol use Figure 1 presents profiles of each latent class on the indicator variables and Table 2 summarizes key race and gender differences among the classes. Among male adolescents, the number of classes and patterns of use were comparable for white, African American, and Hispanic males; the LCAs suggested a five class solution was the best fitting model across racial/ethnic groups. For white males, identification of the optimal number of classes was straightforward as the AIC, the adjusted BIC, and the Vuong-Lo-Mendell-Rubin likelihood ratio test all suggested a five class solution was optimal. However, among African American and Hispanic males, evidence from different fit indices was somewhat contradictory. For these groups, both the adjusted BIC and AIC values suggested the five class solution was appropriate. However, the adjusted Lo-Mendell-Rubin likelihood ratio test did not indicate the five class solution was a significant improvement over the four class solution. The concordance of the adjusted

216

M. R. Khan et al.

Am J Drug Alcohol Abuse, 2014; 40(3): 213–224

Table 1. Alcohol behaviors among white (n ¼ 9548), Hispanic (n ¼ 3184), and African American (n ¼ 4005) adolescents in the US*.

Am J Drug Alcohol Abuse Downloaded from informahealthcare.com by University of Laval on 06/30/14 For personal use only.

African White Hispanic American Total % % % % Past year drinking Never Once a month Two to three times a month Once a week or more Number of drinks/occasion Never drank One drink Two or three Four or more Past year binge drinking Never Once a month More than once a month Past year getting drunk Never Once a month More than once a month Past year physical/social problems Hung over Sick Regret about drinking Problems with parent(s) Problems with dating Regret sex Fighting Problems with friend(s) Problems with school Peer drinkers Never drank No best friends drink monthly One or more best friends drink monthly

47.69 32.00 9.76 10.55

50.13 31.81 7.62 10.44

64.40 23.11 4.66 7.82

52.14 29.84 8.14 9.88

48.08 9.90 14.35 27.68

50.65 10.08 15.59 23.68

65.04 11.80 12.71 10.45

52.61 10.39 14.19 22.82

68.47 17.96 13.57

70.57 17.61 11.82

86.60 7.07 6.33

73.20 15.29 11.51

65.74 71.13 22.11 18.97 12.15 9.90 from drinking 22.25 20.96 21.50 17.93 16.26 14.22 11.04 10.90 9.74 10.90 9.25 6.77 7.41 7.42 7.48 6.86 3.43 2.69

83.39 10.34 6.27

70.97 18.71 10.32

10.35 9.86 8.42 4.97 5.20 4.69 3.27 2.88 1.98

19.16 18.05 14.00 9.56 8.87 7.69 6.42 6.26 2.95

47.72 8.75 43.53

64.53 7.92 27.55

52.19 8.90 38.91

50.15 10.61 39.24

*Differences between whites and African-Americans were tested using the Chi-square statistic with adjustment for Add Health’s weighted sampling design. All differences were significant at p50.001.

BIC and AIC, with considerations of adequate prevalence levels and good interpretability of classes resulted in identifying the five class solution as most appropriate among minority males. Review of fit indices also suggested the five class solution was optimal among female adolescents in each racial/ ethnic group. The five class solution was chosen for white adolescent females given the fit statistics and the fact that each of the five groups had at least 5% of the sample. However, among African American and Hispanic women, the five class solution resulted in low prevalence (55%) in some alcohol latent classes and poor interpretability of some of the five groups; for these reasons the four class solution was considered optimal for African American and Hispanic females. For each subgroup of white, African American, and Hispanic males and females, entropy levels of the identified LCA solutions were very high (0.93 or higher), with individuals having a high probability of membership in one class and low probabilities of membership in all other classes, suggesting good model fit.

[abstainers (white: 49.1%, African American: 64.4%, Hispanic: 52.2%), experimenters (white: 16.6%, African American: 14.1%, Hispanic: 16.0%), occasional heavy drinkers (white: 16.8%, African American: 9.3%, Hispanic: 15.5%), frequent heavy drinkers with moderate risk of problems (white: 10.2%, African American: 5.1%, Hispanic: 7.8%), frequent heavy drinkers with high risk of problems (white: 7.3%, African American: 7.1%, Hispanic: 8.5%)]. Given these classifications, approximately 17.5% of white males, 12.2% of African American males, and 16.3% of Hispanic males were frequent heavy and/or problem drinkers. Among white, African American, and Hispanic males who were categorized as frequent heavy drinkers with a high risk of problems, there was a nearly universal endorsement of binge drinking and getting drunk at least once per month, and approximately 70–80% reported drinking at least weekly. Males in this category commonly endorsed feelings of regret about drinking and multiple social consequences of drinking, with white and Hispanic adolescent males in this class more likely to report most social problems than African American males (Table 3a). For example, one of the most common problems among males was problems with parents, being reported by higher levels of white (62%) and Hispanic (60%) males than African American males (46%). Likewise, problems with dating, with fighting, and with friends were commonly reported with higher levels observed among white (dating: 57%, fight: 57%, friends: 52%) and Hispanic (dating: 75%, fight: 65%, friends: 53%) males versus African American males (dating: 43%, fight: 48%, friends: 28%). White males also reported high levels of regrettable sexual experiences due to alcohol (55%), followed by African Americans (47%), and Hispanics (37%). The racial groups had comparable levels of problems with school due to alcohol us (white: 28%; African American: 24%; Hispanic: 23%). The LCAs indicated that among male adolescents, there was a group of frequent heavy drinkers with moderate risk of problems. Levels of weekly drinking, binge drinking, and getting drunk were comparable though reports of problems related to alcohol were lower compared with frequent heavy drinkers with high risk of problems. White occasional heavy drinkers drank less frequently than problem drinkers (78% were drinking once a month or less; 17% 2–3 times per month; 5% once per week); on a typical drinking occasion they had a comparable number of drinks as frequent heavy drinkers (approximately four or more drinks). Binging and getting drunk occurred less than once per month. These drinkers reported social consequences of alcohol use including problems with parents and dating, but at lower levels than observed among problem drinkers. Most occasional heavy drinkers had a best friend who drank at least once per month. Finally, experimenters had low frequency and quantity of drinking and were unlikely to report physiological or social consequences. However, the majority had friends who drank at least once per month.

Males

Females

Analyses yielded a five-class solution with similar patterns of use among white, African American, and Hispanic males

Analyses yielded a five-class solution among white females [abstainers (48.5%), experimenters (17.5%), occasional heavy

0

20

40

60

80

Percentage 100

0

20

40

60

80

100

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

Male: Experimenter

●

●

●

●

●

●

●

● ●

●

40

80

60

0

100

20

40

60

80

● ● ●

● ● ● ● ● ●

●

Indicator

● ●

20

0

40

60

80

We e k l Bin y Dri nk ge ing > 1 /M Dr o un k> 1/M 4+ D o r ink s/O Pa r cc e nt as ion S c ho o Fri l en d Da t i ng R e gre t Hu ng o Sic ver k Se x Fig hti n g rD rin k

●

●

●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

Male: Frequent Heavy − Moderate Risk of Problems

● ● ● ●

We e kl Bin y Dri n ge >1 king / Dr M o un k> 1/M 4+ Dr o in k s/ P a ren Occ as t ion Sc ho o Fri l en d D a tin g Re gre t Hu ng ov er S i c k Se x Fig

●

●

ht i n g rD rin k

Pe e

ing

100

80

White

60

40

Pe e

ing

100

Female: Occasional Heavy − High Risk of Problems

Figure 1. Profiles of alcohol latent classes by gender and race.

● ●

Male: Occasional Heavy

100

ing

100 Female: Occasional Heavy − Moderate Risk of Problems

80

20

Hispanic

60

40

0

100

●

20

80

60

●

●

● ●

●

Male: Frequent Heavy − High Risk of Problems

● ●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

Female: Frequent Heavy − High Risk of Problems

in

African−American

20

0

40

W e ek l Bin y Dri n ge >1 king / Dr M o un k> 1/M 4+ D o ir nk s/O Pa re cc n as t ion Sc ho ol F r i en d Da tin g Re g ret H u ng o Sic ver k Se x Fig h t i ng Pe er Dr ink

●

Dr

0

Dr

Dr

20

S

Dr

Female: Experimenter

er

er

er

0

go

er

Am J Drug Alcohol Abuse Downloaded from informahealthcare.com by University of Laval on 06/30/14 For personal use only.

W e ek ly D r Bin inkin g ge >1 Dr / M un 4+ o k> D r 1 ink /M s o /O cc as io Pa n r e n Sc t ho ol Fri en d Da tin g R e g Hu ret ng ov er Sic k Se x F i g hti ng ink Pe

ing

We e k ly Dr Bin inkin g g e> 1/M Dr un 4+ o k> Dr 1 in k s/O /Mo cc as ion P a ren Sc t ho ol Fr i e nd Da tin g Re g Hu ret ng o ve r Sic k S e Fig x h tin g ink ing Pe

We e kly D r Bin inkin g ge >1 Dr / M un 4+ o k> D 1 ir nk /M s o / O cc as io Pa n r e n Sc t ho ol F r i en d Da tin g Re g Hu ret ng ov er S i c k Se x F i g hti ng ink Pe

ing

W e ek ly Dr Bin inkin g ge > 1 Dr /M un 4+ o k D > r 1 ink /M s o /O cc as io Pa n r e n Sc t ho ol Fri e n d Da tin g Re g r e t ve r Hu n

ick

Se x hti ng

Fig

Pe

●

●

kin g

W e ek l Bin y Dri nk ge ing > 1 /M Dr o un k> 1/M 4+ D o ir nk s/ Pa ren Occ as t ion Sc h oo Fri l en d Da t i ng R e gre t H u ng o Sic ver k Se x Fig hti ng Pe er Dr ink

ing

Adolescent alcohol classes and adulthood drugs

DOI: 10.3109/00952990.2014.892950

217

218

M. R. Khan et al.

Am J Drug Alcohol Abuse, 2014; 40(3): 213–224

Table 2. Race and gender differences in patterns of adolescent alcohol use among adolescents in the US. Number of classes

Description of classes

Differences by race

Differences by gender

Abstainers; experimenters; occasional heavy drinkers; frequent heavy drinkers with low risk of problems; frequent heavy drinkers with high risk of problems

 Whites more likely than other groups to be drinkers including heavy drinkers.  White frequent heavy drinkers had higher risk of social problems than Hispanic/African Americans frequent heavy drinkers.  Hispanic frequent heavy drinkers had higher risk of problems with dating and fighting than white/African Americans frequent heavy drinkers.  African American frequent heavy drinkers had lower risk of most social problems than white/Hispanic frequent heavy drinkers.

Among males, frequent heavy drinkers had disproportionate vulnerability to a broad range of alcohol-related social problems compared with other drinking groups and abstainers.

5

Abstainers; experimenters; occasional heavy drinkers with moderate risk of problems; occasional heavy drinkers with high risk of problems; frequent heavy drinkers with high risk of problems

Among females, both frequent and occasional heavy drinkers had disproportionate risk of alcoholrelated social problems. In some groups occasional heavy drinkers had the same or higher risk of problems than frequent heavy drinkers. Among females, the most common alcoholrelated problems involved dating/sex.

Hispanic

4

Abstainers; experimenters; occasional heavy drinkers with high risk of problems; frequent heavy drinkers with high risk of problems

African American

4

 Whites more likely than other groups to be drinkers including heavy drinkers. White frequent heavy drinkers had very high levels of consumption compared with Hispanics/African Americans in this class, and comparable to male frequent heavy drinkers.  White frequent heavy drinkers had higher risk of social problems than Hispanic/African Americans frequent heavy drinkers.  Hispanic frequent heavy drinkers had lower risk of problems with parents than white/African Americans frequent heavy drinkers.  African American frequent heavy drinkers had lower risk of most social problems than white/Hispanic frequent heavy drinkers.

Males White

5

5 Hispanic

Am J Drug Alcohol Abuse Downloaded from informahealthcare.com by University of Laval on 06/30/14 For personal use only.

5 African American

Females White

drinkers with moderate risk of problems (18.3%), occasional heavy drinkers with high risk of problems (10.2%), frequent heavy drinkers with high risk of problems (7.3%)]. Among African American and Hispanic women, the optimal choice was a four-class solution [abstainers (African American: 63.4%, Hispanics: 55.1%), experimenters (African American: 20.7%, Hispanics: 22.2%), occasional heavy drinkers with high risk of problems (African American: 9.9%, Hispanics: 14.5%), frequent heavy drinkers with high risk of problems (African American: 6.1%, Hispanics: 8.2%)]. White female frequent heavy drinkers with high risk of problems had nearly universal endorsement of binge drinking, defined as four or more drinks on occasion (90%) and of getting drunk at least once per month (86%). These levels of alcohol use were comparable to those observed among white, African American, and Hispanic males and much higher than observed among Hispanic (binge: 79% monthly drunk: 66%) and African American (binge: 63%, monthly drunk: 80%) females in this class. Approximately 60% of white, Hispanic,

and African Americans in this group endorsed weekly drinking. African American and Hispanic female frequent heavy drinkers reported lower frequency and quantity of use than their white counterparts and were somewhat less likely to report physiological or social consequences compared with white female frequent heavy drinkers. Among female frequent heavy drinkers, high levels of regret about drinking were reported in each racial/ethnic group (white: 65%, Hispanic: 58%, African American: 56%) (Table 3b). While male frequent heavy drinkers reported a broad range of social outcomes, among female frequent heavy drinkers, the most commonly reported problems included problems with a dating situation (white: 36%, Hispanic: 43%, African American: 40%) and/or a sexual situation (white: 44%, Hispanic: 34%, African American: 27%). African American females were less likely to have problems with school, friends, sexual experiences and physiological symptoms compared to white and Hispanic female frequent heavy drinkers. Hispanic females were least likely to have problems

32 22 17 16 16 12 4

27 30 12 51 22 9 11

Frequent heavy – moderate risk of problems (%) 88 62 57 55 57 52 28

Frequent heavy – high risk of problems (%) 33 35 20 15 13 14 9

Occasional heavy (%) 14 14 17 23 13 0 0

Frequent heavy – moderate risk of problems (%)

Hispanic

75 60 75 37 65 53 23

Frequent heavy – high risk of problems (%) 39 22 25 13 13 11 10

Occasional heavy (%)

3 21 2 20 0 1 0

Frequent heavy – moderate risk of problems (%)

African American

74 46 43 47 48 28 24

Frequent heavy – high risk of problems (%)

Regret about drinking Problem with parents Problem with dating Regrettable sex Fighting Problem with friends Problem with school

24 16 14 11 8 3 10

Occasional heavy – moderate risk of problems (%) 90 42 51 53 23 11 35

Occasional heavy – high risk of problems (%)

White

65 40 36 44 29 19 30

Frequent heavy (%)

53 37 34 24 17 4 32

Occasional heavy (%)

Hispanic

58 27 43 34 27 15 28

Frequent heavy (%)

51 17 23 37 6 5 17

Occasional heavy (%)

56 36 40 27 27 5 20

Frequent heavy (%)

African American

Table 3b. Prevalence levels of social consequences of drinking among white, hispanic, and african american female adolescents who are occasional heavy drinkers versus frequent heavy drinkers in the US.

Regret about drinking Problem with parents Problem with dating Regrettable sex Fighting Problem with friends Problem with school

Occasional heavy (%)

White

Table 3a. Prevalence levels of social consequences of drinking among white, Hispanic, and African American male adolescents who are occasional heavy drinkers versus frequent heavy drinkers in the US.

Am J Drug Alcohol Abuse Downloaded from informahealthcare.com by University of Laval on 06/30/14 For personal use only.

DOI: 10.3109/00952990.2014.892950

Adolescent alcohol classes and adulthood drugs 219

220

M. R. Khan et al.

Am J Drug Alcohol Abuse, 2014; 40(3): 213–224

Table 4a. Adjusted odds ratios (AORs) and 95% Confidence intervals (CIs) of associations between adolescent alcohol use latent class and adolescent marijuana use among white (n ¼ 9548), Hispanic (n ¼ 3184), and African American (n ¼ 4005) adolescents in the US. White AOR Male adolescent alcohol use class Abstainer Experimenter Occasional heavy Frequent heavy – moderate risk of problems Frequent heavy – high risk of problems Female adolescent alcohol use class Abstainer Experimenter Occasional heavy – moderate risk of problems Occasional heavy – high risk of problems Frequent heavy – high risk of problems

Hispanic

a

African American

95% CI

AORa

95% CI

Reference 2.92 11.88 42.01 14.08

Reference 1.61, 5.30 6.82, 20.68 17.55, 100.69 6.69, 29.70

Reference 3.82 8.86 9.44 28.33

Reference 2.13, 6.86 5.39, 14.57 3.87, 23.04 12.99, 61.81

Reference 3.85 – 23.04 39.65

Reference 2.21, 6.69 – 12.42, 42.69 19.61, 80.16

Reference 3.51 – 13.87 26.82

Reference 2.15, 5.73 – 8.59, 22.40 14.82, 48.57

95% CI

AOR

Reference 3.05 11.24 15.66 37.45

Reference 2.28, 4.08 8.44, 14.97 10.64, 23.04 21.98, 63.82

Reference 2.55 12.58 23.08 47.66

Reference 1.81, 3.59 9.40, 16.83 14.43, 36.97 32.36, 70.18

a

Am J Drug Alcohol Abuse Downloaded from informahealthcare.com by University of Laval on 06/30/14 For personal use only.

a

Controlling for gender, poverty, delinquency, and maternal education.

Table 4b. Adjusted odds ratios (AORs) and 95% Confidence intervals (CIs) of associations between adolescent alcohol use latent class and adult marijuana and cocaine use among white (n ¼ 9548), Hispanic (n ¼ 3184), and African American (4005) adolescents in the US. White

Adulthood marijuana use Male adolescent alcohol use class Abstainer Experimenter Occasional heavy Frequent heavy – moderate risk of problems Frequent heavy – high risk of problems Female adolescent alcohol use class Abstainer Experimenter Occasional heavy – moderate risk of problems Occasional heavy – high risk of problems Frequent heavy – high risk of problems Adulthood cocaine use Male adolescent alcohol use class Abstainer Experimenter Occasional heavy Frequent heavy – moderate risk of problems Frequent heavy – high risk of problems Female adolescent alcohol use class Abstainer Experimenter Occasional heavy – moderate risk of problems Occasional heavy – high risk of problems Frequent heavy – high risk of problems

Hispanic

AORa

95% CI

Reference 1.06 2.84 0.29 3.32

Reference 0.30, 3.71 1.36, 5.92 0.06, 1.29 0.81, 13.64

Reference 1.04 1.89 1.80 2.67

Reference 0.61, 1.77 0.90, 3.96 0.54, 6.03 1.23, 5.79

Reference 1.30, 2.12 1.69, 3.34 1.04, 2.48 1.72, 3.55

Reference 1.52 – 1.40 2.28

Reference 0.79, 2.94 – 0.61, 3.23 0.87, 5.95

Reference 2.26 – 3.07 3.77

Reference 1.37, 3.71 – 1.58, 5.96 1.69, 8.40

Reference 1.07 1.83 1.45 3.09

Reference 0.69, 1.67 1.20, 2.80 0.81, 2.62 1.52, 6.30

Reference 0.07 3.95 9.94 5.34

Reference 0.00, 1.75 0.44, 35.38 1.16, 85.46 1.03, 27.61

Reference 1.92 2.02 1.49 2.62

Reference 0.39, 9.55 0.24, 17.27 0.13, 16.83 0.66, 10.49

Reference 1.24 2.39 3.10 3.12

Reference 0.68, 2.28 1.25, 4.59 0.71, 13.48 1.62, 6.01

Reference 3.25 – 1.31 2.25

Reference 0.73, 14.48 – 0.25, 6.83 0.40, 12.57

Reference 2.10 – 5.47 0.42

Reference 0.65, 6.78 – 1.04, 28.91 0.04, 5.04

95% CI

AOR

Reference 1.51 2.50 1.85 3.60

Reference 1.19, 1.91 1.86, 3.37 1.19, 2.86 2.11, 6.15

Reference 1.66 2.38 2.11 2.47

a

African American

95% CI

AOR

a

a

Controlling for gender, poverty, delinquency, maternal education, & adolescent substance use (marijuana, cocaine, inhalants, and injection drugs).

with parents compared to white and African American female frequent heavy drinkers. Among white, African American, and Hispanic adolescent girls, the LCAs identified a sub-population of occasional heavy drinkers with high risk of problems. In this group, levels of weekly drinking, monthly binge drinking, and getting drunk were much lower than the frequent heavy drinkers, with the majority – 62% of white, 74% of Hispanic, and 74% of African American females – drinking once per month or less often. However, when females in this group drank, they had high risk of drinking four or more drinks on occasion and levels of problems associated with alcohol use were comparable or higher than their counterparts who were frequent heavy

drinkers. For example, among white females, occasional heavy drinkers with high risk of problems had higher levels of regret and problems with sexual experiences, dating, parents, and friends (90%, 53%, 51%, 42%, 35%, respectively) than frequent heavy drinkers with high risk of problems (65%, 44%, 36%, 40%, 30%, respectively). Likewise, among African American females, occasional heavy drinkers with high risk of problems had comparable levels of regret and higher levels of problems with sexual experiences (51% and 37%, respectively) compared with frequent heavy drinkers with high risk of problems (56% and 27%, respectively). Among white adolescent girls only, the LCAs identified a second group of occasional heavy drinkers who had moderate

Adolescent alcohol classes and adulthood drugs

DOI: 10.3109/00952990.2014.892950

risk of problems. In this group, the frequency and quantity of use was comparable to other occasional drinkers yet levels of problems related to drinking were lower. Like their male counterparts, white, Hispanic, and African American female experimenters reported low risk of high frequency and quantity of drinking and were unlikely to report physiological or social consequences, yet the majority had friends who drank at least once per month. Approximately 80% or more of the experimenters in each racial/ethnic group drank once a month or less often, with very low levels of binge drinking, and around 5% experienced social or physiological consequences.

221

heavy use was strongly associated with cocaine use including among heavy users with low risk of alcohol-related problems (AOR: 9.94, 95% CI: 1.16, 85.46) and those with high risk of problems (AOR: 5.34, 95% CI: 1.03, 27.61). Among African American males, adolescents who were frequent heavy users with high risk of alcohol-related problems were significantly more likely to use marijuana in adulthood (AOR: 2.67, 95% CI: 1.23, 5.79). Lower severity alcohol use classes were not linked to adulthood marijuana use in this group, while alcohol use was not significantly associated with adulthood cocaine use in this group. Females

Am J Drug Alcohol Abuse Downloaded from informahealthcare.com by University of Laval on 06/30/14 For personal use only.

Associations between adolescent alcohol use latent class and adolescent marijuana use Males Among males, adjusted analyses indicated that being in any of the drinking groups was associated with marked elevations in adolescent marijuana use (Table 4a). A clear dose response relationship was observed between severity of alcohol use class and marijuana use, with being a frequent heavy drinker associated with well over ten times the odds of adolescent marijuana use in each racial/ethnic group (white adjusted odds ratio (AOR): 37.45, 95% CI: 21.98, 63.82; Hispanic AOR: 14.08, 95% CI: 6.69, 29.70; African American AOR: 28.33, 95% CI: 12.99, 61.81). Females A dose response relationship also was observed between severity of alcohol use class and marijuana use in each racial/ ethnic group of females (Table 4a), with being a frequent heavy drinker associated with between 27 to 48 times the odds of with marijuana use in each racial/ethnic group (white AOR: 47.66, 95% CI: 32.36, 70.18; Hispanic AOR: 39.65, 95% CI: 19.61, 80.16; African American AOR: 26.81, 95% CI: 14.82, 48.57). Associations between adolescent alcohol use latent class and young adult drug use Males While longitudinal analyses suggested alcohol use in adolescence was associated with adulthood drug use, differences by race/ethnicity were observed between adolescent alcohol use and adulthood drug use among males (Table 4b). Among white males, analyses controlling for gender, poverty, delinquency, maternal education, and adolescent substance use (marijuana, cocaine, inhalants, and injection drugs) suggested any adolescent alcohol use was associated with elevated odds of adulthood marijuana use, and white males who had been occasional or frequent heavy users experienced elevated odds of adulthood cocaine use. Frequent heavy users with high risk of problems had over three times the odds of adulthood marijuana use (AOR: 3.60, 95% CI: 2.11, 6.15) and cocaine use (AOR: 3.12, 95% CI: 1.62, 6.01). Among Hispanic males, alcohol experimentation was associated with over 2.5 times the odds of adulthood marijuana use (AOR: 2.84, 95% CI: 1.36, 5.92), yet no other class of alcohol use was linked to marijuana use. In this group, frequent

Adolescent alcohol use appeared to be strongly and consistently associated with adulthood drug use among white and African American females. Among white women, each adolescent alcohol use class was associated with increased odds of adulthood marijuana use, with frequent heavy users at greatest odds of adulthood marijuana use (AOR: 2.47, 95% CI: 1.72, 3.55). Among white females, alcohol experimentation was not associated with adulthood cocaine use, though females who had been occasional users or heavy users with alcohol-related problems in adolescence had elevated odds of adulthood cocaine use, with the strongest associations observed among women who had been frequent heavy users with high risk of alcohol-related problems (AOR: 3.12, 95% CI: 1.62, 6.01). Among African American females, like their white female counterparts, each adolescent alcohol use class was associated with adulthood marijuana use, with frequent heavy users at greatest odds of adulthood marijuana use (AOR: 3.77, 95% CI: 1.69, 8.40). African American females who had been occasional users with high risk of problems in adolescents were over 5 times more likely than adolescent abstainers to use cocaine in adulthood (AOR: 5.47, 95% CI: 1.04, 28.91). Though similar trends were observed among Hispanic females, the smaller sample of Hispanics compared with whites and African Americans yielded low power and no significant associations were detected in this group.

Discussion This is one of the only studies, to our knowledge, to describe classes of alcohol users among both male and female white, African American, and Hispanic adolescents, to assess gender as well as race differences in typologies of adolescent alcohol use, and to assess associations between adolescent alcohol use and adolescent and adulthood illicit drug use in the US general population using nationally-representative data. The results corroborated the well-established trend of higher levels of adolescent alcohol use among white versus African American adolescents and suggested that Hispanics have levels of alcohol use comparable to white adolescents, which is inconsistent with findings of prior studies suggesting higher levels of use among whites than Hispanics (41–43). Analyses suggested males and females were characterized by different alcohol use typologies and consequences due to alcohol use, with two striking findings observed. First, while disproportionate vulnerability to alcohol-related problems was observed among males in the highest severity classes of those who drank both heavily and frequently, among females,

Am J Drug Alcohol Abuse Downloaded from informahealthcare.com by University of Laval on 06/30/14 For personal use only.

222

M. R. Khan et al.

even those who drank heavily on occasion experienced very high risk of alcohol-related social problems. In fact in some cases, occasional heavy drinkers had higher risk of problems than frequent heavy drinkers among adolescent girls. A second key gender difference was that while males as a population experienced substantial heterogeneity in the type of social problems experienced, from family to friends to dating, among females, problems with dating and/or sex emerged as key consequences of alcohol use. These findings point to a need for gender-specific alcohol risk reduction programming to most effectively target resources. Specifically, while among males programs can prioritize resources to first reach frequent heavy drinkers, all females with a history of heavy drinking (four or more drinks on an occasion), even if infrequent, should be reached. Second, while programs should provide strategies for how to prevent and/or address a broad range of alcohol-related problems among males, they should focus more time on developing skills to avoid regrettable relationship and sexual situations that result from alcohol use. While race differences in the prevalence of different social problems associated with use were observed, findings do not suggest that race-specific alcohol prevention or treatment is warranted. Rather, results suggest a brief assessment given to all adolescents in programs could accurately classify adolescents with different prevention/intervention needs. Among all adolescents, being a drinker in any alcohol use class was a strong correlate of adolescent marijuana use, with marijuana use risk increasing dramatically with increasing severity of alcohol use class. For example, experimenters had between 2.5 to 4 times the odds of marijuana use depending on the gender and racial/ethnic group while frequent heavy alcohol users had between 5 to 29 times the odds of marijuana use among male adolescents and between 27 to 48 times the odds of use among female adolescents. The findings suggest alcohol risk reduction programs should be coupled with messages that aim to prevent and address adolescent marijuana use, with more intensive screening appropriate in heavier drinkers. In longitudinal studies that controlled for socio-demographic factors and baseline drug use in adolescence, alcohol experimentation and, to a greater extent, moderate and problem use was independently associated with use of illicit drugs in adulthood. The results of these longitudinal analyses suggest that working with adolescent users may help prevent a high-risk trajectory that leads to substance use in adulthood. These results support previous findings indicating that adolescent substance use is linked with adulthood and lifetime drug use (20–22). Research suggests that experimentation with substances has become a normative rite of passage among adolescents in the US (44). Residual confounding by unmeasured personality or environmental factors may present bias when estimating associations between adolescent alcohol use and adulthood illicit drug use (21,22,45–47). Nonetheless, these findings highlight the tremendous importance of targeting adolescent alcohol users in programs that aim to reduce commonly used substances such as marijuana as well as ‘‘harder’’ drugs such as cocaine. Results of this study could be used to more effectively target scarce prevention/intervention resources. LCAs provide

Am J Drug Alcohol Abuse, 2014; 40(3): 213–224

a detailed summary of different profiles of adolescent alcohol users, rather than treating any adolescent who endorses prior alcohol use as a homogenous group. Hence, interventions guided by the results of the LCA can be used to adapt alcohol reduction to needs of unique groups of users. There is a large group of abstainers who do not need to be reached with intensive alcohol or drug risk reduction messaging. While all adolescents who drink should be reminded of the health and social dangers associated with drinking, especially with drinking heavily even if only occasionally, the most intensive prevention/intervention resources could be used with adolescents who are in the most problematic latent classes. Because these results highlight marked gender differences in the number of profiles of users, the results suggest alcohol treatment and prevention should take these sub-group differences in account. Frequent heavy drinkers among males and both frequent and occasional heavy drinkers among females appeared particularly vulnerable to the physiological and social effects of alcohol use and should be prioritized for alcohol risk reduction interventions. Results suggest a brief assessment could accurately classify adolescents with different prevention/intervention needs. Tailored interventions are believed to be more cost-effective and participants exhibit greater compliance (48). The greatest reduction in adverse outcomes will most likely only occur if the right program is targeting the right factors in the right individuals (49). In addition to the important potential for residual confounding, further limitations must be noted. First, uncertainty in true class membership is a limitation. In the LCA, the latent variable does not perfectly predict and perfectly classify individuals into the classes. Hence the potential misclassification should be considered when interpreting results. That said, given the high values for entropy (4¼0.93) we assume such error will be minimal. A second limitation is that we captured alcohol latent classes at one point in time during adolescence, while transitions between alcohol classes occur rapidly, even within even a one-year time period (26). Furthermore, LCA analyses were based on self-reported measures of adolescent alcohol use, which are known to be affected by social desirability bias. Another limitation was that analyses examining adolescent latent class membership and adolescent substance use were based on cross-sectional data; it is not possible to definitively assess whether alcohol use contributed to elevations in use of illicit drugs in adolescence or if other factors drove elevated risk of both alcohol and drug use and hence were the factors underlying an observed association. However, prior studies suggest alcohol is a gateway drug to use of illicit substances in adolescence and hence provide grounds for the proposed hypothesized relationships. Further, in longitudinal analyses examining adolescent latent class and adulthood substance use, we were able to control for adolescent alcohol and drug use. These analyses were among the first that used longitudinal data to evaluate the potential influence of drinking on drug use. Despite these limitations, this study was the first to document gender and race differences in typologies of adolescent alcohol use and the strong independent associations between adolescent alcohol use and both adolescent and young adulthood substance use. The findings indicated gender differences, more so than race differences, in

DOI: 10.3109/00952990.2014.892950

typologies of alcohol use should be considered when designing population-specific programs to address adolescent alcohol use, and that working with adolescent alcohol users may reduce the risk of following a trajectory that leads to substance use in adulthood.

Am J Drug Alcohol Abuse Downloaded from informahealthcare.com by University of Laval on 06/30/14 For personal use only.

Acknowledgements This research was supported by the National Institute on Drug Abuse grant Longitudinal Study of Substance Use, Incarceration, and STI in the US (Maria Khan, PI, R03 DA026735). This research used data from Add Health, a program project designed by J. Richard Udry, Peter S. Bearman and Kathleen Mullan Harris, and funded by a grant P01- HD31921 from the National Institute of Child Health and Human Development, with cooperative funding from 17 other agencies. Special acknowledgement is due to Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. People interested in obtaining data files from Add Health should contact Add Health, Carolina Population Center, 123 W. Franklin Street, Chapel Hill, NC 27516– 2524 (http://www.cpc.unc.edu/addhealth/contract.html).

Declaration of interest The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this paper. This research was supported by the National Institute on Drug Abuse (NIDA) grant Longitudinal Study of Substance Use, Incarceration, and STI in the US. NIDA had no further role in study design; in the collection, analysis and interpretation of data; in the writing of the report; or in the decision to submit the paper for publication.

References 1. U.S. Department of Health and Human Services, Office of the Surgeon General. The Surgeon General’s call to action to prevent and reduce underage drinking; 2007. 2. Centers of Disease Control (CDC). Youth risk behavior surveillance United States, 2009. MMWR 2010;59(SS-5):1–142. Available from: www.cdc.gov/mmwr/pdf/ss/ss5905.pdf [Accessed December 2013]. 3. Field MST, Wiers RW. Cognitive processes in alcohol binges: a review. Curr Drug Abuse Rev 2008;1:263–179. 4. Corte CM, Sommers MS. Alcohol and risky behaviors. Annu Rev Nurs Res 2005;23:327–360. 5. Hartley DE, Elsabagh S, File SE. Binge drinking and sex: effects on mood and cognitive function in healthy young volunteers. Pharmacol Biochem Behav (US) 2004;611–619. 6. Squeglia LM, Jacobus J, Tapert SF. The influence of substance use on adolescent brain development. Clin EEG Neurosci 2009;40: 31–38. 7. Townshend JM, Duka T. Binge drinking, cognitive performance and mood in a population of young social drinkers. Alcohol Clin Exp Res (US) 2005;317–325. 8. Brown SA, Tapert SF, Granholm E, Delis DC. Neurocognitive functioning of adolescents: effects of protracted alcohol use. Alcohol Clin Exp Res 2000;24:164–171. 9. de Wit H. Impulsivity as a determinant and consequence of drug use: a review of underlying processes. Addict Biol (UK) 2009;14:22–31. 10. Norris J, Stoner SA, Hessler DM. et al. Cognitive mediation of alcohol’s effects on women’s in-the-moment sexual decision making. Health Psychol (US) 2009;20–28. 11. Fromme K, D’Amico EJ, Katz EC. Intoxicated sexual risk taking: an expectancy or cognitive impairment explanation? J Stud Alcohol 1999;60:54–63.

Adolescent alcohol classes and adulthood drugs

223

12. Miller JW, Naimi TS, Brewer RD, Jones SE. Binge drinking and associated health risk behaviors among high school students. Pediatrics (US) 2007;119:76–85. 13. Maney DW, Higham-Gardill DA, Mahoney BS. The AlcoholRelated Psychosocial and Behavioral Risks of a Nationally Representative Sample of Adolescents. J Sch Health 2002;72: 157–163. 14. Swahn MH, Simon TR, Hammig BJ, Guerrero JL. Alcoholconsumption behaviors and risk for physical fighting and injuries among adolescent drinkers. Addict Behav 2004;29:959–963. 15. Baskin-Sommers A, Sommers I. The co-occurrence of substance use and high-risk behaviors. J Adolesc Health 2006;38:609–611. 16. Miller JW, Naimi TS, Brewer RD, Jones SE. Binge drinking and associated health risk behaviors among high school students. Pediatrics 2007;119:76–85. 17. The MMWR Surveillance Summary Youth Risk Behavior Surveillance-United States, 2011;64:4. June 8, 2012. 18. Chassin L, Flora DB, King KM. Trajectories of alcohol and drug use and dependence from adolescence to adulthood: the effects of familial alcoholism and personality. J Abnormal Psychol 2004;113: 483–498. 19. Dierker LC, Vesel F, Sledjeski EM, Costello D, Perrine N. Testing the dual pathway hypothesis to substance use in adolescence and young adulthood. Drug Alcohol Depend 2007;87:83–93. 20. Grant BF, Dawson DA. Age of onset of drug use and its association with DSM-IV drug abuse and dependence: results from the National Longitudinal Alcohol Epidemiologic Survey. J Subst Abuse (US) 1998;10:163–173. 21. Grant JD, Scherrer JF, Lynskey MT, Lyons MJ, Eisen SA, Tsuang MT, True WR, et al. Adolescent alcohol use is a risk factor for adult alcohol and drug dependence: evidence from a twin design. Psychol Med (UK) 2006;36:109–118. 22. Palmer RH, Young SE, Hopfer CJ, Stallings MC, Crowley TJ, Hewitt JK. Developmental epidemiology of drug use and abuse in adolescence and young adulthood: evidence of generalized risk. Drug Alcohol Depend 2009;102:78–87. 23. Townshend JM, Duka T. Patterns of alcohol drinking in a population of young social drinkers: a comparison of questionnaire and diary measures. Alcohol Alcohol 2002;37:187–192. 24. Goodman LA. Exploratory latent structure analysis using both identifiable and unidentifiable models. Biometrika 1974;61: 215–231. 25. Hesselbrock VM, Hesselbrock MN. Are there empirically supported and clinically useful subtypes of alcohol dependence? Addiction 2006;101:97–103. 26. Bucholz KK, Heath AC, Madden PA. Transitions in drinking in adolescent females: evidence from the Missouri adolescent female twin study. Alcohol Clin Exp Res 2000;24:914–923. 27. Chung T, Martin CS, Clark DB. Concurrent change in alcohol and drug problems among treated adolescents over three years. J Stud Alcohol Drugs 2008;69:420–429. 28. Reboussin BA, Song EY, Shrestha A, Lohman KK, Wolfson M. A latent class analysis of underage problem drinking: evidence from a community sample of 16–20 year olds. Drug Alcohol Depend 2006; 83:199–209. 29. Fergusson DM, Horwood LJ, Lynskey MT. The prevalence and risk factors associated with abusive or hazardous alcohol consumption in 16-year-olds. Addiction 1995;90:935–946. 30. Dauber S, Hogue A, Paulson JF, Leiferman JA. Typologies of alcohol use in white and African American adolescent girls. Subst Use Misuse 2009;44:1121–1141. 31. Flory K, Brown TL, Lynam DR, Miller JD, Leukefeld C, Clayton RR. Developmental patterns of African American and Caucasian adolescents’ alcohol use. Cultur Divers Ethnic Minor Psychol 2006; 12:740–74–76. 32. Bureau, U.S.C. Hispanic Heritage Month 2013. 33. The National Longitudinal Study of Adolescent Health, Carolina Population Center, University of North Carolina at Chapel Hill website. Available from: www.cpc.unc.edu/addhealth [Accessed December 2013]. 34. Bearman PS, Jones J, Udry JR. 1997. The National Longitudinal Study of Adolescent Health, Carolina Population Center, The University of North Carolina at Chapel Hill website. Available from: www.cpc.unc.edu/projects/addhealth [Accessed 16 March 2007].

Am J Drug Alcohol Abuse Downloaded from informahealthcare.com by University of Laval on 06/30/14 For personal use only.

224

M. R. Khan et al.

35. Udry JR. 1991. References, instruments, and questionnaires consulted in the development of the Add Health in-home adolescent interview, Carolina Population Center, The University of North Carolina at Chapel Hill website. Available from: www.cpc.unc.edu/ projects/addhealth/files/refer.pdf. [Accessed 16 March 2007]. 36. Resnick MD, Bearman PS, Blum RW, Bauman KE, Harris KM, Jones J, Tabor J, et al. Protecting adolescents from harm. Findings from the National Longitudinal Study on Adolescent Health. Jama 1997;278:823–832. 37. Chantala K, Tabor J. 1999. Strategies to perform a design-based analysis using the Add Health data, Carolina Population Center, The University of North Carolina at Chapel Hill website. Available from: www.cpc.unc.edu/projects/addhealth/files/weight1.pdf. [Accessed 16 March 2007]. 38. Sieving RE, Beuhring T, Resnick MD, Bearinger LH, Shew M, Ireland M, Blum RW. Development of adolescent self-report measures from the National Longitudinal Study of Adolescent Health. J Adolesc Health 2001;28:73–81. 39. Cohen M, Feng S, Florey FA, Ford CA, Harris KM, Hewitt JK, Hobbs MM, et al. 2003. Biomarkers in Wave III of the Add Health Study, Carolina Population Center, The University of North Carolina at Chapel Hill website. Available from: www.cpc.unc.edu/projects/addhealth/files/biomark.pdf. [Accessed 16 March 2007]. 40. Muthe´n B, Muthe´n BO. Mplus User’s Guide. 5th ed. Los Angeles, CA: Muthe´n & Muthe´n; 1998–2007. 41. Substance Abuse and Mental Health Services Administration (SAMHSA). Results from the 2009 National Survey on Drug Use

Am J Drug Alcohol Abuse, 2014; 40(3): 213–224

42.

43. 44. 45.

46.

47.

48. 49.

and Health: national findings. Rockville, MD: Office of Applied Studies, SAMHSA; 2010. Blum RW, Beuhring T, Shew ML, Bearinger LH, Sieving SE, Resnick MD. The effects of race/ethnicity, income, and family structure on adolescent risk behaviors. Am J Public Health 2000;90: 1879–1884. Griffin KW, Scheier LM, Botvin GJ, Diaz T. Ethnic and gender differences in psychosocial risk, protection, and adolescent alcohol use. Prev Sci 2000;1:199–212. Vega WA, Gil AG. Drug use and ethnicity in early adolescence. New York: Plenum; 1998. Kendler KS, Myers J, Prescott CA. Specificity of genetic and environmental risk factors for symptoms of cannabis, cocaine, alcohol, caffeine, and nicotine dependence. Arch Gen Psychiatry (US) 2007;64:1313–1320. Rhee SH, Hewitt JK, Young SE, Corley RP, Crowley TJ, Neale NC, Stallings MC. Comorbidity between alcohol dependence and illicit drug dependence in adolescents with antisocial behavior and matched controls Drug Alcohol Depend (Ireland) 200684:;85–92. Young SE, Rhee SH, Stallings MC, Corley RP, Hewitt JK. Genetic and environmental vulnerabilities underlying adolescent substance use and problem use: general or specific? Behav Genet 2006;36: 603–615. Collins LM, Murphy SA, Bierman KL. A conceptual framework for adaptive preventive interventions. Prev Sci 2004;5:185–196. Lanza ST, Rhoades BL. Latent class analysis: an alternative perspective on subgroup analysis in prevention and treatment. Prev Sci 2013;14:157–168.