Download - Pathways From ADHD to Early Drug Use

Pathways From ADHD to Early Drug Use

HOWARD D. CHILCOAT, Sc.D., AND NAOMI BRESLAU, PH.D.

ABSTRACT

Objective: This study tests whether attention-deficit hyperactivity disorder (ADHD) increases the risk of early drug use.

Method: A Community-based sample of 412 low birth weight and 305 normal birth weight children and their mothers initially

were assessed when the children were 6 years old with a follow-up assessment at age 11. Results: The relationship of

ADHD with drug use varied by level of externalizing problems. Regardless of ADHD status, children with a low level of exter-

nalizing problems had a low risk of drug use, and those with the highest level of externalizing problems had a high risk. At

the middle level of externalizing problems, ADHD increased the incidence of drug use to the magnitude observed at the

high level of externalizing problems, and children with ADHD were at significantly higher risk than those without ADHD

(odds ratio = 2.1, p = .03). Findings were similar for low and normal birth weight children. Low parent monitoring and high

peer drug use signaled increased risk of drug use for children, independent of ADHD status. Psychostimulant treatment for

ADHD was unrelated to risk of drug use. Conclusions: Risk for early drug use in children with ADHD depends on level of

associated externalizing problems. Parent monitoring and peer drug use appear to be potential targets for drug prevention

for children with ADHD, as well as children in general. J. Am. Acad. Child Adolesc. Psychiafry; 1999, 38(11):1347-1354. Key

Words: attention-deficit hyperactivity disorder, drug use, smoking, alcohol.

In this report, bringing together 2 distinct areas of research, we investigate whether children with attention- deficit hyperactivity disorder (ADHD) are at increased risk of early drug initiation. The first considers the role of ADHD as an individual-level factor that predisposes children to be involved with psychoactive drugs. The second measures the extent to which specific environmental factors influence the risk of drug involvement. We use data from a prospective study to estimate the risk of early

et al., 1993), yet little is known about the impact of ADHD on early drug use. The past 2 decades of research have clarified the importance of early-onset drug use (i.e., before the mid-teen years). Drug use that starts in childhood is atypical and more likely to lead to later drug problems than drug use that starts later in adolescence or early adulthood (Anthony and Petronis, 1995; Breslau et al., 1993; Escobedo et al., 1993; Fleming et al., 1982; Robins and Przybeck, 1985; Yamaguchi and Kandel, 1984).

drug use associated with ADHD and test whether factors in the social environment, specifically, parental monitoring and drug use by peers, affect the risk of drug use in children with ADHD in the same way that they have been found to affect children in general.

Early Drug Use

ADHD and Substance Abuse

A number of studies, mostly of clinical samples, link ADHD with substance use disorders. Retrospective studies have found elevated rates of ADHD among substance abusers (Carroll and Rounsaville, 1993; DeMilio, 1989; Eyre et al., 1982; Goodwin et al., 1975; Horton et al., 1987; Tarter et al., 1977; Wood et al., 1983). Similarly, studies comparing the prevalence of substance use disorders in clinical samples of patients with ADHD and controls have found associations between ADHD and adult drug use disorders (Biederman et al., 1995; Mannuzza et al., 1993).

orders in adolescence are inconsistent (Biederman et al.,

Most studies of ADHD and drug use have focused on drug use disorders (Biederman et al., 1997; Mannuzza

Accepted May 26, 1999. From the Department .f I'ychiany, Hen y Ford Health Sciences Center, Dehoit. This researcb was supported by NIDA grant DA11952 and NZMHgrant

Reprint quests to Dr. Cbilcoat, Psychiatry Research, Henry Ford Health Sciences Center, I Ford Place, 3A, Detroit, MI 48202; e-mail: [email protected].

Results on the association of ADHD with drug use dis- MH44586. The authors thank Dr. Rachel Klein for helpfir1 comments.

1997; M~~~~~~ et al., 1991). Studies that have focused on

0890-8567/99/3811-134701~99 by the American Academy of Child and use, rather than have Produced conflicting as Adolescent Psychiatry. well (Boyle et al., 1993; Milberger et al., 1997a,b). How-

J . A M . A C A D . CHILD A D O L E S C . PSYCHIATRY, 38:11, N O V E M B E R 1999 1347

C H I L C O A T A N D BRESLAU

ever, none of these studies examined the earliest stage of children’s drug use, that is,jrst use of drugs, usually involv- ing alcohol or tobacco.

Externalizing Problems and Drug Use

Another area of controversy is the role of externalizing behavior problems, including aggressiveness, acting-out and disruptive behaviors, as well as symptoms of conduct and oppositional defiant disorders, as mediators in the pathway from ADHD to drug use. Externalizing problems in childhood, regardless of disruptive behavior diagnosis, have been identified consistently as one of the strongest predictors of drug use and abuse (Hawkins et al., 1992). In addition, clinical (e.g., Barkley et al., 1990; Biederman et al., 1991; Steingard et al., 1992) and epidemiological studies (e.g., Jensen et al., 1993; Lynskey and Fergusson, 1995; Offord et al., 1992) indicate that ADHD is associated with dimensional measures of externalizing, as well as conduct disorder. Klein and colleagues (Gittelman et al., 1985; Mannuzza et al., 1991) reported no direct link between childhood hyperactivity per se and later substance use disorders. Instead, the onser of conduct disorder or antisocial personality disorder in children with ADHD regularly preceded drug use problems. Similarly, Lynskey and Fergusson (1995) found no association between ADHD and drug use once externalizing problems were taken into account. In contrast, Milberger et al. (1997a) found that ADHD was a significant risk factor for smoking, independent of conduct disorder. Studies of boys in treatment for conduct disorder indicate that ADHD is associated with an increased number of substance dependence symptoms (Thompson et al., 1996; Whitmore et al., 1998). The discrepancies across these studies point to the need for further longitudinal research into the specific roles of ADHD and externalizing problems in the pathway to drug use.

ADHD and Protective/Risk Factors

Several studies have identified drug use by peers as a risk factor for drug use (Hawkins et al., 1992), and numerous intervention strategies have been developed to reduce the influence of peers on children’s drug use (Botvin et al., 1995; Ellickson and Bell, 1998). Recently, effective monitoring and supervision by parents have emerged as specific aspects of child-rearing that might prevent or delay the onset of children’s drug-taking (Barnes and Farrell, 1992; Brown et al., 1993; Chilcoat and Anthony, 1996; Chilcoat et al., 1995; Dishion et al., 1988; Richardson et al., 1989).

Despite the consistent evidence that these risk and protective factors influence drug-taking behavior for children in generaL, little is known about their effects on children with specific vulnerabilities to drug use, such as ADHD.

Research Questions

There are important gaps in previous research that limit the understanding of the relationship between ADHD and drug use. First, most studies have focused on drug use disorders, rather than earlier stages. Thus, it is difficult to segregate the role of ADHD and externalizing problems at each stage in the progression of drug use. Second, most of the evidence of the interrelationship of ADHD, conduct disorder, and drug disorder comes from studies of clinical samples, which are highly susceptible to selection bias, especially with respect to comorbidity (Berkson, 1946).

To fill these gaps, we address the following questions, using data collected prospectively from a community- based sample of children: (1) Does ADHD signal increased risk of early drug use? (2) What is the role of externalizing problems in the pathway from ADHD to drug use in childhood? (3) What is the impact of social environmental factors, including peer drug use and parent monitoring, on early drug use among children with and without ADHD?

METHOD

Sample Data come from a sample of 717 children who were assessed at ages

6 and 11 years in a longitudinal study of the neuropsychiatric consequences of low birth weight. The sample and study have been described in detail elsewhere (Breslau et al., 1996a,b; Chilcoat et al., 1996). In brief, 628 low birth weight (LBW, <2,500 g) and 477 normal birth weight (NBW) children were selected from the 1983 to 1985 lists of newborn discharges from 2 major hospitals in southeast Michigan, one in the city of Detroit and the other in a nearby suburb. Children with severe neurological impairment were excluded. A total of 823 children (473 LBW and 350 NBW; 75% of those eligible) completed the initial assessment at age 6 years. Of this cohort of children, 717 (87%) were reassessed when they were 11 years old. Overall, the characteristics of the follow-up sample were similar to those of the initial sample (Table 1).

Assessments Neuropsychological performance, child psychiatric disorders, and

behavior problems were measured at age 6 and age 11. Children’s self- reports of drug use, parent monitoring, and peer drug use were obtained at age 11 only. Mothers’ history of psychiatric disorder was measured with the Diagnostic Interview Schedule when the children were 6 years old.

When the children were 6 years old, the Diagnostic Interview Sched- ule for Children (DISC) (Shaf5er et al., 1988) used mothers’ reports to measure ADHD, oppositional defiant disorder, and anxiety disorders in

1348 I . AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 38:11, N O V E M B E R 1999

ADHD AND EARLY DRUG USE

TABLE 1 Characteristics of the Sample at Baseline and Follow-up Assessments

Characteristic ( n = 823) ( n = 717)

African-American (Yo) 42.9 46.2 Mother‘s education (%)

<High school 16.9 17.1 High school 27.5 26.5 Some college 37.3 38.0 College 18.3 18.4

Maternal age, mean (SD) 33.1 (5.6) 33.0 (5.6) Single mother (Yo) 23.8 25.7 Urban site of birth (Yo) 50.3 52.9

Age 6 Age 11

Low birth weight (<2,500 g) (Yo) 57.4 57.4

children according to DSM-ZZZ-R criteria. At the age 11 assessment, the DISC was used to interview both children and mothers.

At both assessments, mothers and teachers rated children’s behavior problems, using the Child Behavior Checklist (CBCL) and the Teacher‘s Report Form, respectively (Achenbach, 1991). In this report we use the CBCL Externalizing scale as a continuous measure of children’s conduct problems. This composite scale contains 2 subscales, Delin- quent Behavior and Aggressive Behavior. Raw scores on the CBCL were converted to Tscores, based on age/sex-defined percentiles from normative samples (Achenbach, 1991).

Children’s use of drugs, peer drug use, and level of parent monitoring were measured at the follow-up assessment, using child self-reports. We used selected items from the Johns Hopkins Prevention Center Youth Interview (Chilcoat and Anthony, 1996; Chilcoat et al., 1995) to inquire about children’s use of tobacco, alcohol, marijuana, and inhalants. In this report, children were counted as drug users if they had ever used tobacco, alcohol (without parents’ permission), marijuana, or inhalants at least once. Items in the parent monitoring scale concerned patents’ supervisory rules and surveillance of the child’s whereabouts and behaviors outside the home. Peer drug use items measured friends’ use of alcohol, tobacco, and marijuana; friends’ atti- tudes toward drug use; and exposure to other people’s use of marijuana, cocaine, or other drugs.

Data Analysis Contingency tables tested crude associations of drug use with

ADHD and other factors. Logistic regression estimated relative odds (RO) of drug use for ADHD and other factors, adjusting for possible confounders. We also used generalized additive models (GAM) (Hastie andTibshirani, 1990) to estimate the incidence of drug use in relation to continuous predictors. This nonpararnetric approach minimizes the possible bias that can occur when a parametric model using a continuous predictor is misspecified (for example, when a linear term is used to model a curvilinear relationship). These models maximize information and precision that is otherwise lost when continuous variables are treated as categories by using cutoffs and have been used by our research group to shed light on associations that were obscured by using linear models (Chilcoat and Schiitz, 1996).

RESULTS

Incidence of Drug Use by Age 11

Of the 711 children with complete data, 137 (19.1%) had used drugs at least once. Incidence was the highest for

tobacco and alcohol, 10.6% and 10.1%, respectively. Of the 117 children who had used tobacco or alcohol, 31 had used both. A small number of children had used inhalants by age 11 (3.8%), and only 7 had used marijuana. The majority of children who reported inhalant use (20 of 27) had done so without ever smoking or drinking.

ADHD and Drug Use

As shown inTable 2, ADHD at age 6 signaled increased risk of drug use by age 11 (x2 = 11.0, df = 1, p < .001). When we adjusted for sex, birth weight, and site (urban versus suburban) in a logistic regression model, the odds of drug use for children with ADHD relative to those without ADHD was 1.7 (95% confidence interval [CI] =

1.1-2.7). It is interesting that even with ADHD status held constant, low birth weight emerged as a vulnerability marker for early drug use (RO = 1.6, 95% CI = 1.1-2.4). Nonetheless, the magnitude of the association between ADHD and drug use was identical for the LBW and NBW groups (RO = 1.7 for each), although reduced sample sizes resulted in less precise estimates, especially for the NBW group (Z= 2.09, df = 1 ,p = .04, LBW; Z=

To test whether ADHD might be manifested differently in LBW versus NBW children, we compared the profile of ADHD symptoms for LBW and NBW children with ADHD. As shown in Figure 1, the profile of symptoms was nearly identical for both birth weight groups.

We used GAM curves to estimate the extent to which number of symptoms of ADHD, as an indicator of ADHD severity, signaled increased risk of early drug use (Fig. 2). The risk of drug use remained constant and at

1.32, df = 1 ,p = .19, NBW).

TABLE 2 Incidence and RO of Drug Use by ADHD and Birth Weight Status

Incidence Adjusted RO n (%I (95% CI)

All children ADHD 146 28.8 1.7 (1.1-2.7)“ No ADHD 565 16.6 1 .o ADHD 100 32.0 1.7 ( 1 .0-3.0)b No ADHD 307 19.2 1 .o

ADHD 46 21.7 1.7 (0.8-3.4)b No ADHD 258 13.6 1 .o

Low birth weight

Normal birth weight

Note: RO = relative odds; ADHD = attention-deficit hyperactivity

a Adjusted for sex, site of birth, and birth weight. disorder; CI = confidence interval.

Adjusted for sex and site of birth.

J . AM. ACAD. C H I L D ADOLESC. PSYCHIATRY, 38:11, NOVEMBER 1999 1349

C H I L C O A T A N D BRESLAU

P 1.1 rn I 1111 111 I n II in11 mp 111 111 I I I - . . . . . .... . .

20 i

-, ............. A ....

......... .........

- LBW ----NBW

8 C a, -0 0 C

.- -

t 0

c - Fig. 1 Prevalence of attentlon-deficit hyperactivity disorder (ADHD) symptoms among low birth weight (LBW) and normal birth weight (NBW) children who met criteria for A D H D at age 6 years. Fidgets = often fidgets or squirms; Seated = difficulty remaining seated; Distracted = easily distracted; Waiting = difficulty awaiting turn; Blurts = often blurts out answers; Instructions = difficulty following through on instructions; Attention =

difficulty sustaining attention; Activities = often shifts from one uncompleted activity to another; Playing = difficulty playing quietly; Talks = often talks excessively; Interrupts = often interrupts or intrudes; Listen = often does not seem to listen; Loses = often loses things necessary for tasks; Danger = often engages in dangerous activities without considering consequences.

relatively low levels across the range of symptom counts below the diagnostic threshold for ADHD (i.e., 0-7 symptoms). There was an increase in the risk of drug use in the 8+ symptom range, which peaked at 10 symptoms with estimated incidence of 35%. The small number of observations at the maximum number of symptoms renders estimates at this level imprecise, as indicated by the broad CIS.

Role of Behavior Problems

There was a strong association between ADHD and children's behavior problems, as measured by the total Externalizing scale on the CBCL. The mean Tscore for Externalizing among children with ADHD was more than 1 SD higher than that for children without ADHD (mean Externalizing Tscore = 58.7 and 47.4, SD = 9.9 and 9.1 for children with and without ADHD, respectively; t = 13.1, df= 710,p < .001).

To compare the incidence of drug use by ADHD at specific levels of Externalizing problems, we plotted sep- arate GAM curves for children with and without ADHD

0 2 4 6 8 1 0 1 2 1 4 Total ADHD Symptoms

Fig. 2 Incidence of drug use by age 11 years by number of artention-deficit hyperactivity disorder (ADHD) symptoms at age 6 years, as estimated by generalized additive models. The solid line represents the estimated incidence of drug use, and the dotted lines represent the corresponding 95% confidence intervals. Hatch marks at the top and bottom of the figure cor- respond to the number of A D H D symptoms for each child with and without a history of drug use, respectively.

(Fig. 3). These curves indicate an interaction between ADHD and Externalizing problems. The GAM curve for the non-ADHD group is linear, in which the risk of drug use increases incrementally from approximately 10% at the lowest level of Externalizing to more than 30% at the highest level. In contrast, the GAM curve for children with ADHD is nonlinear (test for departure from line- arity, x2 = 7.45, df = 2.8, p = .05): children with ADHD and low levels of Externalizing problems were unlikely to have initiated drug use, whereas the incidence of drug use

" _ 0

2 - 7

0'

30 40 50 60 70 Total Externalizing T score

Fig. 3 Incidence of drug use by total Externalizing T score for children with and without attention-deficit hyperactivity disorder (ADHD), as estimated by generalized additive models.

1350 J . A M . A C A D . CHILD A D O L E S C . PSYCHIATRY, 38:11, N O V E M B E R 1999

A D H D AND EARLY DRUG USE

increased dramatically with increasing Externalizing problems within the moderate range and leveled off in the bor- derline clinical range of Externalizing problems. Thus, the largest difference between children with and without ADHD was found at moderate levels of Externalizing problems (Tscore = 50-59), in which the odds of drug initiation in children with ADHD versus without ADHD was 2.09 (34.5% versus 20.1%; x 2 = 4.86, df = I , p =

.027). LBW and NBW children with ADHD had similar patterns of association between Externalizing problems and drug use. There was no association between drug use and Internalizing problems at age 6.

Environmental RisWProtective Influences on Drug Use

Parent monitoring and peer drug use were associated with children’s drug use in a model that included ADHD/ externalizing status, sex, site of birth, and birth weight status. Low levels of parent monitoring signaled a 2-fold increase in the risk of drug initiation (RO = 2.2; 95% CI =

1.3-3.6 for the lowest versus highest tertiles (thirds) of parent monitoring). We found no evidence of an interaction between ADHD status and parent monitoring (x’ = 1.38,

Increasing levels of drug use by peers signaled increased risk of early drug use. The risk of drug use for children at the highest versus lowest tertiles of peer drug use was nearly 6-fold (RO = 5.7; 95% CI = 3.2-10.2), and there was no statistically significant interaction between ADHD and peer drug use (2’ = 2.88, df= 2,p = .237). In addition, we included mothers’ lifetime history of depression, anxiety disorder, and substance abuse or dependence (alcohol andlor illicit drugs), because prior research suggested that mothers with a history of psychiatric disorder might over- report and overgeneralize their children’s behavior problems (Chilcoat and Breslau, 1997). We found no change in the relationship of ADHD with children’s drug use, when mother‘s psychiatric history was controlled.

Additional Analysis: Drug Use and ADHD Pharmacotherapy

We tested whether children receiving medication for treatment of ADHD at age 11 were more likely to use drugs than children who did not take medication. Of the children with ADHD at age 6,30 (20.2%) received pharmacotherapy at age 11, based on mothers’ reports. Nearly all of these children were treated with methylphenidate (Ritalin@). The risk of drug use was nearly identical for children with ADHD who did and did not receive pharmacotherapy (incidence = 31 .O% and 28.2%, respectively).

df= 2,p = .501).

A small proportion of children who did not have ADHD diagnosed at age 6 years received pharmacological treatment for ADHD at age 11 years (n = 24,4.2%). The incidence of drug use in these children did not differ from the incidence in children not receiving treatment (incidence =

16.6% for both groups). Models that controlled for severity of ADHD symptoms did not detect a difference in the risk of drug use in treated and untreated children.

DISCUSSION

LBW and NBW children with ADHD were at increased risk for drug use by age 11 years. There was evidence of an interaction between ADHD and level of externalizing problems. Regardless of ADHD, children with low levels of externalizing problems were at low risk of initiation and children with high levels were at substantially higher risk. Only at moderate levels of externalizing problems did children with ADHD have a significantly higher risk of drug use, compared with children without ADHD. Parent monitoring and peer drug use appear to influence drug use in children with ADHD in the same way that they have been found previously to influence drug use among children in general. There was no evidence that medication used to treat ADHD increased children’s risk of drug initiation.

The findings from this prospective study add to the existing evidence that ADHD signals increased risk of drug use. Our results extend findings from previous reports in 2 important respects. First, to our knowledge, this is the only study that examined ADHD in relation to children’s initial drug use. Second, data were collected prospectively, from a community-based sample of children, using reports from multiple informants. By meas- uring ADHD and conduct problems when the children were 6 years old and drug use at age 11 years, we were able to gain control of the temporal sequence in which these phenomena occurred. The use of multiple informants, that is, mothers’ reports of ADHD and externalizing problems at age 6 and children’s reports of drug use at age 11, reduces the possibility of a reporting bias when this information is obtained from a single respondent. Furthermore, use of a community-based sample avoids the selection bias inherent in clinical samples.

It is clear from these results that externalizing problems play an important role in the observed relationship between ADHD and early drug use. The finding that at moderate levels of externalizing problems children with ADHD were at high risk of early drug use-as high as that

J . AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 38:11, NOVEMBER 1999 1351

CHILCOAT AND BRESLAU

of a h e n a hi& level of externalizing problems- s-6 &at ADHD lowers the threshold for high risk of drug use associated with externalizing problems. It has

suggested that children with ADHD are at increased & for condua problems, which in turn increase the risk for drug use (Mannuzza et al., 1993). We found that a strong association between ADHD and externalizing problems existed by the time the children in our sample were 6 Y ~ S old. If externalizing problems acted as a mediator in the causal pathway from ADHD to drug use, ADHD must have exerted a causal influence on conduct problems well before 6 years of age. On the other hand, ADHD might not have preceded the externalizing problems but instead both might be aspects of a constellation of behavior problems that put children at increased risk for drug use.

There is need for replication of our finding that at moderate levels of externalizing problems children with ADHD were at high risk of drug use. The use of GAM curves enabled the detection of a difference in the risk of drug use-had we relied on parametric models with a linear term for externalizing problems, we would have found no increase in risk of drug use due to ADHD, as did Lynskey and Fergusson (1995). Replication of this finding is needed using new data, as well as reanalysis of existing data using nonlinear statistical models.

Limitations

In this study we relied on mothers’ reports of ADHD and conduct problems when the children were 6 years old. Overall, the prevalence of ADHD, based on the DISC, was higher than estimates reported in other population- based studies (Goldman et al., 1998). We expected that the prevalence of ADHD in this study might be elevated due to the overrepresentation of LBW and urban children in our sample (Breslau, 1995) and because the prevalence of ADHD is highest early in childhood and declines as children grow older (Cohen et al., 1993; Szatmari et al., 1989). Nonetheless, the possibility that the DISC might have been overly sensitive in assessing ADHD in this sample cannot be ruled out. However, we would expect that measurement error due to overdiagnosing a disorder would tend to bias results in the direction of the null hypothesis (i.e., toward RO = I). In fact, if we required the presence of 9 symptoms as a more stringent definition of ADHD, the magnitude of the association with drug use would have been greater than that observed using the DSM-III-R threshold of 8 symptoms. It is possible that there could be a bias in favor of an association if mothers’

reports of ADHD on the DISC reflected a general tend- ency to report behavior problems in their children. In this case, we would expect that there would be no residual association between drug use and ADHD, once Ex- ternalizing problems were taken into account. Our findings indicate that this is not the case.

A potential concern is the possibility that ADHD might be manifested differently in LBW children than in NBW children. We found no support for this concern. Despite the higher incidence of drug use in LBW compared with NBW children, the association of ADHD with drug use was identical in both groups. Furthermore, we found that the profile of ADHD symptoms was identical in LBW and NBW children.

We focused on early drug use as a key outcome given prior evidence of a link between drug use that starts in childhood and later drug problems. Further follow-up will be needed to confirm this link in our sample and to measure the relationship between ADHD and later, more extensive drug involvement.

Clinical Implications

This study has several important implications for intervention. Early intervention to reduce drug use should tar- get children with ADHD who have moderate, as well as high, levels of early externalizing problems. Furthermore, regardless of ADHD status, children with high levels of externalizing problems at age 6 years are at very high risk for early drug initiation later in childhood. Preventive interventions should start early in childhood, inasmuch as the risk factors for early drug use observed in this study emerge by the time children enter elementary school. The presence of individual-level vulnerability markers, such as ADHD and conduct problems, does not override the influence of parent monitoring and peer drug use. Thus, interventions that seek to increase parent monitoring and reduce children’s association with drug-using peers might be effective in children at risk because of the presence of ADHD and behavior problems.

There was no evidence that stimulant treatment of ADHD increased children’s risk of early drug use, despite growing speculation of such an association (Goldman et al., 1998). It also has been suggested that psychostimulant medication in children with ADHD might have a protective effect (Loney, 1998). Our findings show neither an adverse nor a beneficial effect on the risk for drug initiation.

1352 J . AM. ACAD. C H I L D ADOLESC. PSYCHIATRY, 38:11, NOVEMBER 1999

A D H D A N D EARLY D R U G USE

Conclusion

Findings in this study fill a gap in the current understanding about the relationship of ADHD to children’s early drug use, an important and frequently overlooked stage in the development of drug use disorders. Further longitudinal research will extend these findings to test how ADHD influences children’s progression to more ad- vanced stages of drug use. Additional study is needed to understand the contributions of specific elements of ADHD, such as impulsivity, that might increase children’s vulnerability to drug use. Such an understanding can shed light on the mechanisms that lead to drug use, which might be relevant to preventing drug use in children in general.

REFERENCES Achenbach T M (1991), Manual for the Child Behavior Checklist/4-18 and

1991 Profile. Burlington: University of Vermont Department of Psychiatry Anthony JC, Petronis KR (1995), Early-onset drug use and risk of later drug

problems. Drug Alcohol Depend 40:9-15 Barkley RA, Fischer M, Edelbrock CS, Smallish L (l990), The adolescent out-

come of hyperactive children diagnosed by research criteria, I: an 8-year prospective follow-up study. / A m Acad ChildAdoLesc Pychiahy 29546-557

Barnes GM, Farrell M P (1992), Parental support and control as predictors of adolescent drinking, delinquency, and related problem behaviors. /Marriage Fam 54:763-776

Berkson J (1946), Limitations of the application of fourfold table analysis to hospital data. Biometrics 2:47-53

Biederman J, Newcorn J , Sprich S (1991), Comorbidiry of attention deficit hyperactivity disorder with conduct, depressive, anxiety, and other disorders. Am J Psychiatry 148:564-577

Biederman J, Wilens T, Mick E et al. (l997), Is A D H D a risk factor for psychoactive substance use disorders? Findings from a four-year prospective follow-up study. ] A m Acad ChildAdolesc Psychiatry 36:2 1-29

Biederman J, WilensT, Mick E, Milberger S, SpencerTJ, Faraone SV (1995), Psychoactive substance use disorders in adults with attention deficit hyperactivity disorder (ADHD): effects of A D H D and psychiatric comorbidity. Am /Psychiatry 152:1652-1658

Botvin GJ, Baker E, Dusenbury L, Bowin EM, Diaz T (1995), Long-term follow-up results of a randomized drug abuse prevention trial in a white middle-class population. / M A 273: 1106-11l2

Boyle MH, Offord DR, Racine YA, Fleming JE, Szatmari P, Links PS (1993), Predicting substance use in early adolescence based on parent and teacher assessments of childhood psychiatric disorders: results from the Ontario Child Health Study follow-up. J Child Psychol Psychiatry 34:535-544

Breslau N (1995), Psychiatric sequelae of low birth weight. Am J Epidemiol 17:96-106

Breslau N , Brown GG, DelDotto JE et al. (1996a), Psychiatric sequelae of low birth weight at six years of age. JAbnorm ChildPsychol24:385-400

Breslau N , Chilcoat H D , DelDotto J , Andreski P, Brown G (1996b), Low birthweight and neurocognitive status at six years of age. Biol Psychiatry 40:389-397

Breslau N, Fenn N, Peterson EL (1993), Early smoking initiation and nicotine dependence in a cohort of young adults. DrugAlcohol Depend33:129-137

Brown BB, Mounts N , Lamborn SD, Steinberg L (1993), Parenting practices and peer group affiliation in adolescence. Child Dev 64:467-482

Carroll KM, Rounsaville BJ (1993), History and significance of childhood attention deficit disorder in treatment-seeking cocaine abusers. Compr Psychiatry 34: 7 5-82

Chilcoat HD, Anthony J C (1996), Impact of parent monitoring on initiation of drug use through late childhood. J Am Acad ChiM Adolesc Psychiatry 35:9 1-100

Chilcoat HD, Breslau N (1997), Does psychiatric history bias mothers’ reports? An application of a new analytic report. / A m Acad ChildAdoLescfiychiatry 36~971-979

Chilcoat H D , Breslau N , Anthony JC (1996), Potential barriers to parent monitoring: social disadvantage, marital status, and maternal psychiatric disorder. / A m Acad Child Adolesc Psychiatry 35: 1673-1682

Chilcoat H D , Dishion TJ, Anthony J C (1995), Parent monitoring and the incidence of drug sampling in urban elementary school children. A m / Epidemioll4 1 : 2 5-3 1

Chilcoat H D , Schiitz C G (1996), Age-specific prevalence of hallucinogen use in the US household population. Drug Alcohol Depend 43: 143-153

Cohen P, Cohen J, Kasen S et al. (1993), An epidemiological study of disorders in late childhood and adolescence, I: age- and gender-specific prevalence. J Child Psychol Psychiatry 34:85 1-867

DeMilio L (1989), Psychiatric syndromes in adolescent substance abusers. Am JPsychiatry 146:1212-1214

DishionTJ, Patterson GR, Reid JR (1988), Parent and peer factors associated with drug sampling: implications for treatment. In: Implications for Prevention (NIDA Research Monograph Val 56, D H H S Publication ADM 91-1813). Washington, DC: US Government Printing Office

Ellickson PL, Bell RM (1998), Drug prevention in junior high: a multi-site longitudinal test. Science 247: 1299-1305

Escobedo LG, Marcus SE, Holtzman D, Giovino GA (1993), Sports partic- ipation, age at smoking initiation, and the risk of smoking among US high school students. /AMA 269:1391-1395

Eyre SL, Rounsaville BJ, Kleber H D (1982), History of childhood hyperactivity in a clinic population of opiate addicts.] Nerv Ment Dis 170:522-529

Fleming JP, Kellam SG, Brown C H (1982), Early predictors of age at first use of alcohol. Drug Alcohol Depend 9:285-303

Gittelman R, Mannuzza S, Shenker R, Bonagura N (1985), Hyperactive boys almost grown up, I: psychiatric status. Arch Gen Psychiaq 42:937-947

Goldman LS, Gene1 M, Bezman RJ, Slanetz PJ (1998), Diagnosis and rreat- ment of attention-deficit hyperactivity disorder in children and adolescents. JAMA 279: 1100-1 107

Goodwin DW, Schulsinger F, Hermansen L, Guze S, Winokur G (1975), Alcoholism and the hyperactive child syndrome. J Nerv Ment Dis 160: 349-3 5 3

Hastie TJ, Tibshirani RJ (1990), Generalized Additive Models. London: Chapman and Hall

Hawkins JD, Catalan0 RF, Miller JY (1992), Risk and protective factors for alcohol and other drug problems in adolescence and early adulthood: implications for abuse prevention. Psychol Bull 112:64-105

Horton A M , Fiscella RA, O’Connor K, Jackson M, Slone D (1987), Revised criteria for detecting alcoholic patients with attention deficit disorder, residual type. J Nerv Ment Dis 175:371-372

Jensen PS, Salzberg AD, Richters JE, Wacanbe HK (1993), Scales, diagnoses, and child psychopathology, I: CBCL and DISC relationships. / A m Acad Child Adolesc Psychiatry 32:397-406

Loney J (1998), Risk of treatment versus nontreatment. In: NIH Consensus Confirenre: Diagnosis and Treatment o f Attention Deficit Hyperactivitj Dis- order. Washington, DC: National Institutes of Health, pp 175-179

Lynskey MT, Fergusson DM (1995), Childhood conduct problems, attention deficit behaviors, and adolescent alcohol, tobacco, and illicit drug use. J Abnorm Child Psychol 23:28 1-302

Mannuzza S, Klein RG, Bessler A, Malloy R LaPadula M (1993), Add: outcome of hyperactive boys: educational achievement, occupational rank, and psychiatric status. Arch Gen Psychiatry 50565-576

Mannuzza S, Klein RG, Bonagura N , Malloy R Giampino TL, Addalli KA (1991), Hyperactive boys almost grown up. Arch Gen Psychiahy48:77-83

Milberger S, Biederman J, Faraone SV, Chen L, Jones J (1997a), A D H D is associated with early initiation of cigarette smoking in children and adolescents. J Am Acad Child Adolesc Psychiatry 36:37-44

Milberger S, Biederman J, Faraone SV, Chen L, Jones J (1997b), Further evidence of an association between attention-defidhyperactivity disorder and cigarette smoking. Am ]Addict 6:205-217

Offord DR, Boyle MH, Racine YA et al. (1992), Outcome, prognosis, and risk in a longitudinal follow-up study.JAm Acad ChildAdolesc Psychiatry 31 :916-923

J . A M . A C A D . C H I L D A D O L E S C . PSYCHIATRY, 38:11, N O V E M B E R 1993 1353

CHILCOAT A N D BRESLAU

Richardson JL, D y e r K, McGuigan K et al. (1989), Substance abuse among eighth-grade students who take care of themselves after school. Pediatrics 84:556-566

Robins LN, Przybeck TR (1985), Age at onset of drug use as a factor in drug and other disorders. In: Implicationsfor Prevention (NIDA Research Mono- graph 56). Washington, DC: US Government Printing Office, pp 178-192

Shaffer D, Schwab-Stone M, Fisher P, Davies M, Piacentini J , Giola P (1988). A Revised I/ersion of the Diagnostic Interview Schedule for Children (DISC- R): Results of a Field Trial and Proposakfor a New Instrument (DISC-2). Rockville, MD: National Institute of Mental Health

Steingard R, Biederman J, Doyle A, Sprich-Buckminster S (1992), Psychi- atric comorbidity in attention deficit disorder: impact on the interpreta- tion of Child Behavior Checklist results. ] A m Acad ChildAdolesr Psychiany 31 :449-454

Szatmari P, Offord DR, Boyle MH (1989), Ontario Child Health Study: prevalence of attention deficit disorder with hyperactivity. 1 ChildPsychol Psychiatry 30:2 19-230

Tarter R, McBride H, Buonpane N, Schneider D (1977), Differentiation of alcoholics according to childhood history of minimal brain dysfunction, family history, and drinking pattern. Arch Gen Prychiatry 34:761-768

Thompson LL, Riggs PD, Mikulich SK, Crowley TJ (1996), Contribution of ADHD symptoms to substance problems and delinquency in conduct- disordered adolescents. JAbnorm Child Psychol24:325-347

Whitmore EA, Mikulich SK, Thompson LL, Riggs PD, Aarons GA, Crowley TJ (1998), Influences on adolescent substance dependence: conduct disorder, depression, attention deficit hyperactivity disorder, and gender. Drug Alcohol Depend 47237-98

Wood D, Wender PH, Reimher RW (1983), Prevalence of attention deficit disorder, residual type, or minimal brain dysfunction, in a population of male alcoholic patients. A m ] Psychiatry 140:95-98

Yamaguchi K, Kandel DB (1984), Patterns of drug use from adolescence to young adulthood, 111: predictors of progression. Am 1 Public Health 74:673-68 1

Drinking and Driving Among US High School Seniors, 1984-1997. Patrick M. O’Malley, PhD, Lloyd D. Johnston, PhD

0bjectiver:This article reports the prevalence of, and trends in, driving after drinking and riding in a car with a driver who has been drinking among American high school seniors, based on data from more than a decade (1984-1997) of annual national surveys. Method: Logistic regressions were used to assess the effects of demographic factors (gender, region of country, population density, parental education, and race/ethnicity) and selected “lifestyle” factors (religious commitment, high school grades, truancy, illicit drug use, evenings out per week, and miles driven per week). Results; Rates of adolescent driving after drinking and riding with a driver who had been drinking declined significantly from the mid-1980s to the early or mid-1990s, but the declines have not con- tinued in recent years. Rates of driving or riding after drinking were higher among high school seniors who are male, White, living in the western and northeastern regions of the United States, and living in rural areas. Truancy, number of evenings out, and illicit drug use all related Significantly positively with the dependent variables, whereas grade point average and religious commitment had a negative relationship. Miles driven per week related positively to driving after drinking. Am J Public Health 1999;89:678-684. Copyright 1999 by the American Public Health Association.

1354 J . AM. ACAD. C H I L D ADOLESC. PSYCHIATRY, 38:11, NOVEMBER 1999