Findings From the Northwestern Juvenile Project

The study reported in this Bulletin uses self-report and urinalysis results from the Northwestern Juvenile Project to compare the veracity of self-reported drug use across demographic groups and to identify the best approach for assessing drug use among juvenile detainees. The study focuses on self-report and urinalysis for three reasons: they are the most widely used, best understood measures of substance use; they can be administered by trained lay personnel (a practical consideration for juvenile justice facilities); and the level of agreement and disagreement between the two measures is relatively well documented.

This section describes the study’s sample and methods, presents an overview of findings, and then looks at findings for the two substances most commonly used by juvenile detainees—cannabis and cocaine—by subject characteristics (gender, race/ethnicity, age, prior treatment for substance abuse, and recent drug charges).

Sample and Methods

Sample. The Northwestern Juvenile Project is an ongoing longitudinal study of high-risk youth sampled from the Cook County Juvenile Temporary Detention Center (CCJTDC) in Chicago, IL. CCJTDC was selected for study because of its urban location (most juvenile detainees nationwide are in urban areas) and because Cook County is ethnically diverse and has a large Hispanic population (Hispanics are the largest minority group in the United States and are overrepresented in the juvenile justice system). In addition, CCJTDC’s size (approximately 8,500 admissions each year, daily census of 650, and daily intake of 20) ensured adequate numbers of subjects for key subgroups such as females and Hispanics.

The CCJTDC sample was stratified by gender, race/ethnicity (African American, non-Hispanic white, Hispanic), age (10–13 or 14 and older), and legal status (processed as a juvenile or as an adult). All estimates reported in this Bulletin were weighted to reflect the CCJTDC sample (Little and Schenker, 1995; Cochran, 1977).

Initial sampling and baseline interviews were conducted between November 1995 and June 1998. The final sample size for the project was 1,829. Additional information on the sample and study design is available elsewhere (Teplin et al., 2002; Abram et al., 2003; Teplin et al., 2003).

Interviews and urine samples. Trained interviewers used the Diagnostic Interview Schedule for Children (DISC 2.3) to gather information on substance use. Subjects were assured that the information they revealed would remain confidential and would have no bearing on their legal proceedings. Interviewers built rapport with subjects during the questions that preceded the substance use items (basic demographic information, education, life circumstances, and DISC mental health items). Female subjects were assigned female interviewers, and Spanish-speaking subjects were assigned bilingual interviewers.

The DISC instrument measures use of alcohol, cannabis/hashish, uppers/speed, downers (e.g., barbiturates), tranquilizers (e.g., Valium, Librium, and Ativan), heroin and opiates (including morphine, methadone, and codeine), cocaine/crack cocaine, hallucinogens (including LSD and PCP), and inhalants (e.g., glue). The instrument asks about lifetime use, age at first use, frequency of use in the past year, any use in the past 6 months, and treatment. Because it is designed to assess patterns of use, it does not ask about use in the past 2 or 3 days.

Of the 1,829 subjects interviewed, 1,745 (about 95 percent) provided urine samples.6 Some subjects were unwilling or unable to provide samples, and some samples were unavailable for administrative reasons.7 The final sample size for the analyses reported in this Bulletin is 1,742 (3 subjects did not complete the DISC portion of the interview).8

Urinalysis. Enzyme-Multiplied Immunoassay Tests (EMIT) were used to identify illicit drug use by the subjects. The EMIT–10 panel tests for the presence of amphetamines, barbiturates, benzodiazepines (diazepam, nordiazepam, oxazepam, chlordiazepoxide, and norflurazepam), cannabis, cocaine (benzoylecgonine is a cocaine metabolite found in the urine), methaqualone (Quaaludes), opiates (including codeine and morphine), PCP, and propoxyphene (Darvon). Some EMIT–10 panel categories are not exact matches for self-report categories.9

Drug charges. Data on recent drug charges (e.g., possession or sale of controlled substances) against subjects were collected from the Cook County Court Clerk’s computer system. The period of coverage for these data was from 90 days prior to intake at CCJTDC to 30 days after intake (to account for lagtime between arrests and charges).

Overview of Findings

Table 2 presents an overview of findings—self-reported use, urinalysis results, and computations of veracity, prevalence, and bias (see “Definitions” sidebar)—among the total sample of all detained juveniles.

Table 2: Drug Use by Detained Juveniles in Cook County, IL: Overview of Findings (%)
 

Self-Reported Substance Use

Substance Detected
by
Urinalysis

Veracity
Minimum Prevalence

Minimum Bias for Self-Reporting

Minimum Bias for Urinalysis
Drug
6-Month
Lifetime
6-Month
Lifetime
6-Month
Lifetime
6-Month
Lifetime
6-Month
Lifetime
Any substance
77.3
90.1
66.4
87.7
94.1
85.4
94.0
8.2
3.9
19.0
27.6
Cannabis
77.1
90.1
65.9
87.6
94.1
85.3
94.0
8.2
3.9
19.4
28.1
Other than cannabis
8.0
13.0
5.8
27.5
37.8
12.2
16.6
4.2
3.6
6.4
10.9
    Cocaine
4.4
6.2
4.8
21.7
28.1
8.1
9.6
3.6
3.4
3.4
4.8
    Opiates
1.2
2.6
0.2
65.7
65.7
1.3
2.6
0.1
0.1
1.1
2.5
    Uppers
0.8
1.8
0.0
0.8
1.8
    Downers
0.3
0.5
0.0
0.3
0.5
    Tranquilizers
0.3
0.6
0.0
0.3
0.6
    Hallucinogens
5.5
9.4
1.1
33.3
39.3
6.1
10.1

Note: N=1,742. For explanations of veracity, minimum prevalence, and minimum bias, see “Definitions” sidebar.

* Veracity cannot be computed because the EMIT urinalysis result is zero.

• Bias figures are not presented because self-reported use is very limited and bias computations would not, therefore, be meaningful.

‡ Bias figures are not presented because the EMIT urinalysis panel detects PCP but not the other hallucinogens in the DISC self-report instrument and bias computations would, therefore, be misleading.

Self-reported use. Self-reported use of any substance was quite common. Use in the past 6 months was reported by 77.3 percent of youth, lifetime use by 90.1 percent. The figures for cannabis were about the same (77.1 percent and 90.1 percent, respectively), an indication that almost all youth who report any substance use report cannabis use. Self-reported use of substances other than cannabis was much less common: 8.0 percent and 13.0 percent (6-month and lifetime, respectively) for any other substance, 5.5 percent and 9.4 percent for hallucinogens, 4.4 percent and 6.2 percent for cocaine, and even less for the other categories measured (opiates, uppers, downers, and tranquilizers).

Definitions

The measures reported in tables 2, 3, and 4 and analyzed in the accompanying text are defined below.

Self-reported use: percentage of detained youth who reported substance use.

Urinalysis results: percentage of detained youth who tested positive for substances in EMIT–10 urinalysis.

Veracity: percentage of detained youth who tested positive by urinalysis who also reported use.

Minimum prevalence: estimated prevalence of substance use among detained youth, based on combined self-report and urinalysis results.

Minimum bias for self-report: estimated percentage of detained youth who use drugs and would go undetected if self-report is used without urinalysis (calculated as the difference between minimum prevalence and self-reported use).

Minimum bias for urinalysis: estimated percentage of detained youth who use drugs and would go undetected if urinalysis is used without self-report (calculated as the difference between minimum prevalence and urinalysis results).

Self-reported use and the related measures of veracity, prevalence, and bias are presented separately for use in the past 6 months and ever (lifetime). Prevalence and bias estimates are “minimum” because both self-report and urinalysis measures are expected to underestimate true rates of substance use.

Urinalysis results. In many ways, patterns of use detected by EMIT–10 urinalysis were similar to patterns of self-reported use. However, use rates based on urinalysis results generally were lower than those based on self-reports of 6-month and lifetime use. This is not surprising, given that in urinalysis, the window of sensitivity to drugs generally is only 2–3 days. Two-thirds (66.4 percent) of detainees tested positive for any drug, 65.9 percent for cannabis, 5.8 percent for any substance other than cannabis, 4.8 percent for cocaine, and 1.1 percent for hallucinogens.

Veracity. Overall, self-reporting had a surprisingly high level of veracity. Among all detainees with positive urinalysis results, 87.7 percent reported use of any substance in the past 6 months and 94.1 percent reported lifetime use of any substance. However, although the veracity figures for cannabis (87.6 percent for 6 months and 94.1 percent for lifetime) were nearly identical to the overall figures, those for any drug other than cannabis were much lower (27.5 percent and 37.8 percent). This means that cannabis alone was responsible for the high level of veracity in self-reporting of overall drug use. More than three in five detainees who tested positive for use of illicit substances other than cannabis did not accurately report their use.

Minimum prevalence. Based on a combination of self-report and urinalysis results, the minimum prevalence estimates show that at least 85.4 percent of detained youth had used some kind of illicit substance in the past 6 months and at least 94 percent had used an illicit substance at some point in their lifetime. For substances other than cannabis, however, minimum prevalence was much lower: 12.2 percent for 6 months and 16.6 percent for lifetime.

Minimum bias for self-report. This measure, which is the difference between the minimum prevalence estimate and the self-report estimate, indicates that self-reporting without urinalysis overlooked at least 8.2 percent of detained youth who had used some kind of illicit substance in the past 6 months, or almost 1 in 10 youth in detention. This bias drops to 3.9 percent for lifetime use. For substances other than cannabis, self-reporting without urinalysis overlooked 4.2 percent of youth who had used a substance in the past 6 months and 3.6 percent of those with lifetime use. These values were low because overall use of these substances was low. Nevertheless, as the bias and prevalence figures show, self-reporting overlooked one-fourth (3.6/16.6) to one-third (4.2/12.2) of juveniles who had used substances other than cannabis.

For cocaine specifically, 6-month self-reporting overlooked 3.6 percent of detainees and lifetime self-reporting overlooked 3.4 percent of detainees. The 6-month minimum prevalence for cocaine use was only 8.1 percent, so overlooking 3.6 percent means missing more than one-third of confirmed cocaine users.

For uppers, downers, and tranquilizers, self-reported use was less than 2 percent and EMIT urinalysis findings were zero. Therefore, veracity could not be computed for these substances, and table 2 does not present self-report bias figures for them.

Because very few detainees tested positive for hallucinogens, very few users of hallucinogens would be missed by self-reporting alone. As noted in endnote 9, the EMIT panel detects PCP but not the other hallucinogens in the DISC questionnaire. Therefore, the minimum prevalence estimates for hallucinogens (6.1 percent and 10.1 percent for 6-month and lifetime use, respectively) are biased downward relative to estimates for the other categories of substances. For these reasons, table 2 does not present self-report bias figures for hallucinogens.

Minimum bias for urinalysis. Minimum biases were greater for urinalysis than for self-reporting, in part because urinalysis has a limited window of sensitivity to drugs. When not combined with self-reporting, urinalysis overlooked one in five detained youth (19.0 percent) who reported using any substance in the past 6 months and one in four (27.6 percent) who reported lifetime use. For substances other than cannabis, urinalysis overlooked 6.4 percent of self-reported 6-month users and 10.9 percent of lifetime users. Table 2 does not present urinalysis bias figures for uppers, downers, tranquilizers, and hallucinogens for the same reasons it does not present self-report bias figures for these substances (see above). (The urinalysis bias for hallucinogens was quite large, but this is probably because of the limited scope of the EMIT test for hallucinogens.)

Age at Onset of Substance Use

Juvenile detainees who reported using drugs were asked about age at first use. Their response is disturbing. Ten percent of the youth who reported using drugs said they first used them at or before age 11, and 25 percent reported first use at or before age 12. Ten percent of youth who reported using cocaine said they first used it before age 11, and 50 percent reported first use before age 15.

Analysis: Cannabis

This section analyzes findings for cannabis. See table 3 and “Definitions” sidebar.

Table 3: Drug Use by Detained Juveniles in Cook County, IL: Analysis of Findings for Cannabis (%)
 
Self-Reported Substance Use

Substance Detected
by
Urinalysis

Veracity
Minimum Prevalence
Minimum Bias for Self-Reporting
Minimum Bias for Urinalysis
 
6-Month
Lifetime
6-Month
Lifetime
6-Month
Lifetime
6-Month
Lifetime
6-Month
Lifetime
N
MALE
77.3
90.2
67.4
87.6
94.2
85.7
94.1
8.4
3.9
18.2
26.7
1,126
Race
African American
77.1
89.9
69.9
86.8
93.7
86.3
94.3
9.2
4.4
16.4
24.4
552
Non-Hispanic white
81.0
91.7
59.5
94.6
97.8
84.2
93.0
3.2
1.3
24.4
33.6
199
Hispanic
78.1
91.6
58.9
91.1
97.3
83.3
93.2
5.2
1.6
24.4
34.3
372
Age
10–13
56.7
70.3
46.6
70.5
81.4
70.5
79.0
13.8
8.7
23.9
32.3
306
14–15
75.5
88.2
69.0
85.0
90.6
85.9
94.6
10.4
6.5
16.9
25.7
344
16+
82.5
95.6
70.0
91.8
98.7
88.3
96.5
5.7
0.9
18.2
26.5
476
Prior treatment
No
95.6
98.2
74.8
97.5
99.0
97.4
99.0
1.8
0.7
22.6
24.2
616
Yes
99.3
99.7
87.9
100.0
100.0
99.3
99.7
0.0
0.0
11.4
11.9
158
Recent drug charges
None
74.7
88.5
60.6
88.2
95.0
81.8
91.5
7.1
3.0
21.2
30.9
857
Any
82.4
93.6
81.2
86.5
93.0
93.4
99.3
10.9
5.7
12.1
18.1
269
2+
84.9
97.7
77.5
89.8
98.8
92.8
98.6
7.9
0.9
15.3
21.1
87
FEMALE
75.1
88.5
45.5
88.6
92.7
80.3
91.8
5.2
3.3
34.9
46.4
616
Race
African American
71.6
86.1
46.1
87.1
90.8
77.6
90.3
6.0
4.2
31.5
44.2
404
Non-Hispanic white
85.4
96.3
47.5
92.3
97.4
89.0
97.6
3.7
1.2
41.6
50.1
81
Hispanic
81.2
93.1
46.6
93.0
97.8
84.5
94.1
3.3
1.0
37.9
47.5
130
Age
10–13
50.8
68.3
22.4
71.1
71.1
57.3
74.8
6.5
6.5
34.9
52.4
50
14–15
76.5
89.1
44.8
90.4
93.6
80.8
92.0
4.3
2.9
36.0
47.1
334
16+
78.4
92.1
51.2
88.1
93.7
84.5
95.3
6.1
3.3
33.3
44.1
232
Prior treatment
No
94.4
98.5
55.5
98.9
99.4
95.0
98.8
0.6
0.3
39.4
43.3
311
Yes
97.8
100.0
57.4
97.8
100.0
99.1
100.0
1.3
0.0
41.7
42.6
94
Recent drug charges
None
74.2
88.0
42.6
89.4
93.1
78.7
91.0
4.5
2.9
36.1
48.4
547
Any
82.6
92.4
67.2
84.7
90.9
92.8
98.5
10.2
6.1
25.7
31.3
69
2+
86.7
92.2
67.9
88.5
88.5
94.5
100.0
7.8
7.8
26.6
32.1
14
Note: For explanations of veracity, minimum prevalence, and minimum bias, see “Definitions” sidebar.

Gender

  • Although self-reported use of cannabis was about the same for boys and girls (around 75 percent for 6-month use and 90 percent for lifetime use), urinalysis results differed dramatically: 67.4 percent for boys and 45.5 percent for girls. This suggests either that cannabis use was less common among girls than boys or that the temporal association between using cannabis and being arrested was stronger for boys than for girls.

  • Veracity in reporting cannabis use was about the same for boys and girls, i.e., most detainees with positive urinalysis results also reported use. However, because boys had a higher rate of positive urinalysis results, their 6-month self-report bias (8.4 percent) was greater than that for girls (5.2 percent). In other words, cannabis use during the past 6 months was more likely to go undetected among boys than among girls if self-reporting was used without urinalysis. Lifetime self-report bias was similar for boys (3.9 percent) and girls (3.3 percent).

  • Urinalysis biases were much higher for girls (34.9 percent for 6 months and 46.4 percent for lifetime) than for boys (18.2 percent 6 months, 26.7 percent lifetime), which means that cannabis use would be much more likely to go undetected among girls than among boys if urinalysis testing is used without self-reporting. This again suggests that recent cannabis use was less common among girls than among boys.

Race/ethnicity

  • Among males, self-reported cannabis use was roughly equal for all three racial/ethnic groups.

  • Boys in all three racial/ethnic groups had reasonably good veracity for reporting cannabis use, although 6-month veracity was slightly lower for African American boys (86.8 percent) than for other boys (94.6 percent for non-Hispanic whites and 91.1 percent for Hispanics). Because the rate of positive urinalysis results was higher for African American boys than other boys, the 6-month self-report bias was greater for African Americans (9.2 percent) than for others (3.2 percent for non-Hispanic whites and 5.2 percent for Hispanics).

  • Among boys, urinalysis bias for detecting cannabis use was lower for African Americans (24.4 percent for lifetime use) than for other groups (33.6 percent for non-Hispanic whites and 34.3 for Hispanics). The proportion of lifetime cannabis users who would be overlooked when urinalysis is used without self-report methods is smaller for African Americans than for other racial/ethnic groups.

  • Unlike males, female detainees in the three racial/ethnic groups differed substantially in their rates of self-reported cannabis use. Only 71.6 percent of African American girls reported use in the past 6 months, compared with 85.4 percent of non-Hispanic whites and 81.2 percent of Hispanics. Lifetime self-report rates also were lower for African American girls than for other girls.

  • Because of the differences in self-reported use, the 6-month self-report bias for cannabis was greater for African American girls (6 percent) than for other girls (3.7 percent for non-Hispanic whites and 3.3 percent for Hispanics). The 6-month urinalysis bias was greater for non-Hispanic white girls (41.6 percent) than for African American girls (31.5 percent) and Hispanic girls (37.9 percent). In short, self-reporting alone was most likely to overlook cannabis use among African American girls, whereas urinalysis alone was most likely to overlook its use among non-Hispanic white girls. This finding suggests that among girls, African Americans’ self-reports of cannabis use are the least reliable.

Age

  • Among both boys and girls, self-reported cannabis use was less common for detainees ages 10–13 than for older detainees. Among detainees ages 10–13, use in the past 6 months was reported by 56.7 percent of boys and 50.8 percent of girls (compared with 77.3 percent of all boys and 75.1 percent of all girls).

  • Among boys, biases for both self-report and urinalysis varied greatly by age: compared with older boys, those ages 10–13 had greater biases (i.e., their cannabis use was more likely to go undetected when either detection method was used alone). In contrast, biases among girls did not vary greatly by age.

Prior treatment for substance abuse

  • Almost all boys and girls who reported prior treatment for substance use had very substantial self-reported cannabis use and consequently had very low self-report bias.

  • Urinalysis detection for cannabis was much more common for boys who reported prior treatment (87.9 percent) than for girls who reported prior treatment (57.4 percent). Because so many boys who reported prior treatment tested positive, their urinalysis bias (11.9 percent for lifetime use) was much smaller than that for girls who reported prior treatment (42.6 percent for lifetime use).

Recent arrests for drug offenses

  • Self-reported cannabis use and positive urinalysis results for cannabis were more common among detainees with recent drug arrests than among those with no recent drug arrests.


  • Detainees with recent drug arrests were quite forthcoming about their cannabis use: veracity for self-reported use in the past 6 months was well above 80 percent for both males and females, and lifetime veracity was above 90 percent.

  • For both 6-month and lifetime use, minimum prevalence exceeded 90 percent for male and female detainees with recent drug arrests.

Analysis: Cocaine

This section analyzes findings for cocaine. Seetable 4 and “Definitions” sidebar.

Table 4: Drug Use by Detained Juveniles in Cook County, IL: Analysis of Findings for Cocaine (%)
 
Self-Reported Substance Use
Substance Detected
by
Urinalysis
Veracity
Minimum Prevalence
Minimum Bias for Self-Reporting
Minimum Bias for Urinalysis
 
6-Month
Lifetime
6-Month
Lifetime
6-Month
Lifetime
6-Month
Lifetime
6-Month
Lifetime
N
MALE
4.0
5.7
4.9
20.8
27.4
7.8
9.2
3.7
3.5
3.0
4.3
1,126
Race
African American
0.4
0.5
3.2
0.0
0.0
3.7
3.8
3.2
3.2
0.4
0.5
552
Non-Hispanic white
21.5
29.5
7.7
61.5
78.6
24.3
31.2
2.8
1.6
17.1
23.5
199
Hispanic
16.2
23.1
12.1
42.0
52.7
22.8
28.8
6.6
5.7
11.5
16.7
372
Age
10–13
1.0
1.6
2.1
5.9
5.9
3.0
3.6
2.0
2.0
0.9
1.5
306
14–15
3.2
4.5
4.5
16.0
19.8
6.8
8.1
3.6
3.6
2.5
3.6
344
16+
5.3
7.4
5.7
25.0
33.9
9.5
11.2
4.2
3.8
3.9
5.5
476
Prior treatment
No
4.7
6.6
4.4
24.9
31.9
7.9
9.6
3.2
3.0
3.6
5.2
616
Yes
10.0
12.9
9.1
33.0
38.1
15.7
18.6
5.7
5.7
7.2
9.4
158
Recent drug charges
None
5.6
7.7
3.8
33.2
40.6
8.0
10.0
2.5
2.3
4.3
6.2
857
Any
1.0
1.5
7.0
7.3
13.0
7.2
7.6
6.3
6.1
0.5
0.6
269
2+
0.4
0.4
5.0
8.7
8.7
5.0
5.0
4.6
4.6
0.0
0.0
87
FEMALE
9.6
12.4
3.7
36.2
39.4
12.0
14.6
2.4
2.3
8.3
10.9
616
Race
African American
1.5
1.5
2.7
0.0
0.0
4.2
4.2
2.7
2.7
1.5
1.5
404
Non-Hispanic white
26.7
44.0
7.3
83.3
100.0
28.0
44.0
1.3
0.0
20.4
36.7
81
Hispanic
24.7
30.2