Transcript Predicting Violent Behavior in Adolescent Cannabis Users

Predicting Violent Behavior in
Adolescent Cannabis Users:
Correlates of and Changes in
Social Environment Over Time
Michelle White, M. S.*, Rod Funk, B. S.,*
Michael Dennis, Ph.D.*, Frank Tims, Ph.D.**
*Chestnut Health Systems, Bloomington, IL and
**Operation PAR, St. Petersburg, FL
Poster Session Presentation at the College on Problems of
Drug Dependence (CPDD) Sixty-Fifth Annual Meeting,
Bal Harbour, FL
June 14-19, 2003.
Abstract
Violence, aggression, and criminal offenses are common among
adolescent substance abusers. Using Moffitt’s (1993) taxonomy of
offending behavior theory, we examine which social environment factors
(e.g. peer group criminality and drug use) are correlated with criminal
activity and violence among adolescents entering substance abuse
treatment. We then predict how changes in social environment factors
affect criminality and violence over time. We use data from the Cannabis
Youth Treatment (CYT) multisite randomized field experiment of 600
adolescents meeting outpatient patient treatment placement criteria.
Follow-up was extended to 30-months post-intake through a CSAT
contract, the Persistent Effects of Treatment Study of Adolescents (PETSA). Our findings were consistent with the two groups hypothesized by
Moffitt in terms of types of crime, correlates, and long-term course of
behaviors. There was mixed evidence for the sensitivity of the high (lifecourse persistent) group to changes in social environment.
Aims
• To summarize Moffitt’s (1993) taxonomy
of behaviors related to criminality and
violence.
• To examine how this taxonomy is
correlated with social environment factors
(e.g. peer group criminality and drug use).
• To examine how well this taxonomy
predicts changes in criminal activity and
social environment factors.
Moffitt’s Taxonomy
• Life-Course Persistent Offenders: begin offending
early in life, commit many crimes & engage in violence,
have psychopathology factors (CD, other), and generally
continue to commit criminal/violent acts in spite of
improved social environment.
• Adolescence Limited Offenders: begin offending in
adolescence, generally engage in crimes of a petty or nonviolent nature (e.g., vandalism, property offenses), do not
have underlying psychopathology problems, and will
generally decrease or stop their illegal behaviors when
there are improvements in their social environment.
Data are from the Cannabis Youth Treatment
(CYT) Randomized Field Experiment
Coordinating Center:
Chestnut Health Systems, Bloomington, IL,
and Chicago, IL
University of Miami, Miami, FL
University of Conn. Health Center, Farmington, CT
Sites:
Univ. of Conn. Health Center, Farmington, CT
Operation PAR, St. Petersburg, FL
Chestnut Health Systems, Madison County, IL
Children’s Hosp. of Philadelphia, Phil. ,PA
Sponsored by: Center for Substance Abuse Treatment (CSAT), Substance Abuse and Mental Health Services Administration
(SAMHSA), U.S. Department of Health and Human Services
Design
• Target Population: Adolescents with marijuana
disorders who are appropriate for 1 to 3 months of
outpatient treatment.
• Inclusion Criteria: 12 to 18 year olds with
symptoms of cannabis abuse or dependence, past 90
day use, and meeting criteria for outpatient treatment.
• Data Sources: self report and collateral reports using
the Global Appraisal of Individual Needs (GAIN),
on-site and laboratory urine testing, therapist alliance
and discharge reports, staff service logs, and cost
analysis.
See www.chestnut.org/li/cyt/ for more information on
CYT and list of articles.
Implementation of Evaluation
• Over 85% of eligible families agreed to participate.
• Quarterly follow-up of 94 to 98% of the adolescents
from 3- to 12-months (88% all five interviews).
• Collateral interviews and urine test data were obtained
at intake, 3- and 6-months on over 85% of participants
(90% of the adolescents who were not incarcerated or
interviewed by phone).
• 90% completion in 30-month follow-up (N=599).
• Baseline taxonomy based on the GAIN’s Crime and
Violence Index (alpha=.9), categorized into three
groups 0-2 (low), 3-6 (moderate) and 7-31 (high).
• Environmental risk based on the GAIN’s Social Risk
Index (alpha=.7).
Crime and Violence Index
During the past 12 months, have you had a disagreement in which
you did the following things? (1=Yes, 0=No)
General Conflict Tactic Index (GCTI) - oral violence subscale
a. Discussed it calmly and settled the disagreement?
b. Left the room or area rather than argue?
c. Insulted, swore or cursed at someone?
d. Threatened to hit or throw something at another person
General Conflict Tactic Index (GCTI) - physical violence subscale
e. Actually threw something at someone?
g. Slapped another person?
h. Kicked, bit, or hit someone?
j. Hit or tried to hit anyone with something (an object)?
k. Beat up someone?
m. Threatened anyone with a knife or gun?
n. Actually used a knife or gun on another person?
During the past 12 months, how many times have you . .
Property Crime Index (PCI) subscale
1. purposely damaged or destroyed property that did not belong to you?
2. bought, received, possessed or sold any stolen goods?
3. passed bad checks, forged (or altered) a prescription or took money from an
employer?
4. taken something from a store without paying for it?
5. other than from a store, taken money or property that didn’t belong to you?
6. broken into a house or building to steal something or just to look around?
7. taken a car that didn’t belong to you?
Interpersonal Crime Index (ICI) subscale
8. used a weapon, force, or strong-arm methods to get money or things from a
person?
9. hit someone or got into a physical fight?
10. hurt someone badly enough they needed bandages or a doctor?
11. used a knife or gun or some other thing (like a club) to get something from
a person?
12. made someone have sex with you by force when they did not want to have
sex?
13. been involved in the death or murder of another person (including
accidents)?
14. intentionally set a building, car or other property on fire?
Drug Crime Index (DCI) subscale
15. driven a vehicle while under the influence of alcohol or illegal drugs?
16. sold, distributed or helped to make illegal drugs?
17. traded sex for food, drugs, or money?
18. been a member of a gang?
19. gambled illegally?
Scored 0 for none and 1 for one or more times. Each subscale is
sum of “types” of crime. Total is sum across all types.
Social Risk Index (SRI)
Of the people you have regularly socialized with or hung
out with in the past year, would you say that none, a few,
some, most or all of them...
a. were employed or in school or training full-time?
b. were involved in illegal activity?
c. weekly got drunk or had 5 or more drinks in a day?
d. used any drugs during the past 90 days?
e. shout, argue, and fight most weeks?
f. have ever been in drug or alcohol treatment?
g. would describe themselves as being in recovery?
With risk items (b,c,d,e) scored 0=none, 1=a few, 2=some,
3=most, 4=all and protective items (a, f, g) scored
4=none, 3=a few, 2=some, 1=most, 0=all
Validation of the CVI scale and subgroups
• The CVI and each of its subscales were internally
consistent (alpha = .7+ for subscales, .9 for total).
• Endorsement of all items and subscales increased with
the shift from low to moderate to high (with all those
with prevalence of 3% or more significant).
• Shifting from low to moderate was associated with
increased oral violence, property crime, and drug
related crime.
• Shifting from moderate to high was associated with
even more of these things, as well as more physical
violence and interpersonal (aka violent) crimes.
Behaviors by Crime/Violence Sub-Groups
Crime/Violence Index
In past year, have you had a disagreement in
which you :
Discussed it calmly and settled it
GCTI –oral violence
Left room or area rather than argue
Insulted or swore at someone
Threatened to hit or throw something
GCTI –Physical violence
Actually threw something at someone
Pushed grabbed or shoved someone
Slapped another person
Kicked, bit or hit someone
Hit or tried to hit anyone with something
Beat up someone
Threatened anyone with gun or knife
Actually used gun or knife on someone
GCI – General Crime Index
During the past year, have you:
GCI – Property Crime Index
Purposely damaged/destroyed property
Passed bad checks/forged prescription
Taken something from store w/o paying
Taken money/property didn't belong to you
Broken into house/building to steal
Taken a car that didn't belong to you
GCI – Interper. Crime Index
Used a weapon/force to get money
Hit someone or got into physical fight
Hurt someone badly they needed MD
Used knife/gun to get something
Made someone have sex with you/force
Been involved in death/murder of person
Intentionally set building/car/etc on fire
GCI – Drug Crime Index
Driven vehicle while under influence of AOD
Sold/distributed/made illegal drugs
Traded sex for food/drugs/money
Been a member of a gang
* p < .05, ** p < .01, *** p < .001.
CVIGP Crime/Violence groups
Low 0-2 Medium 3- High 7-29
(N=172) 6 (N=173) (N=240)
Total
Chi(N=585) Square
Gen. Conflict Tactic Index (GCTI)
8%
8%
5%
0%
0%
1%
1%
1%
0%
0%
0%
1%
61%
68%
73%
29%
12%
34%
10%
24%
6%
20%
1%
1%
68%
84%
94%
75%
49%
82%
48%
70%
37%
65%
10%
5%
49%
57%
61%
39%
24%
44%
23%
36%
17%
33%
4%
3%
df =2
161.26
247.67
349.12
243.57
149.97
273.25
149.04
225.71
117.20
209.74
29.85
9.78
3%
1%
2%
1%
2%
2%
0%
5%
1%
0%
0%
0%
1%
5%
4%
0%
0%
12%
2%
19%
8%
4%
6%
1%
18%
7%
1%
1%
1%
2%
14%
13%
0%
1%
42%
4%
39%
31%
21%
18%
9%
66%
29%
4%
0%
1%
3%
37%
48%
1%
8%
21%
3%
22%
15%
11%
10%
4%
34%
14%
2%
0%
1%
2%
21%
25%
1%
3%
103.65
5.22
79.81
77.79
49.00
34.35
26.82
197.62
76.18
11.96
2.39
1.38
2.20
68.84
122.66
4.34
25.02
***
***
***
***
***
***
***
***
***
***
***
**
***
***
***
***
***
***
***
***
**
***
***
***
Crime/Violence Subscales by Sub-Groups
12
10
8
6
4
2
0
Low 0-2 (N=172)
Medium 3-6 (N=173)
High 7-29 (N=240)
Expected Correlates of CVI subgroups
• There were no significant differences in the
demographic characteristics of the three groups.
• The low CVI group experienced less environmental
risk and fewer problems (including substance use and
HIV risk behaviors), but there were few differences
between the moderate and high group.
• The rates of more pathological problems (including
dependence, mental distress, traumatic distress,
ADHD, and conduct disorder) increased from low to
moderate to high.
Correlates of Crime/Violence Sub-Groups
Low 0-2 Medium 3-6 High 7-29
(N=172) (N=173)
(N=240)
Demographics
Female
African American
Caucasian
Hispanic
Other/Mixed
Non-white
12-14 Years Old
15-18 Years Old
Family
Single Parent Family
Weekly Alcohol Use in Home
Weekly Drug Use in Home
Social Peers
Regular Peer Alcohol Use at Work/School
Regular Peer Alcohol Use Socially
Regular Peer Drug Use at Work/School
Regular Peer Drug Use Socially
Environment
In school
Employed
Current CJ Involvement
Controlled Environment
Ever Been Victimized
Acute Victimization
Ever Homeless/Runaway
Total
(N=585)
Chi-square
14%
37%
56%
2%
5%
44%
17%
83%
20%
24%
66%
4%
6%
34%
15%
85%
18%
27%
64%
5%
5%
36%
14%
86%
18%
29%
62%
4%
5%
38%
15%
85%
2.49
9.34
55%
21%
6%
44%
22%
8%
50%
27%
16%
50%
24%
11%
4.04
2.72
11.59
**
45%
50%
69%
60%
65%
81%
62%
75%
83%
57%
64%
78%
10.00
25.95
8.78
**
***
*
80%
91%
94%
89%
19.47
***
89%
38%
62%
17%
39%
20%
4%
87%
53%
59%
25%
51%
29%
8%
86%
49%
64%
30%
74%
56%
12%
87%
47%
62%
25%
57%
37%
8%
0.83
8.40
0.84
8.47
53.92
58.69
6.66
4.22
1.07
*
*
***
***
*
Prior SA Treatment Episodes
None
One
2+ episodes
Drug Use:
Weekly Any Alcohol or Drug Use\3
Weekly Alcohol Use\3
Weekly Marijuana Use\3
Weekly Crack/Cocaine Use\3
Weekly Heroin/Opiod Use\3
Weekly Other Drug Use\3
Age of First Use Under 15
13+ days in Cont. Env.\3
Substance Severity:\5
No use
Use
Abuse
Dependence
Physiological Dependence
Biomedical:
Acute Health Problems\7
Pregnant within Past Year\8
HIV Risk:
Sexually Active\3
Multiple Sexual Partners \3
Unprotected Sex\3
Any Needle Use\3
Mental Health:
Acute Mental Distress\9
Acute Traumatic Distress\10
ADHD\11
Conduct Disorder\12
* p < .05, ** p < .01, *** p < .001.
78%
13%
9%
75%
13%
12%
73%
19%
9%
75%
15%
10%
4.91
69%
11%
64%
0%
0%
0%
81%
6%
73%
14%
65%
0%
0%
2%
84%
7%
82%
23%
81%
0%
0%
0%
87%
12%
76%
17%
72%
0%
0%
1%
85%
9%
10.01
11.58
19.23
N/A
N/A
4.24
2.92
4.44
**
**
***
0%
5%
59%
4%
31%
0%
4%
49%
5%
43%
0%
1%
38%
5%
55%
0%
3%
48%
5%
44%
27.81
***
20%
10%
26%
7%
30%
14%
26%
11%
5.49
0.39
60%
32%
16%
1%
72%
34%
19%
1%
80%
44%
33%
1%
72%
38%
24%
1%
19.72
8.07
17.69
0.34
***
*
***
15%
5%
25%
26%
24%
10%
36%
47%
37%
22%
49%
76%
27%
14%
38%
53%
27.14
28.35
25.35
102.09
***
***
***
***
Predictive Validity of the CVI subgroups
• The baseline CVI groups can also be used to predict
illegal activity 30 months after intake.
• Both the Moderate and High groups were more likely
than the low group to commit nonviolent crimes
(Odds ratio=1.3 and 1.6 respectively) but these groups
are not significantly different from each other.
• Those in the High group were significantly more
likely to have committed violent crimes than those in
the moderate (Odds ratio=2.8) or low (Odds ratio=4.5)
• Those in the High group were significantly more
likely to have committed 3 or more crimes than those
in the moderate (Odds ratio=3.4) or low (Odds
ratio=4.0) groups.
30 Month Criminal Activity
by Baseline Crime/Violence Sub-Groups
100%
90%
80%
70%
No crime
Incarcerated
Substance Use only
Non-violent crime
Violent crime
60%
50%
40%
X2(8)=18.36, p<.05
30%
20%
10%
0%
Low (n=150)
Moderate (n=158)
High (n=216)
Frequency of Criminal Activity at 30 Months
by Baseline Crime/Violence Sub-Groups
100%
90%
80%
70%
60%
No Crime
50%
1-2 Crimes
3+ Crimes
40%
X2(4)=24.56,
p<.001
30%
20%
10%
0%
Low (n=150)
Moderate (n=158)
High (n=216)
Changes in Social Environment Over Time
• Both the moderate and high group significantly reduced
their illegal activities between intake and 30-months postintake.
• In both groups, increased social risk was associated with
continued illegal activity. Same risk and reduced risk over
time was associated with increasingly larger reductions in
illegal activity.
• The model supports Moffitt’s theory for the moderate
(adolescence-limited) group (criminal activity decreased
with changes in social environment).
• For the high risk group, the findings were mixed:
– The high CVI group was at much higher risk of continued
involvement in spite of changes in social risk (as expected)
– When there were reductions in social environmental risks,
however, the high risk group did reduce their illegal activity (not
expected)
Change in Illegal Activity by CVI group and
Change in Social Environment Over Time
M o d erate C V I
H ig h C V I*
0 .3 5
Illegal A ctivity Index
0 .3 0
0 .2 5
0 .2 0
0 .1 5
0 .1 0
0 .0 5
0 .0 0
In tak e
3 0 M o n th s
In tak e
3 0 M o n th s
M o d - D e c re a se d R isk (N = 3 2 )
H ig h - D e c re a se d R isk (N = 5 4 )
M o d - S a m e R isk (N = 7 8 )
H ig h - S a m e R isk (N = 9 6 )
M o d - In c re a se d R isk (N = 2 1 )
H ig h - In c re a se d R isk (N = 1 5 )
* S ign ifican t tim e b y ch an ge in S o cial R isk . F (2 ,1 6 2 )= 4 .6 2 , p < .0 5 .
Discussion
• The GAIN’s CVI scale appears to be face
valid and internally consistent.
• The CVI’s subgroup typology provides a
simple taxonomy that is consistent with the
two groups hypothesized by Moffitt in
terms of types of crime, correlates, and
long-term course of behaviors.
• It provides mixed evidence for the
sensitivity of the high group to changes in
social environment.
Limitations and Next Steps
• This analysis was limited to self report data
and should ideally be replicated.
• During the coming months we will be doing
more work to examine
– cluster analysis of trajectories across all of the
longitudinal observations at intake, 3, 6, 9, 12
and 30 months.
– PATH analysis of the role of changing
environmental factors on this process.
– the comparability of these measures with
collateral reports and records.
Further Information and
Acknowledgement
• For further information contact: Michelle White, M.
S., Chestnut Health Systems, 720 W. Chestnut St.,
Bloomington, IL 61701, [email protected].
(309) 827-6026
• Preparation of this manuscript was supported by
funding from the Center for Substance Abuse
Treatment (CSAT) through the Persistent Effects of
Treatment Study (PETS, Contract No. 270-97-7011, as
well as Grant #s TI11317, TI11320, TI11321, TI11323,
and TI11324, TI11422, TI11433, and TI11432)
• The content of this publication does not necessarily
reflect the views or policies of the government