Transcript Meta-Analysis of HIV Risk Reduction Interventions Within

Meta-Analysis of HIV Risk
Reduction Interventions Within
Drug Abuse Treatment Programs
By: Dana Trent
June 2001
http://gateway1.ovid.com/ovidweb.cgi
Drug Abuse: Treatment Needed
For A Host of Problems
Unemployment
Increased mortality
Family and social
problems
Our Focus: HIV
Transmission!!!
Goals of this Meta-Analysis
Primary Goal:
Are the HIV risk-reduction
interventions in drug treatment
programs effective?
Secondary Goals:
What will we find if we test
hypotheses regarding moderators
of effect size in past research?
Are there additional variables
impacting program effectiveness
that could direct future research?
Study Selection Criteria
Were intended to reduce HIV
risk behavior
Conducted in the USA or
Canada
Published/Dated between
1985-1998
Written in English
Participants’ primary problem:
illicit drug use/dependence
Study Selection Criteria Cont…
Study Design:
Intervention+Comparison Group
At least one HIV dependent variable
Effect size calculability
Random Assignment or:
Pre-intervention differences/similarities stated
The Source Hunt
Database from past
meta-analysis
University Library
Requests to Authors
Unpublished studies
Dissertation
purchases
92 references
identified, but…
Methodology
Only 18 eligible
studies coded
Coded On:
Study content
Methodology
Participant and
treatment
characteristics
Implementation and
integrity
Dependent variable
characteristics
Effect size calculation
Data Analysis
HIV Interventions
Knowledge, attitudes,
beliefs
Sexual Behavior
Risk-reduction Skills
Injection Practices
Various Moderators
Intervention intensity, total
sample size, researcher
role, etc…
Exploratory Variables
Didactic lecture, ethnic
group, number of
techniques used, etc…
Data Analysis
Effect size
Every HIV risk dependent variable, moderators,
and additional variables.
Corrected and weighted appropriately
Combined for overall average
Tested for significance
Results
Variable
d
K
unweighted
Overall
16
.34
Knowledge
attitudes,
beliefs
11
Sexual
Behavior
SD
d
weighted
95%
Confidence
Interval
Q
Max
Minimum
.16
.64
.03
.31*
.20 to
.42
8.42
.35
.17
.58
.09
.31*
.18 to
.45
7.44
14
.30
.20
.64
-.08
.26*
.15 to
.38
8.22
RiskReduction
Skills
6
.63
.28
.99
.32
.62*
.45 to
.79
9.10
Injection
Practices
5
.09
.14
.29
-.03
.04
-.14 to
.22
1.24
Results
Effective HIV
interventions :
Knowledge, beliefs, and
attitudes
Sexual Behavior
Risk-reduction skills
Non-effective HIV
interventions:
Injection Practices
Results
Moderators of effect
size in past research:
*p<.05
Variable
Overall
Effect Size
Intensity of
Contact
.57*
Results
Variable
Future
Research/
Exploratory
Analysis
*p<.05
**p<.01
Overall
Effect Size
Sexual
Behavior
Ethnic Group Majority
.67*
.72*
Didactic Lecture
.75**
.59*
Separate Gender Sessions
.69**
Self-control (coping) skills
.59*
Peer Group
Discussion/Counseling
.55*
.67**
# of technique types used
.60*
.63*
Stage of drug treatment
when intervention was
delivered
.68*
Discussion
 HIV interventions DO have a clinically relevant
impact!
2/3rds non-significant individually!
– Insufficient power??
– Publication bias??
Discussion
Only one moderator:
Intervention intensity
– Time not as
important?
Discussion
Ethnic Group Majority
Greater impact on
White participants.
Minorities enter with
lower risk.
Didactic Lectures
Intended to teach,
particularly in having
moral instruction as an
ulterior motive.
Non-significant for
Knowledge, attitudes,
and beliefs??
Discussion
Separate for gender
More open discussion
Self-Control/Coping
Aimed at changing sexual
behavior
Peer Group Discussion
Social pressure and
support is powerful, but…
Patients were stabilized
and receiving counseling
in variety of other
contexts!
Discussion
Number of techniques
used
The best had 6 or more
No relation to hours of
contact??
Stage of Drug
Treatment
What about those
outside of treatment
system?
Future Research and
Recommendations
Access outcomes for
longer periods of time
More studies
Designing new HIV
risk-reduction
interventions
Including all that tested
significant