Skip navigation
CLN bookstore

Wsipp Evidence Based Substance Abuse Mental Health Treatment Washington State 2006

Download original document:
Brief thumbnail
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
Washington State
Institute for
Public Policy
110 Fifth Avenue Southeast, Suite 214

•

PO Box 40999

•

Olympia, WA 98504-0999 •

(360) 586-2677

•

www.wsipp.wa.gov

June 2006

EVIDENCE-BASED TREATMENT OF ALCOHOL, DRUG, AND MENTAL HEALTH DISORDERS:
POTENTIAL BENEFITS, COSTS, AND FISCAL IMPACTS FOR WASHINGTON STATE
During the mid-1990s, the Washington State
legislature began to enact statutes to promote an
“evidence-based” approach to several public policies.
While the term evidence-based has not always been
precisely defined in legislation, it has generally been
constructed to describe a program or policy supported
by a rigorous outcome evaluation clearly
demonstrating effectiveness. Additionally, to
determine if taxpayers receive an adequate return on
investment, the legislature has also started to require
benefit-cost analyses of certain state-funded
programs and practices.
Washington’s initial experiments with evidence-based
and cost-beneficial public policies began in the state’s
juvenile justice system. The legislature funded
several nationally known and rigorously researched
programs designed to reduce the reoffending rates of
juveniles. At the same time, the legislature eliminated
the funding of a juvenile justice program when a
careful evaluation revealed that it was failing to
reduce juvenile crime. Thus, the term evidencebased does not mean simply adding new programs, it
also means eliminating programs when research
indicates they do not work.
Following this successful venture into evidencebased public policy, Washington began to introduce
the approach in other fields including adult
corrections, child welfare, and K–12 education.
Extending the Evidence-Based Concept to the
Treatment of Alcohol, Drug, and Mental Health
Disorders. The 2005 Legislature directed the
Washington State Institute for Public Policy (Institute)
to examine the potential benefits Washington could
obtain if it adopted an evidence-based approach for
alcohol, drug, and mental illness treatment. This
report describes our “bottom-line” findings as well as
our research approach.
_____________________________________________
Suggested citation for this report:
Steve Aos, Jim Mayfield, Marna Miller, and Wei Yen. (2006).
Evidence-based treatment of alcohol, drug, and mental
health disorders: Potential benefits, costs, and fiscal impacts
for Washington State. Olympia: Washington State Institute
for Public Policy.

Summary
The Washington State Institute for Public Policy
was directed by the 2005 Washington Legislature
to estimate whether “evidence-based” treatment
for people with alcohol, drug, and mental health
disorders offers economic advantages. Do
benefits outweigh costs? And, if so, what is the
magnitude of the potential fiscal savings to
government, as well as the total net benefits to all
of Washington?
Methods
To answer these questions, we systematically
reviewed the “what works” literature regarding
treatments for people with alcohol, drug, and
mental health disorders. We then estimated the
monetary value of the benefits, including factors
such as improved performance in the job market,
reduced health care and other costs, and reduced
crime-related costs.
Findings
1. Evidence-based treatment works. We found
that the average evidence-based treatment
can achieve roughly a 15 to 22 percent
reduction in the incidence or severity of these
disorders—at least in the short term.
2. The economics look attractive. We found
that evidenced-based treatment of these
disorders can achieve about $3.77 in benefits
per dollar of treatment cost. This is equivalent
to a 56 percent rate of return on investment.
From a narrower taxpayer’s-only perspective,
the ratio is roughly $2.05 in benefits per dollar
of cost.
3. The potential is significant. We estimate
that a reasonably aggressive implementation
policy could generate $1.5 billion in net
benefits for people in Washington ($416
million are net taxpayer benefits). The risk of
losing money with an evidence-based
treatment policy is small.

Background: The Omnibus Treatment of
Mental and Substance Abuse Disorders Act
of 2005
This research assignment originated in a much larger
bill enacted during the 2005 legislative session: the
Omnibus Treatment of Mental and Substance Abuse
Disorders Act.
A major goal of the Act is to reform how publiclyfunded mental health and chemical dependency
programs are provided in Washington. In passing the
omnibus Act, the 2005 Legislature found that:
“Persons with mental disorders, chemical
dependency disorders, or co-occurring mental
and substance abuse disorders are
disproportionately more likely to be confined in a
correctional institution, become homeless,
become involved with child protective services or
involved in a dependency proceeding, or lose
those state and federal benefits to which they
may be entitled as a result of their disorders.” 1
Further, the Legislature found that:
“Prior state policy of addressing mental health and
chemical dependency in isolation from each other
has not been cost-effective and has often resulted
in longer-term, more costly treatment that may be
less effective over time.” 2
Among the several actions adopted in the 2005 Act to
address these general concerns, the Legislature
indicated its intention to:
“Improve treatment outcomes by shifting
treatment, where possible, to evidence-based,
research-based, and consensus-based treatment
practices and by removing barriers to the use of
those practices.” 3

The Basic Questions for the Study
Within the context of the Act’s overall goals, the
language directing the Institute’s study is shown in
the sidebar on this page.

In brief, the Legislature directed the Institute to
answer the following “bottom-line” questions:
9 Does evidence-based treatment for people with
alcohol, drug, or mental health disorders make
economic sense?
9 Do benefits outweigh costs?
9 And, if so, what is the potential magnitude of
the fiscal savings to government, and what are
the total net benefits to all of Washington?
In addition to directing the Institute to answer these
questions, the omnibus Act also required the
Institute to evaluate the effectiveness of the Act’s
pilot programs, which are designed to test several
new implementation approaches (see the sidebar
on page 6 for a brief description of the pilot
program study).

Legislative Study Language
Engrossed Second Substitute Senate Bill 5763,
Chapter 504, Laws of 2005, Sec. 605.
“The Washington state institute for public
policy shall study the net short-run and longrun fiscal savings to state and local
governments of implementing evidencebased treatment of chemical dependency
disorders, mental disorders, and co-occurring
mental and substance abuse disorders. The
institute shall use the results from its 2004
report entitled "Benefits and Costs of
Prevention and Early Intervention Programs
for Youth" and its work on effective adult
corrections programs to project total fiscal
impacts under alternative implementation
scenarios. In addition to fiscal outcomes, the
institute shall estimate the long-run effects
that an evidence-based strategy could have
on statewide education, crime, child abuse
and neglect, substance abuse, and economic
outcomes. The institute shall provide an
interim report to the appropriate committees
of the legislature by January 1, 2006, and a
final report by June 30, 2006.”
The Institute received an appropriation of
$80,000 to conduct the study.

1
2
3

E2SSB 5763, Chapter 504, Laws of 2005, Section 101.
Ibid.
Ibid., Section 101(3).

2

Research Methods
To answer the Legislature’s questions, we followed
the same two-step procedures we have applied to
other recent projects. First, we independently and
systematically assessed the research literature on
“what works,” and then we estimated benefits and
costs. In the Appendix to this report (beginning on
page 7), technical readers can find a detailed
description of our methods. Here, we summarize
our approach.
1. Assessing the research literature: Does
evidence-based treatment of alcohol, drug, and
mental illness reduce the incidence or severity of
these disorders?
We began by reviewing lists of evidence-based
treatments that have been compiled by other
researchers.4 After we reviewed all of the individual
studies associated with these listed treatments, we
then only included the results of “rigorous” evaluation
studies in our review. To be considered rigorous, an
evaluation must have included, at a minimum, a nontreatment comparison group that was well-matched to
the treatment group. We used this restriction
because greater confidence can be placed in causeand-effect conclusions from rigorous comparisongroup studies. Studies that use weaker research
methods do not provide this level of assurance and
were excluded. Thus, our judgment of what
constitutes “evidence” is more restrictive than the
standards used by some other researchers.
In recent years, researchers have developed a set of
statistical tools to facilitate systematic reviews of the
evidence. The set of procedures is called “metaanalysis” and we employed that methodology in this
study. Our meta-analytic review includes 206 studies
(246 trials) of evidence-based treatments for persons
with alcohol, drug, and mental health disorders.
Most of the individual evaluation studies we examined
were conducted outside of Washington State. A
primary purpose of our study is to take advantage of
all evaluations and, thereby, learn whether there are
options that can allow policymakers in Washington to
improve this state’s mental health and chemical
dependency treatment system.

2. Assessing the economics: What are the
benefits and costs of evidence-based treatment
of alcohol, drug, and mental illness?
After calculating the likely effect of an average
evidence-based treatment in reducing disorders, we
then estimated each option’s benefits and costs. To
do this, we used the same methods we have
employed in our earlier reviews of criminal justice and
other prevention programs.5 We estimated the
degree to which reductions in alcohol, substance
abuse, and mental illness disorders improve longevity
and an individual’s economic earnings, reduce health
care and other costs, and reduce crime and crimerelated costs.
As in our previous analyses, impacts were estimated
from two different perspectives: first, we calculated
benefits gained by program participants themselves;
second, we estimated benefits received by taxpayers
and other non-participants. An example of a
participant benefit is the increased economic
earnings stemming from enhanced labor productivity
when a treatment reduces disorder rates. An
example of a taxpayer benefit is the reduced level of
taxes needed to fund hospital emergency room visits
when the evidence-based treatment program
reduces problematic disorders. The perspectives of
both participants and taxpayers are necessary to
provide a full description of fiscal and non-fiscal
benefits and costs.
We then estimated total potential benefits based on
the number of people in Washington who could take
advantage of a particular evidence-based treatment.
We compiled information from a number of sources to
estimate how many people in Washington have a
serious alcohol, drug, or mental illness disorder, and
how many could realistically be expected to benefit
from an evidence-based treatment.
Finally, we varied the estimates and assumptions in
our analysis to gauge the overall level of uncertainty
in the “bottom-line” numbers we present.

5

4

See Appendix A.

See: (a) S. Aos, M. Miller, and E. Drake. (2006). Evidence-based adult
corrections programs. Olympia: Washington State Institute for Public
Policy; (b) S. Aos, R. Lieb, J. Mayfield, M. Miller, and A. Pennucci.
(2004). Benefits and costs of prevention and early intervention programs
for youth. Olympia: Washington State Institute for Public Policy; and
(c) S. Aos, P. Phipps, R. Barnoski, and R. Lieb (2001). The comparative
costs and benefits of programs to reduce crime. Olympia: Washington
State Institute for Public Policy.

3

Findings
How prevalent are alcohol, drug, and serious
mental health disorders?
To estimate the total benefits and costs of evidencebased treatment, we gathered national estimates of
the prevalence of clinically serious alcohol, drug, and
mental health disorders. We focused on serious
disorders because they appear to be the most costly
both to individuals with a disorder and to the rest of
society.6 We focused on adults (18 years and older)
to make the study compatible with current national
prevalence rates and because our previous work
emphasized younger people.7
In this study, we used the following prevalence rates:
9 Alcohol or Drug Disorders. About 7.6 percent of
the adult (18 to 54 years old) population has a
clinically significant alcohol or drug disorder.8
This is equivalent to about 1 in 13 adults. To
account for the comorbidity (two conditions at the
same time) between alcohol and drug disorders,
we also estimated the following:
•

61 percent of these people have an alcoholonly disorder

•

22 percent have a drug-only disorder

•

17 percent have alcohol and drug disorders

9 Serious Mental Illness. About 3.8 percent of the
adult population has a serious mental illness.9
This is equivalent to about 1 in 26 adults. These
serious mental illnesses were defined to include
schizophrenia and other non-affective psychosis,
manic depressive disorder, severe forms of major
depression, and panic disorder.

Does evidence-based treatment of alcohol, drug,
and mental illness reduce the incidence or
seriousness of these disorders?
We found that the average evidence-based
treatment reduces the short-term incidence or
seriousness of alcohol, drug, or mental health
disorders 15 to 22 percent.10
For example, if 75 percent of people with an alcohol
disorder continue to have the disorder without
treatment, then with an average evidence-based
alcohol treatment this percentage can be lowered to
64 percent—a 15 percent improvement in disorder
rates.
Our analysis revealed that in the short-term, the
average evidence-based treatment produces the
following statistically significant decreases in the
probability of these disorders:
9 Alcohol Disorders: a 15 percent reduction
9 Drug Disorders: a 22 percent reduction
9 Serious Mental Illness: an 22 percent reduction
It should be emphasized that these estimates are
based on studies with fairly short-term follow-up
periods—often a year or less. We found few
studies that evaluated effectiveness over the longer
term. To account for this lack of longitudinal
research, in our benefit-cost analyses we
significantly reduce (technically, we “decay”) these
short-term effectiveness rates, since many people
speculate that the beneficial effects of treatment
erode as time passes.11
What are the benefits and costs of evidencebased treatment of alcohol, drug, and mental
illness?

6

See: (a) H. Harwood. (2000). Updating estimates of the economic costs
of alcohol abuse in the United States: Estimates, update methods, and
data. Report prepared by The Lewin Group for the National Institute on
Alcohol Abuse and Alcoholism. Based on estimates, analyses, and data
reported in H. Harwood, D. Fountain, and G. Livermore. (1998). The
economic costs of alcohol and drug abuse in the United States, 1992.
Prepared for the National Institute on Drug Abuse and the National
Institute on Alcohol Abuse and Alcoholism, National Institutes of Health,
Dept. of Health and Human Services. NIH Publication No. 98-4327.
Rockville, MD: National Institutes of Health. http://pubs.niaaa.nih.gov/
publications/economic-2000/index.htm; (b) Office of National Drug Control
Policy. (2004). The economic costs of drug abuse in the United States,
1992-2002. Washington, DC: Executive Office of the President (Publication
No. 207303). http://www.whitehouse drugpolicy.gov/publications/
economic_costs/economic_costs.pdf; and (c) H. Harwood, A. Ameen, G.
Denmead, E. Englert, D. Fountain, and G. Livermore. (2000). The
economic costs of mental illness, 1992. Prepared for the National Institute
of Mental Health.
http://www.lewin.com/NR/rdonlyres/ea3i6g7cjgsvls2ukpupxo7wbjlmu25vh3
nd5rldz3lwyxfab6y6e4smh2zfpcs33wmmuq2cgbp3vg/2487.pdf
7
Aos et al., Benefits and costs of prevention and early intervention
programs for youth.
8
W.E. Narrow, D.S. Rae, L.N. Robins, and D.A. Regier. (2002). Revised
prevalence estimates of mental disorders in the United States: Using a
clinical significance criterion to reconcile 2 surveys' estimates. Archives of
General Psychiatry, 59: 115-123.
9
Harwood et al., The economic costs of mental illness, Table 4.7.

4

We found that the economics of the average
evidence-based treatment for people with serious
alcohol, drug, or mental disorders are quite attractive.
Per dollar of treatment cost, we estimate that
evidence-based treatment generates about $3.77 in
benefits for people in Washington. Expressed as a
return on investment, this is equivalent to roughly a
56 percent rate of return.
When we restrict this analysis to only those
benefits that accrue to taxpayers, the benefit-tocost ratio is $2.05.

10
11

See Appendix A for details behind these estimates.
Ibid.

Of the total benefits to Washington, approximately:
•

35 percent stem from the effect that the reduced
incidence of a disorder has on the person’s
economic earnings in the job market;

•

50 percent are linked to fewer health care and
other costs incurred;

•

7 percent are due to the lowered costs of crime; and

•

8 percent are for miscellaneous benefits.

We also estimated the total potential impact that an
evidence-based strategy could have for Washington
State. This involved first estimating the number of people
in Washington who have a serious disorder (described
above). We then subtracted an estimate of the number of
people in Washington already being treated with an
evidence-based program.12 We further restricted the size
of the potential treatment population by assuming that
only half of those who need treatment (and are not
currently being treated) would ultimately be served.
Under these assumptions, we found that the total net
benefits to Washington would be about $1.5 billion. From
the narrower taxpayer-only perspective, the net benefits
would be about $416 million.
How much uncertainty exists in these estimates
of benefits and costs?
In any estimation of the outcomes of complex human
behavior and human service delivery systems, there is
uncertainty. In our analysis, we estimated the degree to
which our bottom-line estimates could be influenced by
this uncertainty. As described in the Technical Appendix,
we performed an analysis called “Monte Carlo
simulation.” We randomly varied the key factors that
enter our calculations and then re-estimated the results of
our analysis. We did this re-estimation process 10,000
times, each time testing the range of uncertainty in our
findings.
We sought to determine the probability that our estimates
would produce a contrary finding. That is, we tested to
see how often our positive results would turn negative—
that money would be lost not gained.
From the perspective of all of Washington, we found that
the chance that an expansion of evidence-based
treatments would actually lose money (rather than
generate benefits) was less than 1 percent. From the
narrower taxpayer-only perspective, we found that the
chance that an evidence-based strategy would lose
money is approximately 1 percent. That is, about one
time out of a hundred an evidence-based strategy would
end up costing taxpayers more money than it saved.

Next Research Steps
To complete this research project on time and on
budget (the Institute received $80,000 for the study), we
had to adopt several strategies to narrow the study’s
scope. If the legislature decides to initiate a follow-up
study, the following limitations could be addressed:
1. Expand the scope of the study to include people
younger than 18. In this study, we reviewed
published research evaluations of alcohol, drug, and
mental health treatments. These research fields are
vast. In order to make the current study manageable,
we restricted our review to treatments for adults 18
years and older. We also made this restriction
because most of the existing research on the
prevalence and costs of alcohol, drug, and mental
health disorders has been for adult populations.
Additionally, we researched substance abuse
programs for youth in a study we completed in 2004
on prevention programs. A subsequent study could
expand the scope of the current research to identify
the economics of evidence-based treatment for
people 17 years and younger.
2. Expand the scope of the study to include
evidence-based treatment for less serious
alcohol, drug, and mental health disorders. We
restricted our search for evidence-based treatments
to those that focus on people with quite severe,
clinically significant, levels of disorder. We did this
because existing cost studies indicate that the
severe forms of disorder are usually the most costly
to society. A subsequent study could expand the
scope to identify evidence-based treatments for less
severe forms of these disorders. Because of
diminishing returns, however, the returns on
investment will probably not be as large as those
found in this study, but this hypothesis could be
tested in the subsequent study.
3. Identify specific types of evidence-based
treatment. The purpose of the present study was to
explore the total “market” potential of evidencebased treatment; a subsequent study could help
identify specific strategies. We analyzed the
economics of “prototype” evidence-based treatments
for alcohol, drug, or mental health disorders. That is,
we calculated the return on investment for an
average evidence-based treatment. A subsequent
study could focus on specific “name-brand” types of
treatment for alcohol, drug, or mental health
disorders and determine the economic returns
associated with each. This additional detailed
information could offer executive and legislative
public policymakers with “line-item” information on
specific evidence-based treatments.

12

For the purpose of this study, we assume that the vast majority of those
currently being treated are receiving evidence-based treatment.

5

4. Conduct further research regarding the link
between alcohol, drug, and mental health
disorders and child abuse and neglect. This
study contains only rough estimates of how alcohol,
drug, and mental health disorders causally influence
rates of child abuse and neglect. For example, we
included estimates of how substance abuse
disorders affect fetal alcohol syndrome, and we
estimated how all the disorders affect the ability of a
person to perform normal household activities. For
the effect of these disorders on other child welfare
outcomes, however, our current estimates are
probably incomplete and likely underestimate the
actual impact. To overcome this limitation, a
subsequent study could test this linkage further and
develop additional information that could be useful
for public policymakers.

Additional Institute Study From the Omnibus
Treatment of Mental and Substance Abuse
Disorders Act of 2005
Crisis Responder Pilot Evaluation
The same Act that directed the study described in
this report also instructed the Department of Social
and Health Services to establish two pilot sites
where specially trained crisis responders will
investigate and have the authority to detain
individuals considered “gravely disabled or
presenting a likelihood of serious harm” due to
mental illness, substance abuse, or both. The
integration of mental health and substance abuserelated crisis investigations and the establishment
of secure detoxification facilities at the pilot sites
are expected to improve the efficiency of
evaluation and treatment and result in better
outcomes for those involuntarily detained under
this new law. The pilots began operations in May
2006. The Legislature directed the Washington
State Institute for Public Policy to determine if the
pilots cost-effectively improve client mental
health/chemical dependency evaluation, treatment,
and outcomes. A preliminary report by the Institute
is due to the Legislature in December 2007. The
final report is to be completed by September 2008.
For more information on this related project,
contact Jim Mayfield at the Institute:
mayfield@wsipp.wa.gov; 360-586-2783.

6

Technical Appendices
Appendix A: Meta-Analytic Procedures
A1: Study Selection and Coding Criteria
A2: Procedures for Calculating Effect Sizes
A3: Institute Adjustments to Effect Sizes for Methodological Quality, Outcome Measure Relevance, and
Researcher Involvement
A4: Meta-Analytic Results—Estimated Effect Sizes and Citations to Studies Used in the Analyses

Appendix B: Methods and Parameters to Model the Benefits and Costs of Evidence-Based Treatment
B1:
B2:
B3:
B4:
B5:
B6:
B7:
B8:
B9:
B10:
B11:
B12:

General Model Parameters
Program Effectiveness Parameters
Program Design Parameters
Prevalence Parameters
Total Potential Population to Be Treated
Morbidity Parameters and Methods
Lost Household Production Methods
Health Care and Other Costs
Mortality Parameters and Methods
Crime Parameters
Marginal Treatment Effect
Sensitivity Analysis

Exhibits:
A.1:
A.2:
A.3:
A.4:
A.5:
B.1:
B.2:
B.3:
B.4:

Listed Programs, Practices, and Treatments With Studies Meeting Minimum Quality Standards
Meta-Analytic Results of the Effects of EBT on Disordered Alcohol Use
Meta-Analytic Results of the Effects of EBT on Disordered Drug Use
Meta-Analytic Results of the Effects of EBT on Mental Illness
Citations of Studies Used in the Meta-Analysis
The Benefits and Costs of Evidence-Based Treatment: General Model Parameters
The Benefits and Costs of Evidence-Based Treatment: Annual Data Series
The Benefits and Costs of Evidence-Based Treatment: Program-Specific Model Parameters
Meta-Analytic Estimates of Standardized Mean Difference Effect Sizes
B.4a: Citations to Studies in Exhibit B.4
B.5: The Benefits and Costs of Evidence-Based Treatment: Model Parameters Varied in the Monte Carlo Simulations

Appendix A: Meta-Analytic Procedures
To estimate the benefits and costs of evidence-based
treatment (EBT) of alcohol, drug, and mental illness disorders,
we conducted separate analyses of a number of key statistical
relationships. In Appendix A, we describe the procedures we
employed and the results we obtained in estimating the causal
linkage for the following nine relationships:
•

The effect of EBT on serious alcohol disorders

•

The effect of EBT on serious illicit drug disorders

•

The effect of EBT on serious mental illness disorders

•

The effect of serious alcohol disorders on job market
outcomes

•

The effect of serious illicit drug disorders on job market
outcomes

•

The effect of serious mental illness disorders on job
market outcomes

•

The effect of serious alcohol disorders on crime outcomes

•

The effect of serious illicit drug disorders on crime
outcomes

•

The effect of serious mental illness disorders on crime
outcomes

To estimate these nine key relationships, we conducted
reviews of the relevant research literature. In recent years,
researchers have developed a set of statistical tools to facilitate
systematic reviews of evaluation evidence. The set of
procedures is called “meta-analysis” and we employ that
methodology in this study.13 In Appendix A, we describe these
general procedures, the unique adjustments we made to them,
and the results of our meta-analyses.

A1. Study Selection and Coding Criteria
A meta-analysis is only as good as the selection and coding
criteria used to conduct the study.14 Following are the key
choices we made and implemented.
EBT Programs Examined. Due to the broad scope of this
project, we did not conduct a systematic review of all
evaluations of alcohol, drug, and mental illness disorder
treatments. We searched, instead, for studies associated
with treatments that are considered evidence-based
according to the following published sources: the United
States Substance Abuse and Mental Health Services
13

We follow the meta-analytic methods described in: M.W. Lipsey, and
D. Wilson. (2001). Practical meta-analysis. Thousand Oaks: Sage
Publications.
14
All studies used in the meta-analysis are identified in the references
beginning on page 17 of this report. Many other studies were reviewed,
but did not meet standards set for this analysis.

7

Administration (SAMHSA), the University of Washington
Alcohol and Drug Abuse Institute (ADAI), the Washington
Institute for Mental Illness Research and Training (WIMIRT),
and the Cochrane Collaboration. We did not include all
programs listed by these sources, such as prevention
programs for youth, the subject of a previous Washington
15
State Institute for Public Policy (Institute) analysis. We also
excluded gambling, tobacco cessation, and workplace
programs, and programs that exclusively target the elderly.
Exhibit A.1 lists the 57 treatments and practices identified by
the following sources, and for which we found studies that
met our minimum quality standards.
•

•

SAMHSA maintains a list of model, effective, and
16
promising prevention and treatment programs. For
inclusion, we selected programs treating adults with
alcohol, drug, or mental health disorders.
ADAI publishes a list of evidence-based practices for
the prevention and treatment of drug and alcohol
abuse, including several programs for the treatment of
individuals with co-occurring mental health and
substance abuse disorders. We included only the
ADAI-listed programs for adults with alcohol, drug
abuse, or co-occurring disorders.

•

WIMIRT has published several reports identifying
recommended approaches for treating or managing
mental illness in vulnerable populations: children, ethnic
and sexual minorities, the elderly, and those with co17
occurring disorders. We included any program listed
by WIMIRT that focused on the treatment of mentally ill
adults or those with co-occurring disorders.

•

The Cochrane Collaboration conducts and publishes
systematic reviews of the effects of healthcare
18
interventions. Included in this analysis are the results
of their reviews of evidence-based treatments for
serious mental illness. This was our primary source of
evidence for the effects of pharmacological treatments
for mental illness.

Study Selection. As we describe above, the process for
selecting studies of EBT for alcohol, drug, and mental illness
disorders was modified to limit the scope of the literature
review. We used four primary means to locate studies: (a) for
the meta-analysis of EBT programs, we reviewed citations
provided by the organization that recommended a particular
program; (b) we consulted the study lists of other systematic
and narrative reviews of the research literature;19 (c) we
examined the citations in the individual studies themselves;
and (d) we conducted independent literature searches of
research databases using search engines such as Google,
Proquest, Ebsco, ERIC, and SAGE. As we will describe, the
most important criteria for inclusion in our study was that an
evaluation have a control or comparison group. Therefore,
after first identifying all possible studies via these search
methods, we attempted to determine whether the study was
an outcome evaluation that had a comparison group. If a

15

Aos et al., Benefits and costs of prevention and early intervention
programs for youth.
16
http://modelprograms.samhsa.gov/template_cf.cfm?page=model_list
17
http://www.spokane.wsu.edu/research%26service/WIMIRT/content/
documents/Intro%20Book.pdf
18
http://www.cochrane.org/reviews/en/topics/index.html
19
Many studies used in our review of alcohol treatment programs were
identified in W.R. Miller, and P.L. Wilbourne. (2002). Mesa Grande: A
methodological analysis of clinical trials of treatments for alcohol use
disorders. Addiction, 97(2): 265-277. Other similar reviews are
identified with an asterisk in Exhibit A.2.

8

study met these criteria, we then secured a paper copy of the
study for our review.
Peer-Reviewed and Other Studies. We examined all
program evaluation studies we could locate with these search
procedures. Many of these studies were published in peerreviewed academic journals while many others were from
government reports obtained from the agencies themselves.
It is important to include non-peer reviewed studies, because
it has been suggested that peer-reviewed publications may
be biased to show positive program effects. Therefore, our
meta-analysis includes all available studies regardless of
published source.
Control and Comparison Group Studies. Our analysis
only includes studies that had a control or comparison group.
That is, we did not include studies with a single-group, prepost research design. This choice was made because it is
only through rigorous comparison group studies that average
treatment effects can be reliably estimated.
Exclusion of Studies of Program Completers Only. We
did not include a comparison study in our meta-analytic
review if the treatment group was made up solely of program
completers. We adopted this rule because there are too
many significant unobserved self-selection factors that
distinguish a program completer from a program dropout, and
that these unobserved factors are likely to significantly bias
estimated treatment effects. Some comparison group studies
of program completers, however, also contain information on
program dropouts in addition to a comparison group. In these
situations, we included the study if sufficient information was
provided to allow us to reconstruct an intent-to-treat group
that included both completers and non-completers, or if the
demonstrated rate of program non-completion was very small
(e.g. under 10 percent). In these cases, the study still
needed to meet the other inclusion requirements listed here.
Random Assignment and Quasi-Experiments. Random
assignment studies were preferred for inclusion in our review,
but we also included non-randomly assigned control groups.
We only included quasi-experimental studies if sufficient
information was provided to demonstrate comparability
between the treatment and comparison groups on important
pre-existing conditions such as age, gender, and pretreatment characteristics such as prior hospitalizations.
Enough Information to Calculate an Effect Size.
Following the statistical procedures in Lipsey and Wilson
(2001), a study had to provide the necessary information to
calculate an effect size. If the necessary information was
not provided, the study was not included in our review.
Mean-Difference Effect Sizes. For this study, we coded
mean-difference effect sizes following the procedures in
Lipsey and Wilson (2001). For dichotomous measures, we
used the arcsine transformation to approximate the mean
difference effect size, again following Lipsey and Wilson
(2001). We chose to use the mean-difference effect size
rather than the odds ratio effect size because we frequently
coded both dichotomous and continuous outcomes (odds
ratio effect sizes could also have been used with
appropriate transformations).

Exhibit A.1: Listed Programs, Practices, and Treatments With Studies Meeting Minimum Quality Standards
(These treatments are not necessarily recommended by the Institute)
Alcohol and Drug Abuse
12-Step Facilitation Therapy (A)
Behavioral Couples Therapy (A)
Behavioral Self-Control Training (A)
Brief Intervention (S)
Brief Marijuana Dependence Counseling (A)
Cognitive Behavioral Coping Skills Therapy (A)
Cognitive Behavioral Therapy for Alcohol Dependence (O)
Cognitive Behavioral Therapy for Substance Abuse (O)
Community Reinforcement Approach (W)
Contingency Management (A)
Focus on Families (S)
Holistic Harm Reduction (A)
Individual Cognitive Behavioral Therapy (A)
Individual Drug Counseling Approach to Treat Cocaine Addiction (A)
Lower-Cost Contingency Management (A)
Matrix Intensive Outpatient Program for Treatment of Stimulants (A)
Methadone/Opiate Substitution Treatment (A)
Motivational Enhancement Therapy (A)
Multidimensional Family Therapy (A)
Naltrexone (for Alcohol or Opiates) (A)
Relapse Prevention Therapy (A)

Mental Health
Assertive Community Treatment (S)
Behavioral Therapy for Anxiety (O)
Behavioral Treatment of Panic Disorder (W)
Brief Cognitive Behavioral Intervention for Amphetamine Users (A)
Brief Dynamic Psychotherapy for Depression (W)
Cognitive Behavior Therapy (W)
Cognitive Behavior Therapy for Generalized Anxiety Disorder (W)
Cognitive Therapy for Depression (W)
Crisis Intervention for People With Severe Mental Illnesses (C)
Electroconvulsive Therapy for Schizophrenia (C)
Family Intervention (W)
Interpersonal Psychotherapy (W)
Light Therapy for Depression (C
Motivational Interviewing (W)
Multi-Family Group Intervention (W)
Music Therapy for Schizophrenia (C)
Pharmacotherapy for Anxiety Disorder (C)
Pharmacotherapy for Bipolar Disorders (C)
Pharmacotherapy for Depression (C)
Pharmacotherapy for Post Traumatic Stress Disorder (C)
Pharmacotherapy for Schizophrenia (C)
Psychological Treatment of Post-Traumatic Stress Disorder (C)
Mental Health and Substance Abuse
PTSD Stress-Management Therapy (C)
Anger Management for Substance Abuse and Mental Health Clients (A) Supported Employment (S)
Behavioral Treatment for Substance Abuse in Schizophrenia (W)
Treatment of Post Traumatic Stress (S)
DBT for Substance Abusers with Borderline Personality Disorder (W)
Effects of Clozapine on Substance Use Among Schizophrenics (O)
Integrated Group Therapy for Bipolar and Substance Disorders (W)
Integrated Program for Comorbid Schizophrenia & Substance Use (O)
Integrated Treatment for Dual Disorders (W)
Listed by: A = Alcohol and Drug Abuse Institute
C = The Cochrane Collaboration
S = Substance Abuse and Mental Health Services Administration
W = Washington Institute for Mental Illness Research and Training
O = Other Literature Reviews
Note: While practices may be listed by multiple agencies, only one agency is shown.

Unit of Analysis. In most cases, our unit of analysis for this
study was an independent test of a treatment at a particular site.
Some studies reported outcomes for multiple sites; we included
each site as an independent observation if a unique and
independent comparison group was also used at each site. For
certain mental health treatments, we relied on meta-analytic
reviews published by the Cochrane Collaboration. In those
cases, we computed effect sizes from statistics published in the
reviews and the unit of analysis was the review.20
Multivariate Results Preferred. Some studies presented two
types of analyses: raw outcomes that were not adjusted for
covariates such as age, gender, or pre-treatment
characteristics; and those that had been adjusted with
multivariate statistical methods. In these situations, we coded
the multivariate outcomes.
Outcomes Measures of Interest. We only recorded
measures that reflected a change in symptoms, behaviors, or
other outcomes closely related to the treated disorder. In
mental health studies, this includes outcomes such as level of
functioning, symptoms, relapse, psychometric scores,
hospitalizations, and emergency room visits. Relevant
substance abuse outcomes include, for example, quantity
consumed, days of use, abstinence, blood or urine tests,
arrests, employment, and reports of problems due to
substance abuse. We did not record process and quality
20

We tested the validity of this approach by meta-analyzing the
results of 16 individual studies reported in three Cochrane reviews of
treatments for schizophrenia and compared the results to metaanalysis of the three reviews. The resulting standardized effect sizes
differed by only 0.01.

measures such as rates of treatment completion, number of
counseling sessions, client satisfaction, and quality of
services, etc.
Choosing Among Different Outcome Measures. A single
study may report a variety of outcomes. For example, one
study of mental illness treatment may report psychometric
scores and police contacts. A study of an alcohol abuse
treatment may report the quantity of alcohol consumed per
day and arrests. In such cases we recorded the outcome that
most directly reflected the effect of treatment on the primary
disorder: in the examples above, we would have recorded the
treatment effects of psychometric scores and the quantity of
alcohol consumed, respectively.
Averaging Effect Sizes for Similar Outcomes. Some
studies reported similar outcomes: e.g., a variety of
psychometric scores in the case of a mental health treatment,
or a number of different measures of substance use for an
alcohol or drug treatment. In such cases, we calculated an
effect size for each measure and then took a simple average.
As a result, each experimental trial coded in this study is
associated with a single effect size that reflects a general
reduction in the severity or incidence of a given disorder.
Dichotomous Measures Preferred Over Continuous
Measures. Some studies included two types of measures for
the same outcome: a dichotomous (yes/no) outcome and a
continuous (mean number) measure. In these situations, we
coded an effect size for the dichotomous measure. Our
rationale for this choice is that in small or relatively small
sample studies, continuous measures of treatment outcomes
can be unduly influenced by a small number of outliers, while
9

dichotomous measures can avoid this problem. Of course, if a
study only presented a continuous measure, we coded the
continuous measure.
Longest Follow-Up Periods. When a study presented
outcomes with varying follow-up periods, we generally coded
the effect size for the longest follow-up period. The longest
follow-up period allows us to gain the most insight into the
long-run benefits and costs of various treatments.
Occasionally, we did not use the longest follow-up period if it
was clear that a longer reported follow-up period adversely
affected the attrition rate of the treatment and comparison
group samples.
Some Special Coding Rules for Effect Sizes. Most studies in
our review had sufficient information to code exact meandifference effect sizes. Some studies, however, reported some,
but not all the information required. We followed the following
rules for these situations:
•

•

Two-tail p-values. Some studies only reported p-values
for significance testing of program outcomes. When we
had to rely on these results, if the study reported a onetail p-value, we converted it to a two-tail test.
Declaration of significance by category. Some studies
reported results of statistical significance tests in terms of
categories of p-values. Examples include: p<=.01,
p<=.05, or non-significant at the p=.05 level. We
calculated effect sizes for these categories by using the
highest p-value in the category. Thus, if a study reported
significance at p<=.05, we calculated the effect size at
p=.05. This is the most conservative strategy. If the study
simply stated a result was non-significant, we computed
the effect size assuming a p-value of .50 (i.e. p=.50).

A2. Procedures for Calculating Effect Sizes

Me − Mc

A(2): ES m =

SDe2 + SDc2
2

In this formula, ESm is the estimated effect size for the
difference between means from the research information; Me is
the mean number of an outcome for the experimental group;
Mc is the mean number of an outcome for the control group;
SDe is the standard deviation of the mean number for the
experimental group; and SDc is the standard deviation of the
mean number for the control group.
Often, research studies report the mean values needed to
compute ESm in (A2), but they fail to report the standard
deviations. Sometimes, however, the research will report
information about statistical tests or confidence intervals that can
then allow the pooled standard deviation to be estimated. These
procedures are also described in Lipsey and Wilson (2001).
Adjusting Effect Sizes for Small Sample Sizes
Since some studies have very small sample sizes, we follow
the recommendation of many meta-analysts and adjust for this.
Small sample sizes have been shown to upwardly bias effect
sizes, especially when samples are less than 20. Following
23
24
Hedges, Lipsey and Wilson report the “Hedges correction
factor,” which we use to adjust all mean difference effect sizes
(N is the total sample size of the combined treatment and
comparison groups):

[

3 ⎤
⎡
A(3): ES′m = ⎢1 −
⎥ × ES m , or , ES m ( p )
⎣ 4N − 9 ⎦

]

Effect sizes measure the degree to which a program has been
shown to change an outcome for program participants relative to
a comparison group. There are several methods used by metaanalysts to calculate effect sizes, as described in Lipsey and
Wilson (2001). In this analysis, we used statistical procedures
to calculate the mean difference effect sizes of programs. We
did not use the odds-ratio effect size because many of the
outcomes measured in this study are continuously measured.
Thus, the mean difference effect size was a natural choice.

Computing Weighted Average Effect Sizes, Confidence
Intervals, and Homogeneity Tests. Once effect sizes are
calculated for each program effect, the individual measures are
summed to produce a weighted average effect size for a
program area. We calculate the inverse variance weight for
each program effect and these weights are used to compute the
average. These calculations involve three steps. First, the
standard error, SEm of each mean effect size is computed with:25

Many of the outcomes we record, however, are measured as
dichotomies. For these yes/no outcomes, Lipsey and Wilson
(2001) show that the mean difference effect size calculation can
be approximated using the arcsine transformation of the
21
difference between proportions.

A(4): SEm =

A(1): ESm( p ) = 2 × arcsin Pe − 2 × arcsin Pc
In this formula, ESm(p) is the estimated effect size for the
difference between proportions from the research information; Pe
is the percentage of the population that had an outcome such as
re-arrest rates for the experimental or treatment group; and Pc is
the percentage of the population that was re-arrested for the
control or comparison group.
A second effect size calculation involves continuous data
where the differences are in the means of an outcome. When
an evaluation reports this type of information, we use the
22
standard mean difference effect size statistic.
21

Aos et al., Benefits and costs of prevention and early intervention
programs for youth, Table B10, equation 22.
22
Ibid., Table B10, equation 1.

10

′ )2
ne + nc
( ES m
+
ne nc
2(ne + nc )

In equation (A4), ne and nc are the number of participants in
the experimental and control groups and ES'm is from equation
(A3).
Next, the inverse variance weight wm is computed for each
26
mean effect size with:
A(5): wm =

23

1
SEm2

L.V. Hedges. (1981) Distribution theory for Glass’s estimator of effect
size and related estimators. Journal of Educational Statistics, 6: 107-128.
24
Lipsey and Wilson, Practical meta-analysis, 49, equation 3.22.
25
Ibid., 49, equation 3.23.
26
Ibid., 49, equation 3.24.

The weighted mean effect size for a group of studies in program
area i is then computed with:27
A(6): ES =

∑ (w ES ′
∑w
mi

mi

)

mi

Confidence intervals around this mean are then computed by
28
first calculating the standard error of the mean with:
A(7): SE =
ES

1
∑ wmi

Next, the lower, ESL, and upper limits, ESU, of the confidence
interval are computed with:29
A(8): ES L = ES − z(1−α ) ( SE ES )
A(9): ESU = ES + z(1−α ) ( SE ES )
In equations (A8) and (A9), z(1-α) is the critical value for the zdistribution (1.96 for α = .05).
The test for homogeneity, which provides a measure of the
dispersion of the effect sizes around their mean, is given by:30
A(10): Qi = (∑ wi ESi2 ) −

(∑ wi ESi ) 2

∑ wi

The Q-test is distributed as a chi-square with k-1 degrees of
freedom (where k is the number of effect sizes).

In Appendix A3, we describe our rationale for making these
downward adjustments. In particular, we make three types of
adjustments that are necessary to better estimate the results that
we are more likely to achieve in real-world settings. We make
adjustments for: (a) the methodological quality of each study we
include in the meta-analyses; (b) the relevance or quality of the
outcome measure that individual studies used; and (c) the
degree to which the researcher(s) who conducted a study were
invested in the program’s design.
A3.a. Methodological Quality. Not all research is of equal
quality, and this greatly influences the confidence that can be
placed in the results of a study. Some studies are well designed
and implemented, and the results can be viewed as accurate
representations of whether the program itself worked. Other
studies are not designed as well, and less confidence can be
placed in any reported differences. In particular, studies of
inferior research design cannot completely control for sample
selection bias or other unobserved threats to the validity of
reported research results. This does not mean that results from
these studies are of no value, but it does mean that less
confidence can be placed in any cause-and-effect conclusions
drawn from the results.
To account for the differences in the quality of research designs,
we use a 5-point scale as a way to adjust the reported results.
The scale is based closely on the 5-point scale developed by
32
researchers at the University of Maryland. On this 5-point
scale, a rating of “5” reflects an evaluation in which the most
confidence can be placed. As the evaluation ranking gets lower,
less confidence can be placed in any reported differences (or
lack of differences) between the program and comparison or
control groups.
On the 5-point scale as interpreted by the Institute, each study is
rated with the following numerical ratings.

Computing Random Effects Weighted Average Effect Sizes
and Confidence Intervals. When the p-value on the Q-test
indicates significance at values of p less than or equal to .05, a
random effects model is performed to calculate the weighted
average effect size. This is accomplished by first calculating the
random effects variance component, v.31
A(11): v =

Qi − (k − 1)
∑ wi − (∑ wsqi ∑ wi )

•

A “5” is assigned to an evaluation with well-implemented
random assignment of subjects to a treatment group and
a control group that does not receive the
treatment/program. A good random assignment study
should also indicate how well the random assignment
actually occurred by reporting values for pre-existing
characteristics for the treatment and control groups.

•

A “4” is assigned to a study that employs a rigorous
quasi-experimental research design with a program and
matched comparison group, controlling with statistical
methods for self-selection bias that might otherwise
influence outcomes. These quasi-experimental
methods may include estimates made with a convincing
instrumental variables modeling approach, or a
Heckman approach to modeling self-selection.33 A level
4 study may also be used to “downgrade” an
experimental random assignment design that had
problems in implementation, perhaps with significant
attrition rates.

•

A “3” indicates a non-experimental evaluation where the
program and comparison groups were reasonably well
matched on pre-existing differences in key variables.
There must be evidence presented in the evaluation that

This random variance factor is then added to the variance of
each effect size and then all inverse variance weights are
recomputed, as are the other meta-analytic test statistics.

A3. Institute Adjustments to Effect Sizes for
Methodological Quality, Outcome Measure Relevance,
and Researcher Involvement
In Exhibits A.2 – A.4 we show the results of our meta-analyses
calculated with the standard meta-analytic formulas described in
Appendix A2. In the last columns in each exhibit, however, we
list “Adjusted Effect Sizes” that we actually use in our benefit-cost
analysis of each program area: alcohol, drug, and mental illness
treatment. These adjusted effect sizes, which are derived from
the unadjusted results, are always smaller than or equal to the
unadjusted effect sizes we report in the same exhibit.
32

27

Ibid., 114.
Ibid.
29
Ibid.
30
Ibid., 116.
31
Ibid., 134.
28

L. Sherman, D. Gottfredson, D. MacKenzie, J. Eck, P. Reuter, and S.
Bushway. (1998). Preventing crime: What works, what doesn't, what's
promising. Prepared for the National Institute of Justice. Department of
Criminology and Criminal Justice, University of Maryland. Chapter 2.
33
For a discussion of these methods, see W. Rhodes, B. Pelissier, G.
Gaes, W. Saylor, S. Camp, and S. Wallace. (2001). Alternative solutions to
the problem of selection bias in an analysis of federal residential drug
treatment programs. Evaluation Review, 25(3): 331-369.

11

indicates few, if any, significant differences were observed
in these salient pre-existing variables. Alternatively, if an
evaluation employs sound multivariate statistical
techniques (e.g., logistic regression) to control for preexisting differences, and if the analysis is successfully
completed, then a study with some differences in preexisting variables can qualify as a level 3.
•

A “2” involves a study with a program and matched
comparison group where the two groups lack
comparability on pre-existing variables and no attempt
was made to control for these differences in the study.

•

A “1” involves a study where no comparison group is
utilized. Instead, the relationship between a program and
an outcome, i.e., drug use, is analyzed before and after
the program.

We do not use the results from program evaluations rated as a
“1” on this scale, because they do not include a comparison
group and, thus, no context to judge program effectiveness.
We also regard evaluations with a rating of “2” as highly
problematic and, as a result, do not consider their findings in
the calculations of effect. In this study, we only considered
evaluations that rated at least a 3 on this 5-point scale.
An explicit adjustment factor is assigned to the results of
individual effect sizes based on the Institute’s judgment
concerning research design quality. This adjustment is critical
and the only practical way to combine the results of a high
quality study (e.g., a level 5 study) with those of lesser design
quality (level 4 and level 3 studies). The specific adjustments
made for these studies are based on our knowledge of
research in other topic areas. For example, in criminal justice
program evaluations, there is strong evidence that random
assignment studies (i.e., level 5 studies) have, on average,
smaller absolute effect sizes than weaker-designed studies.34
Thus, we use the following “default” adjustments to account for
studies of different research design quality:
•

A level 5 study carries a factor of 1.0 (that is, there is no
discounting of the study’s evaluation outcomes).

•

A level 4 study carries a factor of .75 (effect sizes
discounted by 25 percent).

•

A level 3 study carries a factor of .50 (effect sizes
discounted by 50 percent).

•

We do not include level 2 and level 1 studies in our
analyses.

These factors are subjective to a degree; they are based on
the Institute’s general impressions of the confidence that can
be placed in the predictive power of evaluations of different
quality.

The effect of the adjustment is to multiply the effect size for any
study, ES'm, in equation (A3) by the appropriate research
design factor. For example, if a study has an effect size of -.20
and it is deemed a level 4 study, then the -.20 effect size would
be multiplied by .75 to produce a -.15 adjusted effect size for
use in the benefit-cost analysis.
A3.b. Adjusting Effect Sizes of Studies With Short-Term
Follow-Up Periods. To account for the likelihood that the
effects of treatment do not persist indefinitely for all subjects,
we discount effect sizes, ESm, over time. The majority of
studies coded report only short-term outcomes. Few of the
studies provided outcomes beyond one year post-treatment
and many reported outcomes only during or at the end of a
treatment episode. Therefore, the unadjusted meta-analytic
effect sizes reflect relatively short-term outcomes. To reflect
the likelihood that the effects of a given treatment will decline
over time, we built in a “decay” factor. In Appendix B, we
discuss the methods by which we decay these effects.
A3.c. Adjusting Effect Sizes for Research Involvement in
the Program’s Design and Implementation. The purpose of
the Institute’s work is to identify and evaluate programs that
can make cost-beneficial improvements to Washington’s actual
service delivery system. There is some evidence that
programs closely controlled by researchers or program
developers have better results than those that operate in “real
35
world” administrative structures. In our evaluation of a realworld implementation of a research-based juvenile justice
program in Washington, we found that the actual results were
considerably lower than the results obtained when the
intervention was conducted by the originators of the program.36
Therefore, we make an adjustment to effect sizes, ESm, to
reflect this distinction. As a parameter for all studies deemed
not to be “real world” trials, the Institute discounts ES'm by .5,
although this can be modified on a study-by-study basis.

A4. Meta-Analytic Results—Estimated Effect Sizes
and Citations to Studies Used in the Analyses
Exhibits A. 2, A.3, and A.4 provide technical meta-analytic
results for the effect sizes computed for this analysis. Each
table provides the unadjusted and adjusted effect sizes for
EBT in each of the three program areas, and lists all of the
studies included in each analysis. Exhibit A.5 lists the
citations for all studies used in the meta-analyses.
The meta-analytic results of the effects of EBT on disordered
alcohol use are displayed in Exhibit A.2. The results for
disordered drug use and mental illness are displayed in
Exhibits A.3 and A.4, respectively.

35

34

M.W. Lipsey. (2003). Those confounded moderators in meta-analysis:
Good, bad, and ugly. The Annals of the American Academy of Political and
Social Science, 587(1): 69-81. Lipsey found that, for juvenile delinquency
evaluations, random assignment studies produced effect sizes only 56
percent as large as nonrandom assignment studies.

12

Ibid. Lipsey found that, for juvenile delinquency evaluations, programs in
routine practice (i.e., “real world” programs) produced effect sizes only 61
percent as large as research/demonstration projects. See also:
A. Petrosino, and H. Soydan. (2005). The impact of program developers as
evaluators on criminal recidivism: Results from meta-analyses of
experimental and quasi-experimental research. Journal of Experimental
Criminology, 1(4): 435-450.
36
R. Barnoski. (2004). Outcome evaluation of Washington State's
research-based programs for juvenile offenders. Olympia: Washington
State Institute for Public Policy, available at
<http://www.wsipp.wa.gov/rptfiles/04-01-1201.pdf>.

Exhibit A.2: Meta-Analytic Results of the Effects of EBT on Disordered Alcohol Use
Adjusted
Effect Size
Used in the
Fixed Effects Model
Random Effects Model
BenefitCost
Homogeneity Weighted Mean Effect Size &
Analysis
Weighted Mean Effect Size & p-value
Test
p-value

Alcohol Treatment Effects

Number of trials used in analysis:

Results Before Adjusting Effect Sizes

100

Number of subjects in treatment group: 7,973

Studies Used in the Meta-Analysis

Name of Study
Aalto, et al. (2000)

Essm
-0.097

N Tx

p-value

p-value

ES

p-value

ES

-0.253

0.000

0.000

-0.312

0.000

-0.247

Not
real
Design world
Score
=1

N Cn

39

ES

39

5

0

ESAdj Name of Study

Essm

N Tx

Design Not real
Score world =1

N Cn

ESAdj

-0.097

Lhuintre, et al. (1990)

-0.052

181

175

5

0

-0.052
-0.620

Aalto, et al. (2000)

-0.351

37

39

5

0

-0.351

Maheswaran, et al. (1992)

-0.620

21

20

5

0

Adams (1990)

-0.555

29

16

3

0

-0.277

Mallams, et al. (1982)

-0.666

19

16

5

0

-0.666

Allsop, et al. (1997)

-0.247

15

14

5

0

-0.247

Manwell et al. (2000)

-0.231

103

102

5

1

-0.115

Anderson, et al. (1992)

-0.300

80

74

5

0

-0.300

Marlatt, et al. (1998)

-0.251

174

174

5

0

-0.251

Anton, et al. (1999)

-0.363

68

63

5

0

-0.363

Mason , et al. (1994)

-0.780

7

6

5

0

-0.780

Anton, et al. (2006)

-0.081

917

309

5

0

-0.081

Mason , et al. (1999)

-0.290

70

35

5

0

-0.290

Anton, et al. (2006)

-0.184

157

153

5

0

-0.184

McCrady, et al. (1999)

0.081

24

22

3

1

0.020

Anton, et al. (2006)

-0.092

619

607

5

0

-0.092

McCrady, et al. (1999)

-0.202

24

21

5

0

-0.202

Azrin (1976)

-1.460

9

9

5

1

-0.730

Miller, et al. (1981)

-0.350

19

16

5

0

-0.350

Babor, et al. (1992)

-0.372

350

361

5

0

-0.372

Miller, et al. (1980)

-0.270

19

16

4

1

-0.101

Babor, et al. (1993)

-0.312

350

409

5

0

-0.312

Miller, et al. (1993)

-0.618

14

14

5

1

-0.309

Bien, et al. (1993)

-0.264

18

16

5

0

-0.264

Miller, et al. (2001)

0.178

28

30

5

0

0.178

Bosari, et al. (2000)

-0.615

29

30

5

1

-0.308

Miller, et al. (2001)

-0.040

32

33

5

0

-0.040

Bowers, et al. (1990)

-0.603

15

13

5

0

-0.603

Miller, et al. (2001)

0.158

29

35

5

0

0.158

Brown (1993)

-0.399

14

14

5

0

-0.399

Miller, Taylor, & West (1980)

-0.201

10

10

4

0

-0.151
-0.151

Chaney, O'Leary, Marlatt (1978)

-0.273

14

25

4

1

-0.102

Miller, Taylor, & West (1980)

-0.201

10

10

4

0

Chick (1985)

-0.496

69

64

5

0

-0.496

Monti, et al. (1990)

0.000

23

23

5

0

0.000

Chick, et al. (1988)

-0.189

54

41

5

0

-0.189

Monti, et al. (1993)

-0.538

7

11

5

0

-0.538

Collins, et al., (2002)

0.418

23.97

23.52

5

0

0.418

Murphy, et al. (2001)

-0.183

30

24

5

0

-0.183

Collins, et al., (2002)

-0.533

22.56

24.48

5

0

-0.533

Neighbors, et al. (2004)

-0.326

126

126

5

0

-0.326

Donovan, et al. (1988)

-0.155

20

19

5

0

-0.155

Nelson & Howell (1982-83)

-0.538

16

9

3

0

-0.269

Drake, et al. (1997)

-0.653

69

28

3

0

-0.326

Nilssen (1991)

-0.626

212

108

5

0

-0.626

Drake, et al. (1998) a

-0.033

75

68

5

0

-0.033

Obolensky (1984)

-0.842

9

13

3

0

-0.626

Drake, et al. (1998) b

-0.158

83

73

5

0

-0.158

O'Connell (1987)

-0.074

12

11

3

0

-0.037

Drake, et al. (2000)

-0.944

19

86

3

0

-0.472

Oei & Jackson (1980)

-0.704

16

16

3

0

-0.352

Elvy, et al. (1988)

-0.169

48

72

5

0

-0.169

Oei & Jackson (1982)

-0.867

16

8

3

0

-0.434

Eriksen, Bjornstad, & Gotestam (1986)

-1.139

11

12

3

1

-0.285

Oei & Jackson (1982)

-0.867

16

8

3

0

-0.434

Fals-Stewart, et al. (1996)

-0.174

40

40

5

1

-0.087

O'Farrell, et al. (1993)

-0.578

30

29

5

0

-0.578

Feeney, et al. (2002)

-0.557

50

50

3

0

-0.279

O'Malley, et al. (1992)

-0.819

22

27

5

1

-0.410

Ferrell & Galassi (1981)

-0.951

8

9

5

1

-0.475

Ouimette, et al. (1997)

-0.076

897

1148

4

0

-0.057

Fichter, et al. (1993)

-0.061

45

45

5

0

-0.061

Paille, et al. (1995)

-0.172

173

177

5

0

-0.172

Fleming, et al. (2000)

-0.406

392

382

5

0

-0.406

Persson, et al. (1989)

-0.526

31

23

5

0

-0.526

Graeber, et al. (2003)

-1.332

15

15

4

0

-0.999

Reynolds, et al. (1995)

-0.449

42

36

5

0

-0.449

Handmaker, et al. (1999)

-0.221

18

16

5

1

-0.111

Richmond, et al. (1995)

-0.145

70

61

3

0

-0.073

Harris et al. (1990)

-0.519

9

17

5

1

-0.259

Rohsenhow, Smith, & Johnson (1985)

-0.232

14

20

4

0

-0.174

Heather et al. (1987)

-0.028

34

38

5

1

-0.014

Romelsjo, et al. (1989)

-0.147

41

42

5

0

-0.147

Heather, et al. (1996)

-0.372

47

33

5

0

-0.372

Sanchez-Craig, et al. (1991)

-0.101

29

67

5

0

-0.101

Hedberg, et al. (1974)

-0.683

15

15

5

0

-0.683

Sanchez-Craig, et al. (1996)

-0.006

74

81

5

1

-0.003

Hellerstein, et al. (1995)

-0.776

23

24

5

0

-0.776

Sannibale (1989)

-0.024

31

41

4

1

-0.009

Hester & Delaney (1997)

-0.633

20

20

5

1

-0.317

Sass, et al. (1996)

-0.498

136

136

5

0

-0.498

Hulse, et al. (2002)

-0.719

47

36

4

0

-0.540

Scott (1989)

-0.070

33

39

5

0

-0.070

Hunt & Azrin (1973)

-1.572

8

8

3

1

-0.393

Sisson & Azrin (1986)

-2.479

7

5

5

1

-1.240

James, et al. (2004)

-0.260

29

29

5

0

-0.260

Smith et al. (1998)

-0.470

49

32

4

0

-0.352

Jones, Kanfer, & Lanyon (1982)

-0.884

24

21

4

0

-0.663

Smith, et al. (1999)

-0.275

91

76

3

0

-0.138

Kelly, et al. (2000)

-0.900

11

9

5

1

-0.450

Tomson, et al. (1998)

-0.158

45

30

5

0

-0.158

Kivlahan et al. (1990)

-0.870

15

15

5

1

-0.435

Volpicelli, et al. (1992)

-0.643

35

35

5

0

-0.643

Kuchipudi, et al. (1990)

-0.067

59

55

5

0

-0.067

Wallace, et al. (1988)

-0.424

247

337

5

0

-0.424

Larimer, et al. (2001)

-0.394

60

60

3

0

-0.197

Whitworth, et al. (1996)

-0.257

74

74

4

0

-0.192

Lhuintre, et al. (1985)

-0.56642

33

37

5

0

-0.566

Winters, et al (2002)

-0.435

33

35

5

0

-0.435

13

Exhibit A.3: Meta-Analytic Results of the Effects of EBT on Disordered Drug Use
Adjusted
Effect Size
Used in the
Fixed Effects Model
Random Effects Model
BenefitCost
Homogeneity Weighted Mean Effect Size &
Analysis
Weighted Mean Effect Size & p-value
Test
p-value

Treatment for
Disordered Drug Use

Number of trials used in analysis:

Results Before Adjusting Effect Sizes

44

Number of subjects in treatment group: 3,506

ES

p-value

p-value

ES

p-value

ES

-0.360

0.000

0.000

-0.451

0.000

-0.355

Studies Used in the Meta-Analysis
Name of Study
Avants, et al. (2004)

Essm
-0.232

N Tx
108

N Cn
112

Score
5

real

ESAdj Name of Study
0

-0.232

Johnson, et al. (1992)

Essm
-0.491

N Tx
90

N Cn

Score world =1

60

5

0

ESAdj
-0.491

Azrin, et al. (1996)

-0.651

37

37

3

1

-0.163

Johnson, et al. (1995)

-0.641

90

60

5

0

-0.641

Azrin, et al.(1994)

-0.714

15

11

4

1

-0.268

Johnson, et al. (2000)

-0.487

55

55

5

0

-0.487

Baker, et al. (2001)

-0.688

32

32

3

1

-0.172

Kavanagh, et al. (2004)

-0.725

13

8

5

0

-0.725

Baker, et al. (2005)

-0.472

74

74

4

1

-0.177

Ling, et al. (1998)

-0.443

90

60

5

0

-0.443

Baker, et al. (2005)

-0.494

66

74

4

1

-0.185

Margolin, et al. (2003)

-0.383

45

45

5

0

-0.383

Bellack, et al. (2006)

-0.680

61

49

5

1

-0.340

Marijuana Treatment Project (2004)

-0.216

127

137

5

0

-0.216

Carroll, et al. (1991)

-0.206

21

21

4

1

-0.077

Marijuana Treatment Project (2004)

-0.610

132

137

5

0

-0.610

Carroll, et al. (1994)

-0.461

52

45

4

1

-0.173

Milby, et al. (1996)

-0.044

69

62

3

0

-0.022

Catalano, et al. (2002)

-0.048

63

63

5

1

-0.024

Newman, et al. (1979)

-0.827

50

50

5

0

-0.827

Cornish, et al. (1997)

-0.577

34

17

5

0

-0.577

Petry & Martin (2002)

-1.498

19

23

5

0

-1.498

Ctrits-Christoph, et al. (1999)

-0.237

121

123

4

0

-0.178

Petry, et al. ( 2000)

-0.772

19

23

5

0

-0.772

Dole, et al. (1969)

-2.051

12

16

5

1

-1.026

Piotrowski, et al. (1999)

0.000

51

51

5

0

0.000

Drake, et al. (1997)

-0.113

78

29

3

0

-0.056

Rawson, et al. (1995)

-0.122

41

44

5

0

-0.122

Drake, et al. (1998)a

-0.178

45

40

5

0

-0.178

Schottenfeld, et al. (1997)

-0.291

30

29

5

0

-0.291

Drake, et al. (1998)b

-0.124

45

40

5

0

-0.124

Silverman, et al., (1996)

-0.534

15

15

4

0

-0.401

Drake, et al. (2000)

-0.687

11

54

3

0

-0.344

Silverman, et al., (1998)

-1.554

36

15

4

0

-1.165

Fudala, et al. (2003)

-0.421

214

109

5

0

-0.421

Stephens, et al. (2000)

-0.598

117

86

5

0

-0.598

Gronbladh, et al. (1989)

-0.896

17

17

4

0

-0.672

Stephens, et al. (2000)

-0.497

88

86

5

0

-0.497

Higgins, et al. (2000)

-0.360

36

34

5

1

-0.180

Strain, et al. (1993)

-0.329

84

81

5

0

-0.329

Humphreys, et al., (1999)

-0.189

897

1148

4

0

-0.142

Vanichseni, et al. (1991)

-0.511

120

120

5

0

-0.511

James, et al. (2004)

-0.868

29

29

5

0

-0.868

Woody, et al. (1995)

-0.463

57

27

3

1

-0.116

14

A.4: Meta-Analytic Results of the Effects of EBT on Mental Illness
Our benefit-cost analysis focused on serious mental illness: non-affective psychosis (including schizophrenia), bipolar disorder and
severe forms of panic disorder, and depression. Because studies rarely indicated the severity of subjects’ mental disorders in the
studies, our analysis included all programs for depression, and we estimated effects for panic disorder based on studies for treatments of
anxiety disorders. To derive a single effect size for mental illness treatments, we first calculated effect sizes for four categories of mental
illness: non-affective psychosis, bipolar, anxiety, and major depressive disorders. After weighting according to prevalence among the
populations with serious mental illness, we combined the separate effect sizes into a single average (see the following table).

Disorder

Schizophrenia (Non-affective psychosis)
Bipolar disorder
Anxiety disorders
Major Depressive Disorder
All Mental Illness

Weight
0.079
0.410
0.191
0.321
1.000

Adjusted ES for BenefitCost Analysis
ES
Std Err
-0.323
0.029
-0.382
0.048
-0.404
0.045
-0.280
0.061
-0.360

0.047

Note: Relative prevalence was based on incidence of serious major depression, serious panic disorder, and bipolar I and II reported from the National
37
38
Comorbidity Survey Replication and non-affective psychosis as reported in the National Comorbidity Survey.

Adjusted
Effect Size
Used in the
Fixed Effects Model
Random Effects Model
BenefitCost
Homogeneity Weighted Mean Effect Size &
Analysis
Weighted Mean Effect Size & p-value
Test
p-value
Results Before Adjusting Effect Sizes

Treatments for
Bipolar Disorder

Number of trials used in analysis:
Number of subjects in treatment group:

6

ES

p-value

p-value

ES

p-value

ES

933

-0.386

0.000

-0.549

na

na

-0.382

Studies Used in the Meta-Analysis

Name of Study
Burgess, et al.(2001)

Essm

N Tx

N Cn

Not
real
Design world
Score
=1

-0.300

413

412

Macritchie, et al. (2003)

-0.512

155

161

5

0

-0.512

Rendell, et al. (2003)

-0.426

70

66

5

0

-0.426

16

Studies Used in the Meta-Analysis

Browne, et al. (2002)

Essm
0.045

N Tx
212

-0.300

Rendell, et al. (2003)

N Cn

Design Not real
Score world =1
5

0

ESAdj

-0.524

54

56

-0.524

Rendell, et al. (2003)

-0.454

220

114

5

0

-0.454

Weiss, et al. (2000)

-0.200

21

24

3

1

-0.050

Adjusted
Effect Size
Used in the
Fixed Effects Model
Random Effects Model
BenefitCost
Homogeneity Weighted Mean Effect Size &
Analysis
Weighted Mean Effect Size & p-value
Test
p-value

Number of subjects in treatment group: 1,479

Name of Study

0

N Tx

Essm

Results Before Adjusting Effect Sizes

Treatments for Depression

Number of trials used in analysis:

5

ESAdj Name of Study

N Cn

ES

p-value

p-value

ES

p-value

ES

-0.314

0.000

0.018

-0.323

0.000

-0.315

Not
real
Design world
=1
Score

ESAdj Name of Study

Essm

N Tx

N Cn

Design Not real
Score world =1

ESAdj

196

4

0

0.033

Shea, et al (1992)

-0.194

59

62

5

0

-0.194
-0.008

Fava, et al. (1998)

-0.513

20

20

5

0

-0.513

Shea, et al (1992)

-0.008

61

62

5

0

Lima, et al. (2006)

-0.434

206

179

5

0

-0.434

Simons, et al. (1986)

-0.610

36

16

3

0

-0.305

Lima, et al. (2006)

-0.528

143

155

5

0

-0.528

Tuunainen, et al. (2004)

-0.060

39

32

5

0

-0.060

Lima, et al. (2006)

-0.366

295

305

5

0

-0.366

Ward, et al., (2000)

-0.185

63

67

5

0

-0.185

Moncrieff, et al. (2004)

-0.325

395

355

5

0

-0.325

Wijkstra, et al. (2005)

-0.430

48

101

5

0

-0.430

Reynolds, et al. (2006)

-0.493

25

28

5

0

-0.493

Wijkstra, et al. (2005)

-0.368

100

101

5

0

-0.368

Reynolds, et al. (2006)

0.623

25

29

5

0

-0.623

Wijkstra, et al. (2005)

-0.786

22

17

5

0

-0.786

37
R.C. Kessler, W.T. Chiu, O. Demler et al. (2005), Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the National Comorbidity Survey
Replication. Archives of General Psychiatry, 62(6):617-627.
38
R.C. Kessler, K.A. McGonagle, S. Zhao et al. (1994). Lifetime and 12-month prevalence of DSM-III-R Psychiatric Disorders in the United States. Archives of
General Psychiatry, 51: 8-19.

15

Adjusted
Effect Size
Used in the
Fixed Effects Model
Random Effects Model
BenefitCost
Homogeneity Weighted Mean Effect Size &
Analysis
Weighted Mean Effect Size & p-value
Test
p-value
Results Before Adjusting Effect Sizes

Treatments for
Anxiety Disorders

Number of trials used in analysis:

31

Number of subjects in treatment group: 4,641

Studies Used in the Meta-Analysis

Name of Study

Essm

N Tx

ES

p-value

p-value

ES

p-value

ES

-0.256

0.000

0.000

-0.563

0.000

-0.404

Not
real
Design world
Score
=1

N Cn

ESAdj Name of Study

N Tx

Essm

N Cn

Design Not real
Score world =1

ESAdj

Barlow, et al. (1989)

-1.394

10

15

5

0

-1.394

Cordioli, et al. (2003)

-1.201

23

24

5

0

-1.201

Barlow, et al. (1989)

-0.858

15

15

5

0

-0.858

Dugas, et al. (2003)

-1.364

25

37

5

1

-0.682

Barlow, et al. (1989)

-0.938

16

15

5

0

-0.938

Durham, et al. (1994)

-0.710

35

29

4

1

-0.266

Barlow, et al. (1984)

-1.205

10

10

5

0

-1.205

Kapczinski (2003)

-0.378

277

280

5

0

-0.378

Barlow, et al. (2000)

-0.553

60

22

5

0

-0.553

Ladouceur, et al. (2000)

-1.571

14

12

5

1

-0.785

Barlow, et al., (1992)

-1.588

24

10

3

1

-0.397

Lindsay, et al. (1987)

-1.200

10

10

5

1

-0.600

Beck, et al. (1992)

-0.507

17

16

3

1

-0.127

Linehan, et al. (1999)

-0.289

12

16

5

1

-0.144

Bisson & Andrew (2005)

-0.375

79

77

4

0

-0.281

Marks, et al., (1993)

-0.909

23

17

3

0

-0.454

Bisson & Andrew (2005)

-0.426

266

187

5

0

-0.426

Mortberg, et al. (2005)

-1.005

12

12

5

1

-0.502

Bisson & Andrew (2005)

-1.006

44

42

5

0

-1.006

Pittler, et al. (2003)

-0.201

197

183

5

0

-0.201

Blomhoff, et al. (2001)

-0.256

91

88

5

0

-0.256

Stein, et al. (2000)

-0.146

1872

1824

5

0

-0.146

Blomhoff, et al. (2001)

-0.508

88

88

5

0

-0.508

Stein, et al (2006 )

-0.177

1270

1237

5

0

-0.177

Borkovec & Costello (1993)

-0.342

18

20

4

1

-0.128

White & Keenan (1992)

-0.119

26

10

3

1

-0.030

Borkovec & Mathews (1988)

-0.410

10

10

5

0

-0.410

White & Keenan (1992)

-0.354

31

10

3

1

-0.089

Borkovec, et al (1987)

-0.367

16

14

4

0

-0.275

White & Keenan (1992)

-0.318

31

10

3

1

-0.080

Butler, et al., (1991)

-1.203

19

19

5

0

-1.203

Adjusted
Effect Size
Used in the
Fixed Effects Model
Random Effects Model
BenefitCost
Homogeneity Weighted Mean Effect Size &
Analysis
Weighted Mean Effect Size & p-value
Test
p-value

Treatments for Non-Affective
Psychosis (Including
Schizophrenia)
Number of trials used in analysis:

Results Before Adjusting Effect Sizes

49

Number of subjects in treatment group: 3,926

ES

p-value

p-value

ES

p-value

ES

-0.370

0.000

0.000

-0.423

0.000

-0.324

Studies Used in the Meta-Analysis
Name of Study

Essm

N Tx

N Cn

Score

real

ESAdj Name of Study

Essm
-0.542

N Tx

N Cn

Score world =1

ESAdj

Aber-Wistedt, et al. (1995)

-0.557

20

20

5

0

-0.557

Haddock, et al., (2006)

Barrowclough, et al. (2001)

-0.633

17

15

5

1

-0.317

Hoult, et al. (1984)

Bigelow, et al. (1991)

-0.812

15

7

3

0

-0.406

James, et al. (2004)

Bond, et al. (1988)

-0.460

84

83

5

0

-0.460

Joy, et al. (2004)

-0.202

228

237

5

0

-0.202

Bond, et al. (1988)

-0.516

84

83

5

0

-0.516

Lehman, et al. (1994)

-0.201

359

302

3

0

-0.101

Bond, et al. (1990)

-0.743

42

40

5

0

-0.743

Lehman, et al. (1997)

-0.354

77

75

5

0

-0.354

Bond, et al. (1991)

-1.098

30

10

3

1

-0.274

Lewis, et al. (2005)

-0.308

40

38

5

0

-0.308

Bond, et al. (1995)

-0.502

39

35

3

0

-0.251

Macias, et al. (1994)

-0.802

19

18

4

0

-0.602

Bush. et al. (1990)

-0.832

14

14

5

0

-0.832

Marques (2004)

-0.266

208

207

5

0

-0.266

-0.541
-0.60129

15

14

5

1

-0.271

26

25

5

0

-0.541

29.000

29

5

0 -0.60129

Curtis, et al. (1996)

-0.023

147

145

5

0

-0.023

McFarlane (2002)

-0.514

27

14

5

0

-0.514

Chandler, et al. (1996)

-0.450

115

108

5

0

-0.450

McFarlane (2002)

-0.291

50

50

5

0

-0.291

Chandler, et al. (1997)

-0.431

105

105

3

1

-0.108

McFarlane (2002)

-0.297

34

34

5

0

-0.297

Chandler, et al. (1997)

-0.431

105

105

3

1

-0.108

McFarlane, et al. (1995)

-0.289

83

89

5

0

-0.289

Drake et al. (1998)

-0.017

105

98

5

0

-0.017

McFarlane, et al. (2000)

0.585

37

32

5

0

0.585

Drake, et al. (1996)

-0.660

39

35

3

0

-0.330

Morse, et al. (1997)

-0.421

90

45

4

0

-0.316

Drake, et al. (1999)

-1.166

74

76

3

1

-0.292

Mota Neto, et al. (2002)

-0.590

159

83

5

0

-0.590

Dyck, et al. (2002)

-0.428

55

51

5

0

-0.428

Quinlivan, et al. (1995)

-0.510

30

30

5

0

-0.510

Dyck, et al. (2002)

-0.150

56

150

3

0

-0.075

Shern, et al. (2000)

-0.453

91

77

5

0

-0.453

El-Sayeh & Morganti (2006)

-0.452

155

155

5

0

-0.452

Test, et al. (1980)

-0.321

54

57

3

0

-0.160

Essock, et al. (1995)

-0.503

58

50

5

0

-0.503

Test, et al. (1991)

-0.680

75

47

5

0

-0.680

Fekete, et al. (1998)

-0.534

58

50

3

0

-0.267

Tharyan, et al. (2005)

-0.363

214

178

5

0

-0.363

Ford, et al. (1996)

-0.375

47

47

3

0

-0.188

Thornley, et al. (2003)

-0.229

264

248

5

0

-0.229

Gervey, et al. (1994)

-1.584

17

17

3

1

-0.396

Wilson, et al. (1995)

-0.511

26

33

5

0

-0.511

Goering, et al. (1988)

-0.189

82

82

3

0

-0.094

Wood, et al. (1995)

-0.753

32

32

3

0

-0.376

Gold, et al. (2005)

-0.678

99

81

5

0

-0.678

Note: Treatments in Assertive Community Treatment were predominantly schizophrenics but included people with other serious mental illness.

16

Exhibit A.5: Citations of Studies Used in the Meta-Analysis
Aalto, M., Saksanen, R., Laine, P., Forsstrom, R., Raikaa, M., Kiviluoto, M., et al. (2000). Brief intervention for female heavy drinkers in routine general practice: A 3-year
randomized controlled study. Alcoholism: Clinical and Experimental Research, 24(11): 1680-1686.
Åberg-Wistedt, A., Cresell, T., Lindberg, Y., Liljenberg, B., & Ösby, U. (1995). Two-year outcome of team-based intensive case management for patients with schizophrenia.
Psychiatric Services, 46(12): 1263-1266.
Adams, D.R., (1990). An early counseling intervention program for problem drinkers contrasting group and individual delivery formats (group treatment). Doctoral
dissertation, University of British Columbia, Canada.
Allsop, S., Saunders, B., Phillips, M., & Carr, A. (1997). A trial of relapse prevention with severely dependent male problem drinkers. Addiction, 92(1): 61-74.
Anderson, P. & Scott, E. (1992). The effect of general practitioners advice to heavy drinking men. British Journal of Addiction, 87: 891-900.
*Anton, R.F., Brady, K.T., & Moak, D.H. (1999). Pharmacotherapy. In P.J. Ott, R.E. Tarter, & R.T. Ammerman (Eds.), Sourcebook on substance abuse: Etiology,
epidemiology, assessment, and treatment (pp. 303-314). Boston: Allyn and Bacon.
Anton, R.F., Moak, D.H., Waid, L.R., Latham, P.K., Malcolm, R.J. & Dias, J.K. (1999). Naltrexone and cognitive behavioral therapy for the treatment of outpatient alcoholics:
Results of a placebo-controlled trial. American Journal of Psychiatry, 156(11): 1758-1764.
Anton, R.F., O'Malley, S.S., Ciraulo, D.A., Cisler, R.A., Couper, D., Donovan, D.M., et al. (2006). Combined pharmacotherapies and behavioral interventions for alcohol
dependence: The COMBINE study - A randomized controlled trial. Journal of the American Medical Association, 295(17): 2003-2017.
Avants, S.K., Margolin, A., Usubiaga, M.H. & Doebrick, C. (2004). Targeting HIV-related outcomes with intravenous drug users maintained on methadone: A randomized
clinical trial of a harm reduction group therapy. Journal of Substance Abuse Treatment, 26(2): 67-78.
Azrin, N.H. (1976). Improvements in the community-reinforcement approach to alcoholism. Behavior Research and Therapy, 14(5): 339-348.
Azrin, N.H., Acierno, R., Kogan, E., Donohue, B., Besalel, V. & McMahon, P.T. (1996). Follow-up results of supportive versus behavioral therapy for illicit drug abuse.
Behavioral Research and Therapy, 34(1): 41-46.
Azrin, N.H., Donohue, B., Besalel, V.A., Kogan, E.S. & Acierno, R. (1994). Youth drug abuse treatment: A controlled outcome study. Journal of Child and Adolescent
Substance Abuse 3(3): 1-16.
Babor, T.F., Lauerman, R., Kranzler, H., McRee, B., Korner, P., Wiber, C., et al. (1992). Farmington, USA. In T.F. Babor & M. Grant (Eds.), Project on identification and
management of alcohol-related problems report on phase II: A randomized clinical trial of brief interventions in primary care (pp. 191-209). Geneva, Switzerland: World
Health Organization.
Baker, A., Boggs, T.G., & Lewin, T.J. (2001). Randomized controlled trial of brief cognitive-behavioural interventions among regular users of amphetamine. Addiction, 96(9):
1279-1287.
Baker, A., Lee, N.K., Claire, M., Lewin, T.J., Grant, T., Pohlman, S., et al. (2005). Brief cognitive behavioural interventions for regular amphetamine users: A step in the right
direction. Addiction, 100(3): 367-378.
Barlow, D.H., Cohen, A.S., Waddell, M.R., Vermilyea, B.B., Klosko, J.S., Blanchard, E.B., et al. (1984). Panic and generalized anxiety disorder: Nature and treatment.
Behavior Therapy, 15: 431-449.
Barlow, D.H., Craske, M.G., Cerny, J.A., & Klosko, J.S. (1989). Behavioral treatment of panic disorder. Behavior Therapy, 20: 261-282.
Barlow, D.H., Gorman, J.M., Shear, M.K., & Woods, S.W. (2000). Cognitive-behavioral therapy, imipramine, or their combination for panic disorder: A randomized controlled
trial. Journal of the American Medical Association, 283: 2529-2536.
Barlow, D.H., Rapee, R.M., & Brown, T.A. (1992). Behavioral treatment of generalized anxiety disorder. Behavior Therapy, 23: 551-570.
Barrowclough, C., Haddock, G., Tarrier, N., Lewis, S., Moring, J. & OBrien, R. (2001). Randomized controlled trial of motivational interviewing, cognitive behavior therapy,
and family intervention for patients with comorbid schizophrenia and substance use disorders. American Journal of Psychiatry, 158(10): 1706-1713.
Beck, A.T., Sokol, L., Clark, D.A., Berchick, R., & Wright, F. (1992). A crossover study of focused cognitive therapy for panic disorder. American Journal of Psychiatry,
149(6): 778-783.
Bellack, A.S., Bennett, M.E., Gearon, J.S., Brown, C.H., & Yang, Y. (2006). A randomized clinical trial of a new behavioral treatment for drug abuse in people With severe
and persistent mental illness. Archives of General Psychiatry, 63(4): 426-432.
Bien, T. H., Miller, W. R., & Boroughs, J. M. (1993). Motivational interviewing with alcohol outpatients. Behavioural and Cognitive Psychotherapy, 21(4): 347-356.
Bigelow, D.A., & Young, D.J. (1991). Effectiveness of a case management program. Community Mental Health Journal, 27: 115-123.
Bisson, J. & Andrew, M. (2005). Psychological treatment of post-traumatic stress disorder (PTSD). The Cochrane Database of Systematic Reviews, 2: CD003388.
Blomhoff, S., Haug, T.T., Hellstrom, K., Holme, I., Humble, M., Madsbu, H.P., et al. (2001). Randomised controlled general practice trial of sertaline, exposure therapy, and
combined treatment in generalized social phobia. British Journal of Psychiatry, 179(1): 23-30.
*Bond, G.R., Becker, D.R. & Drake, R.E. (2001). Implementing supported employment as an evidence-based practice. Psychiatric Services, 52: 313-322.
Bond, G.R., Dietzen, L.L., McGrew, J.H., et al: (1995). Accelerating entry into supported employment for persons with severe psychiatric disabilities. Rehabilitation
Psychology, 40: 91-111.
*Bond, G.R., Drake, R.E., Mueser, K.T., & Becker, D.R. (1997). An update on supported employment for people with severe mental illness. Psychiatric Services, 48(3): 335346.
Bond, G.R., Miller, L.D., Krumwied, R.D., & Ward, R.S. (1988). Assertive case management in three CMHCs: a controlled study. Hospital and Community Psychiatry, 39:
411-418.
Bond, G.R., Pensec, M., Dietzen, L.L., McCafferty, D., Giemza, R., & Sipple, H.W. (1991). Intensive case management for frequent users of psychiatric hospitals in a large
city: a comparison of team and individual caseloads. Psychosocial Rehabilitation Journal, 15(1): 90-98.
Bond, G.R., Witheridge, T.F., Dincin, J., Wasmer, D., Webb, J., & DeGraaf-Kaser, R. (1990). Assertive community treatment for frequent users of psychiatric hospitals in a
large city: a controlled study. American Journal of Community Psychology, 18(6): 865-891.
Borkovec, T.D. & Costello, E. (1993). Efficacy of applied relaxation and cognitive-behavioral therapy in the treatment of generalized anxiety disorder. Journal of Consulting
and Clinical Psychology, 61(4): 611-619.
Borkovec, T.D. & Mathews, A.M. (1988). Treatment of nonphobic anxiety disorders: A comparison of nondirective, cognitive, and coping desensitization therapy. Journal of
Consulting and Clinical Psychology, 56(6): 877-884.
Borkovec, T.D., Mathews, A.M., Chambers, A., Ebrahimi, S., Lytle, R., & Nelson, R. (1987). The effects of relaxation training with cognitive therapy or nondirective therapy
and the role of relaxation-induced anxiety in the treatment of generalized anxiety. Journal of Consulting and Clinical Psychology, 55(6): 883-888.
Borsari, B. & Carey, K.B. (2000). Effects of a brief motivational intervention with college student drinkers. Journal of Consulting and Clinical Psychology, 68(4): 728-733.
Bowers, T.G. & Al-Redha, M.R. (1990). A comparison of outcome with group/marital and standard/individual therapies with alcoholics. Journal of Studies on Alcohol, 51(4):
301-309.
Brown, J. M., & Miller, W. R. (1993). Impact of motivational interviewing on participation and outcome in residential alcoholism treatment. Psychology of Addictive Behaviors,
7(4): 211-218.
Browne, G., Steiner, M., Roberts, J., Gafni, A., Byrne, C., Dunn, E., et al. (2002). Sertaline and/or interpersonal psychotherapy for patients with dysthymic disorder in
primary care: 6-month comparison with longitudinal 2-year follow-up of effectiveness and costs. Journal of Affective Disorders, 68: 317-330.
Burgess, S., Geddes, J., Hawton, K., Townsend, E., Jamison, K., & Goodwin, G. (2001). Lithium for maintenance treatment of mood disorders. The Cochrane Database of
Systematic Reviews, 3: CD003013.
Bush, C.T., Langford, M.W., Rosen, P., & Gott, W. (1990). Operation outreach: Intensive case management for severely psychiatrically disabled adults. Hospital and
Community Psychiatry, 41: 647-649.
Butler, G., Fennell, M., Robson, P., & Gelder, M. (1991). Comparison of behavior therapy and cognitive behavior therapy in the treatment of generalized anxiety disorder.
Journal of Consulting and Clinical Psychology, 59(1): 167-175.
Carroll, K.M., Rounsaville, B.J. & Gawin, F.H. (1991). A comparative trial of psychotherapies for ambulatory cocaine abusers: Relapse prevention and interpersonal therapy.
American Journal of Drug and Alcohol Abuse, 17(3): 229-247.
Carroll, K.M., Rounsaville, B.J., Nich, C., Gordon, L.T., Wirtz, P.W. & Gawin, F.H. (1994). One-Year Follow-up of Psychotherapy and Pharmacotherapy for Cocaine
Dependence: Delayed Emergence of Psychotherapy Effects. Archives of General Psychiatry, 51(12): 989-997.
Catalano, R.F., Haggerty, K.P., Fleming, C.B., Brewer, D.D., & Gainey, R.R. (2002). Children of substance abusing parents: Current findings from the Focus on Families
project. New York: Kluwer Academic Press/Plenum Publishers.
Chandler, D., Meisel, J., Hu, T., et al. (1997). A capitated model for a cross-section of severely mentally ill clients: employment outcomes. Community Mental Health
Journal, 33: 501-516.

17

Exhibit A.5: Citations of Studies Used in the Meta-Analysis, continued
Chandler, D., Meisel, J., McGowen, M., Mintz, J., & Madison, K. (1996). Client outcomes in two model capitated integrated service agencies. Psychiatric Services, 47(1):
175-180.
Chaney, E.F., OLeary, M.R. & Marlatt, A.G. (1978). Skill training with alcoholics. Journal of Consulting and Clinical Psychology, 46: 1092-1104.
Chick, J., Lloyd, G. & Crombie, E. (1985). Counseling problem drinkers in medical wards: A controlled study. British Medical Journal, 290: 965-967.
Chick, J., Ritson, B., Connaughton, J., Stewart, A., & Chick, J. (1988). Advice versus extended treatment for alcoholism: A controlled trial. British Journal of Addiction, 83(2):
159-170.
Collins, S.E., Carey, K.B., & Sliwinski, M.J. (2002). Mailed personalized normative feedback as a brief intervention for at-risk college drinkers. Journal of Studies on Alcohol,
63(5): 559-567.
Cordioli, A.V., Heldt, E., Bochi, D.B., Margis, R., de Sousa, M.B., Tonello, J.F., et al. (2003). Cognitive-behavioral group therapy in obsessive-compulsive disorder: a
randomized clinical trial. Psychotherapy and Psychosomatics, 72(4): 211-6.
Cornish, J.W., Metzger, D., Woody, G.E., Wilson, D., McLellan, A.T., Vandergrift, B., et al. (1997). Naltrexone pharmacotherapy for opioid dependent federal probationers.
Journal of Substance Abuse Treatment, 14(6): 529-534.
Crits-Christoph, P., Siqueland, L., Blaine, J., Frank, A., Luborsky, L. & Onken, L.S. (1999). Psychosocial treatments for cocaine dependence: National institute on drug
abuse collaborative cocaine treatment study. Archives of General Psychiatry, 56(6): 493-502.
Curtis, J.L., Millman, E.J., Struening, E., et al. (1996). Deaths among former psychiatric inpatients in an outreach case management program. Psychiatric Services, 47(4):
398-402.
Dole, V.P., Robinson, J.W., Orraca, J., Towns, E., Searcy, P., & Caine, E. (1969). Methadone treatment of randomly selected criminal addicts. New England Journal of
Medicine, 280(25): 1372-1375.
Donovan, D.M. & Ito, J.R. (1988). Cognitive behavioral relapse prevention strategies and aftercare in alcoholism rehabilitation. Psychology of Addictive Behaviors, 2(2): 74.
Drake, R.E., McHugo, G.J., Bebout, R.R., Becker, D.R., Harris, M., Bond, G.R., et al. (1999). A Randomized Clinical Trial of Supported Employment for Inner-city Patients
With Severe Mental Disorders. Archives of General Psychiatry, 56(7): 627-633.
Drake, R.E., McHugo, G.J., Becker, D.R., Anthony, W.A., & Clark, R.E. (1996). The New Hampshire study of supported employment for people with severe mental illness.
Journal of Consulting and Clinical Psychology, 64(2): 391-39.
Drake, R.E., McHugo, G.J., Clark, R.E., et al: (1998a). Assertive community treatment for patients with co-occurring severe mental illness and substance use disorder: a
clinical trial. American Journal of Orthopsychiatry, 68(2): 201-215.
Drake, R.E., Mercer-McFadden, C., Mueser, K.T., McHugo, G.J. & Bond, G.R. (1998b). Review of integrated mental health and substance abuse treatment for patients With
dual disorders. Schizophrenia Bulletin, 24(4): 589-608.
*Drake, R.E., Mueser, K.T., Brunette, M.F., & McHugo, G.J. (2004). A review of treatments for people with severe mental illnesses and co-occurring substance use
disorders. Psychiatric Rehabilitation Journal, 27(4): 360-374.
Drake, R.E., Xie, H., McHugo, G.J. & Green, A.I. (2000). The effects of clozapine on alcohol and drug use disorders among patients With schizophrenia. Schizophrenia
Bulletin, 26(2): 441-449.
Drake, R.E., Yovetich, N.A., Bebout, R.R., Harris, M. & McHugo, G.J. (1997). Integrated treatment for dually diagnosed homeless Adults. Journal of Nervous and Mental
Disease, 185(5): 298-305.
Dugas, M.J., Ladouceur, R., Leger, E., Freeston, M.H., Langlois, F., Provencher, M.D., et al. (2003). Group cognitive-behavioral therapy for generalized anxiety disorder:
treatment outcome and long-term follow-up. Journal of Consulting and Clinical Psychology, 71(4): 821-825.
Durham, R.C., Murphy, T. Allan, T., Richard, K., Treliving, L. R., & Fenton, G.W. (1994). Cognitive therapy, analytic psychotherapy and anxiety management training for
generalized anxiety disorder. British Journal of Psychiatry, 165: 315-323.
Dyck, D.G., Hendryx, M.S., Short, R.A., Voss, W.D., & McFarlane, W.R. (2002). Service use among patients with schizophrenia in psychoeducational multiple-family group
treatment. Psychiatric Services, 53(6): 749-754.
El-Sayeh, H.G. & Morganti, C. (2006). Aripiprazole for schizophrenia. The Cochrane Database of Systematic Reviews, 2: CD004578.
Elvy, G.A., Wells, J.E. & Baird, K.A. (1988). Attempted referral as intervention for problem drinking in the general hospital. British Journal of Addiction, 83(1): 83-89.
Eriksen, L., Björnstad, S., & Götestam, K. G. (1986). Social skills training in groups for alcoholics: One-year treatment outcome for groups and individuals. Addictive
Behaviors, 11(3): 309-330.
Essock, S.M., & Kontos, N. (1995). Implementing assertive community treatment teams. Psychiatric Services, 46(7): 679-683.
Fals-Stewart, W., Birchler, G.R., & Ofarrell, T.J. (1996). Behavioral couples therapy for male substance-abusing patients: Effects on relationship adjustment and drug-using
behavior. Journal of Consulting and Clinical Psychology, 64(5): 959-972.
Fava, G.A., Rafanelli, C., Grandi, S., Canestrari, R., & Morphy, M.A. (1998) Six-year outcomes for cognitive behavioral treatment of residual symptoms in major depression.
The American Journal of Psychiatry, 155(10): 1443-1445.
Feeney, G., Young, R., Connor, J., Tucker, J., & McPherson, A. (2002). Cognitive behaviour therapy combined with the relapse-prevention medication acamprosate: are
short-term treatment outcomes for alcohol dependence improved? The Australian and New Zealand Journal of Psychiatry, 36(5): 622-628.
Fekete, D.M., Bond, G.R., McDonel, E.C., Salyers, M.P., Chen, A., & Miller, L. (1998). Rural assertive community treatment: a field experiment. Psychiatric Rehabilitation
Journal, 21(4): 371-379.
Ferrell, W. L., & Galassi, J. P. (1981). Assertion training and human relations training in the treatment of chronic alcoholics. International Journal of the Addictions, 16(5):
959-968.
Fichter, M.M., & Frick, U. (1993). The key relative’s impact on treatment and course of alcoholism. European Archives of Psychiatry and Clinical Neuroscience, 243(2): 87-94.
Fleming, M., Mundt, M.P., French, M.T., Manwell, L.B., Stauffacher, E.A., & Barry, K.L. (2000). Benefit-cost analysis of brief physician advice with problem drinkers in
primary care settings. Medical Care, 38(1): 7-18.
Ford, R., Ryan, P., & Norton, P., et al. (1996). Does intensive case management work? Clinical, social, and quality of life outcomes from a controlled study. Journal of
Mental Health, 5(4): 361-368.
Fudala, P.J., Bridge, T.P., Herbert, S., Williford, W.O., Chiang, C.N., Jones, K., et al. (2003). Office-based treatment of opiate addiction with a sublingual-tablet formulation
of buprenorphine and naloxone. New England Journal of Medicine, 349(10): 949-958.
Gervey, R. & Bedell, J.R. (1994). Supported employment in vocational rehabilitation. In J.R. Bedell (Ed.), Psychological Assessment and Treatment of Persons With Severe
Mental Disorders (pp. 151-175). Washington, DC, Taylor & Francis.
Goering, P.N., Farkas, M., Wasylenki, D., et al. (1988). Improved functioning for case management clients. Psychosocial Rehabilitation Journal, 12(1): 3-17.
Gold, C., Heldal, T.O., Dahle, T., & Wigram, T. (2005). Music therapy for schizophrenia or schizophrenia-like illnesses. The Cochrane Database of Systematic Reviews, 2:
CD004025.
Graber, R.A. & Willer, W.R. (1988). Abstinence or controlled drinking goals for problem drinkers: A randomized clinical trial. Psychology of Addictive Behaviors, 2(1): 20-33.
Gronbladh, L. & Gunne, L. (1989). Methadone-assisted rehabilitation of Swedish heroin addicts. Drug and Alcohol Dependence, 24(1): 31-37.
Haddock, G., Barrowclough, C., Tarrier, N., Moring, J., O'Brien, R., Schofield, N., et al. (2006). Cognitive-behavioural therapy and motivational intervention for schizophrenia
and substance misuse: 18-month outcomes of a randomized controlled trial. British Journal of Psychiatry, 183: 418-426.
Handmaker-Sheehy, N., Miller, W.R., & Manicke, M. (1999). Findings of a pilot study of motivational interviewing with pregnant drinkers. Journal of Studies on Alcohol,
62(2): 285-287.
Harris, K.B. & Miller, W.R. (1990). Behavioral self-control training for problem drinkers: Mechanisms of efficacy. Psychology of Addictive Behaviors, 4(2): 82-90.
Heather, N. (1996). The public health and brief interventions for excessive alcohol consumption: The British experience. Addictive Behaviors, 21(6): 857.
Heather, N., Campion, P.D., Neville, R.G. & Maccabe, D. (1987). Evaluation of a controlled drinking minimal intervention for problem drinkers in general practice (The
DRAMS Scheme). Journal of the Royal College of General Practitioners, 37(301): 358-363.
Hedberg, A. G., & Campbell, L. M. (1974). A comparison of four behavioral treatment approaches to alcoholism. Journal of Behavior Therapy and Experimental Psychiatry,
5(3/4): 251-256.
Hellerstein, D.J., Rosenthal, R.N. & Miner, C.R. (1995). Prospective study of integrated outpatient treatment for substance-abusing schizophrenic patients. American Journal
on Addictions, 4(1): 33-42.
Hester, R.K. & Delaney, H.D. (1997). Behavioral self-control program for windows: Results of a controlled clinical trial. Journal of Consulting and Clinical Psychology, 65(4):
686-693.
Higgins, S.T., Wong, C.J., Badter, G.J., Haug Ogden, D.E. & Dantona, R.L. (2000). Contingent reinforcement increases cocaine abstinence during outpatient treatment and
1 year of follow-up. Journal of Consulting and Clinical Psychology, 68(1): 64-72.

18

Exhibit A.5: Citations of Studies Used in the Meta-Analysis, continued
Hoult, J. & Reynolds, I. (1984). Schizophrenia: A comparative trial of community orientated and hospital orientated psychiatric care. Acta Psychiatrica Scandinavica, 69(5):
359-372.
Hulse, G.K. & Tait, R.J. (2002). Six-month outcomes associated with brief alcohol intervention for adult in-patients with psychiatric disorders. Drug and Alcohol Review, 21:
105-112.
Humphreys, K., Huebsch, P.D., Finney, J.W., & Moos, R.H. (1999). A comparative evaluation of substance abuse treatment: V. Treatment can enhance the effectiveness of
self-help groups. Alcoholism: Clinical and Experimental Research, 23(3): 558-563.
Hunt, G.M. & Azrin, N.H. (1973). A community reinforcement approach to alcoholism. Behavior Research and Therapy, 11(1): 91-104.
James, W., Preston, N.J., Koh, G., Spencer, C., Kisely, S.R., & Castle, D.J. (2004). A group intervention which assists patients with dual diagnosis reduce their drug use: A
randomized controlled trial. Psychological Medicine, 34(6): 983-90.
Johnson, R.E., Chutuape, M.A., Strain, E.C., Walsh, S.L., Stitzer, M.L. & Bigelow, G.E. (2000). A comparison of levomethadyl acetate, buprenorphine, and methadone for
opioid dependence. The New England Journal of Medicine, 343(18): 1290-1297.
Johnson, R.E., Eissenberg, T., Stitzer, M.L., Strain, E.C., Liebson, I.A., & Biglow, G.E. (1995). A placebo controlled clinical trial of buprenorphine as a treatment for opioid
dependence. Drug and Alcohol Dependence, 40(1): 17-25.
Johnson, R.E., Jaffe, J.H., & Fudala, P.J. (1992). A controlled trial of buprenorphine treatment for opioid dependence. Journal of the American Medical Association,
267(20):2750-2755.
Jones, S. L., Kanfer, R., & Lanyon, R. I. (1982). Skill training with alcoholics: A clinical extension. Addictive Behaviors, 7(3): 285-290.
Joy, C.B., Adams C.E., & Rice K. (2004). Crisis intervention for people with severe mental illnesses. The Cochrane Database of Systematic Reviews, 4: CD001087.
Kapczinski, F., Lima, M.S., Souza, J.S., Cunha, A., & Schmitt, R. (2003). Antidepressants for generalized anxiety disorder. The Cochrane Database of Systematic Reviews,
2: CD003592.
Kavanagh, D.J., Young, R., White, A., Saunders, J.B., Wallis, J., & Shockley, N. (2004). A brief motivational intervention for substance misuse in recent-onset psychosis.
Drug and Alcohol Review, 23(2): 151-155.
Kelly, A.B., Halford, W.K., & Young, R.M. (2000). Maritally distressed women with alcohol problems: The impact of short-term alcohol-focused intervention on drinking
behavior and marital satisfaction. Addiction, 95(10): 1537-1549.
Kivlahan, D.R., Marlatt, G.A., Fromme, K., Coppel, D.B., & Williams, E. (1990). Secondary prevention with college drinkers: Evaluation of an alcohol skills training program.
Journal of Consulting and Clinical Psychology, 58(6): 805-810.
Kristenson, H., Ohlin, H., Hulten-Nosslin, M.B., Trell, E. & Hood, B. (1983). Identification and intervention of heavy drinking in middle-aged men: Results and follow-up of 2460 months of long-term study with randomized controls. Alcoholism: Clinical and Experimental Research, 7(2): 203-209.
*Kreek, M.J. & Vocci, F.J. (2002). History and current status of opioid maintenance treatments: Blending conference session. Journal of Substance Abuse Treatment, 23(2):
93-105.
Kuchipudi, V., Hobein, K., Flickinger, A. & Iber, F.L. (1990). Failure of a 2-hour motivational intervention to alter recurrent drinking behavior in alcoholics with gastrointestinal
disease. Journal of Studies on Alcohol, 51(4): 356-360.
Ladouceur, R., Dugas, M., Freeston, M.H., Leger, E., Gagnon, F., & Thibodeau, N. (2000). Efficacy of a cognitive-behavioral treatment for generalized anxiety disorder:
Evaluation in a controlled clinical trial. Journal of Consulting and Clinical Psychology, 68(6): 957-964.
Larimer, M.E., Turner, A.P., Anderson, B.K., Fader, J.S., Kilmer, J.R., Palmer, R.S., et al (2001). Evaluating a brief alcohol intervention with fraternities. Journal of Studies
on Alcohol, 62(3): 370-380.
*Latimer, E.A. (1999). Economic impacts of assertive community treatment: A review of the literature. Canadian Journal of Psychiatry, 44(5): 443-454.
Lehman, A.F., Dixon, L.B., Kernan, E., et al. (1997). A randomized trial of assertive community treatment for homeless persons with severe mental illness. Archives of
General Psychiatry, 54(11): 1038.
Lehman, A.F., Postrado, L.T., McNary, S.W., & Goldman, H.H. (1994). Continuity of care and client outcomes in the Robert Wood Johnson Foundation program on chronic
mental illness. Milbank Quarterly, 72: 105-122.
Lewis, R., Bagnall, A.M., & Leitner, M. (2005). Sertindole for schizophrenia. The Cochrane Database of Systematic Reviews, 3: CD001715.
Lhuintre, J.P., Daoust, M., Moore, N. D., Chretien, P., Saligaut, C., Tran, G., et al. (1985). Ability of calcium bis acetyl homotaurine, a GABA agonist, to prevent relapse in
weaned alcoholics. Lancet, 1(8436): 1014-1016.
Lhuintre, J.P., Moore, N., Tran, G., Steru, L., Langrenon, S., Daoust, M., et al. (1990). Acamprosate appears to decrease alcohol intake in weaned alcoholics. Alcohol and
Alcoholism, 25(6): 613-622.
Lima, M.S., Moncrieff, J., & Soares, B.G.O. (2006). Drugs versus placebo for dysthymia. The Cochrane Database of Systematic Reviews, 2: CD001130.
Lindsay, W.R., Gamsu, C.V., McLaughlin, E., Hood, E.M., & Espie, C.A. (1987). A controlled trial of treatments for generalized anxiety. British Journal of Clinical
Psychology, 26(1): 3-15.
Linehan, M.M., Schmidt, H., Dimeoff, L.A., Craft, J.C., Kanter, J. & Comtois, K.A. (1999). Dialectical behavior therapy for patients with borderline personality disorder and
drug-dependence. American Journal on Addictions, 8(4): 279-292.
Ling, W., Charuvastra, C., et al. (1998). Buprenorphine maintenance treatment of opiate dependence: A multicenter, randomized clinical trial. Addiction, 93(4): 475.
Macias, C., Kinney, R., Farley, O.W., Jackson, R., & Vos, B. (1994). The role of case management within a community support system: partnership with psychosocial
rehabilitation. Community Mental Health Journal, 30: 323-339.
Macritchie, K., Geddes, J.R., Scott, J., Haslam, D., de Lima, M., Goodwin, G. (2003). Valproate for acute mood episodes in bipolar disorder. The Cochrane Database of
Systematic Reviews, 1: CD004052.
Maheswaran, R., Beevers, M. & Beevers, D.G. (1992). Effectiveness of advice to reduce alcohol consumption in hypertensive patients. Hypertension, 19(1): 79-84.
Mallams, J. H., Godley, M. D., Hall, G. M., & Meyers, R. J. (1982). A social-systems approach to resocializing alcoholics in the community. Journal of Studies on Alcohol,
43(11): 1115-1123.
Manwell, L.B., Fleming, M.F., Mundt, M.P., Staffacher, E.A., & Barry, K.L. (2000). Treatment of problem alcohol use in women of childbearing age: Results of a brief
intervention trial. Alcoholism: Clinical and Experimental Research, 24(10): 1517-1524.
Margolin, A., Avants, S.K., Warburton, L.A., Hawkins, K.A. & Shi, J. (2003). A randomized clinical trial of a manual-guided risk reduction intervention for HIV-positive
injection drug users. Health Psychology, 22(2): 223-228.
Marijuana Treatment Project Research Group. (2004). Brief treatments for cannabis dependence: Findings from a randomized multisite trial. Journal of Consulting and
Clinical Psychology, 72(3): 455-466.
Marks, I.M., Swinson, R.P., Basaglu, M., et al. (1993). Alprazolam and exposure alone and combined in panic disorder with agoraphobia: A controlled study in London and
Toronto. British Journal of Psychiatry, 162: 776-787.
Marlatt, G.A., Baer, J.S., Kivlahan, D.R., Dimeff, L.A., Larimer, M.E., Quigley, L.A., et al. (1998). Screening and brief intervention for high-risk college student drinkers:
Results from a 2-year follow-up assessment. Journal of Consulting and Clinical Psychology, 66(4): 604-615.
Marques, L.O., Lima, M.S., & Soares, B.G.O. (2004). Trifluoperazine for schizophrenia. The Cochrane Database of Systematic Reviews, 1: CD003545.
Mason, B. J., Ritvo, E. C., Morgan, R. O., Salvato, F. R., Goldberg, G., Welch, B., et al. (1994). A double-blind, placebo-controlled pilot study to evaluate the efficacy and
safety of oral nalmefene HCl for alcohol dependence. Alcoholism: Clinical and Experimental Research, 18(5): 1162-1167.
Mason, B.J., Salvato, F.R., Williams, L.D., Rivto, E.C., & Culter, R.B. (1999). A double-blind, placebo controlled study of oral nalmefene for alcohol dependence. Archives of
General Psychiatry, 56(8): 719-724.
McCrady, B., Epsteien, E.E., & Hirsch, L.S. (1999). Maintaining change after conjoint behavioral alcohol treatment for men: Outcomes at 6 months. Addiction, 94(9): 13811396.
McFarlane, W.R. (2002). Empirical studies of outcome in multifamily groups. In W.R. McFarlane (Ed.), Multifamily groups in the treatment of severe psychiatric disorders
(pp. 49-70). New York: The Guilford Press.
McFarlane, W.R., Dushay, R.A., Deakins, S.M., et al. (2000). Employment outcomes in family-aided assertive community treatment. American Journal of Orthopsychiatry,
70(2): 203-214.
McFarlane, W.R., Lukens, E., Link, B., Dushay, R., Deakins, S.A. & Newmark, M. (1995). Multiple-family groups and psychoeducation in the treatment of schizophrenia.
Archives of General Psychiatry, 52(8): 679-687.
Milby, J.B., Schumacher, J.E., Raczynski, J.M., Caldwell, E., Engle, M., Michael, M., et al. (1996). Sufficient conditions for effective treatment of substance abusing
homeless persons. Drug and Alcohol Dependence, 43(1): 39-47.
Miller, W. R., Benefield, R. G., & Tonigan, J. S. (1993). Enhancing motivation for change in problem drinking: A controlled comparison of two therapist styles. Journal of
Consulting and Clinical Psychology, 61(3): 455-461.

19

Exhibit A.5: Citations of Studies Used in the Meta-Analysis, continued
Miller, W.R., Gribskov, C.J. & Mortell, R.L. (1981). Effectiveness of a self-control manual for problem drinkers with and without therapist contact. International Journal of the
Addictions, 16(7): 1247-1254.
Miller, W.R., Meyers, R.J., Tonigan, J.S., & Grant, K.A. (2001). Community reinforcement and traditional approaches: Findings of a controlled trial. In R.J. Meyers & W.R.
Miller (Eds.), A community reinforcement approach to addition treatment (pp. 79-103). Cambridge UK: Cambridge University Press.
Miller, W.R., Taylor, C.A. & West, J.C. (1980). Focused versus broad-spectrum behavior therapy for problem drinkers. Journal of Consulting and Clinical Psychology, 48(5):
590-601.
*Miller, W. & Wilbourne, P. (2002). Mesa Grande: A methodological analysis of clinical trials of treatments for alcohol use disorders. Addiction, 97(2): 265-277.
Moncrieff, J., Wessely, S., & Hardy, R. (2004). Active placebos versus antidepressants for depression. The Cochrane Database of Systematic Reviews, 1: CD003012.
Monti, P. M., Abrams, D. B., Binkoff, J. A., Zwick, W. R., Liepman, M. R., Nirenberg, T. D., et al. (1990). Communication skills training, communication skills training with
family and cognitive behavioral mood management training for alcoholics. Journal of Studies on Alcohol, 51(3): 263-270.
Monti, P. M., Rohsenow, D. J., Rubonis, A. V., Niaura, R. S., Sirota, A. D., Colby, S. M., et al. (1993). Cue exposure with coping skills treatment for male alcoholics: A
preliminary investigation. Journal of Consulting and Clinical Psychology, 61(6): 1011-1019.
Morse, G.A., Calsyn, R.J., Klinkenberg, W.D., et al. (1997). An experimental comparison of three types of case management for homeless mentally ill persons. Psychiatric
Services, 48(4): 497-503.
Mörtberg, E, Karlsson, A., Fyring, C. & Sundin, Ö. (2005). Intensive cognitive-behavioral group treatment (CBGT) of social phobia: A randomized controlled study. Journal of
Anxiety Disorders, in press.
Mota Neto, J.I.S., Lima, M.S., & Soares, B.G.O. (2002). Amisulpride for schizophrenia. The Cochrane Database of Systematic Reviews, 2: CD001357.
*Mueser, K.T., Bond, G.R., Drake, R.E., & Resnick, S.G. (1998). Models of community care for severe mental Illness: A review of research on case management.
Schizophrenia Bulletin, 24(1): 37-74.
Murphy, J.G., Duchnick, J.J., Vuchinich, R.E., Davison, J.W., Karg, R.S., Olson, A.M., et al. (2001). Relative efficacy of a brief motivational intervention for college student
drinkers. Psychology of Addictive Behaviors, 15(4): 373-379.
Neighbors, C., Larimer, M.E., & Weis, M.A. (2004). Targeting misperceptions of descriptive drinking norms: Efficacy of a computer-delivered personalized normative
feedback intervention. Journal of Consulting and Clinical Psychology, 72(3): 434-447.
Nelson, J.E., & Howell, R.J. (1983). Assertiveness training using rehearsal and modeling with male alcoholics. American Journal of Drug and Alcohol Abuse, 9(3): 309-323.
Newman, R.G. & Whitehill, W.B. (1979). Double-blind comparison of methadone and placebo maintenance treatments of narcotic addicts in Hong Kong. Lancet, 2: 485.
Nilssen, O. (1991). The Tromso study: identification of and a controlled intervention on a population of early-stage risk drinkers. Preventive Medicine, 20(4): 518-528.
Obolensky, M. R. (1984). Recidivism among male DWI first offenders and intervention. Doctoral dissertation, University of Rhode Island (UMI No. 8508028).
O’Connell, J.M. (1987). Effectiveness of an alcohol relapse prevention program. Doctoral dissertation, Fordham University (UMI No. 8725685).
Oei, T. P. S., & Jackson, P. R. (1980). Long-term effects of group and individual social skills training with alcoholics. Addictive Behaviors, 5(2): 129-136.
Oei, T. P. S., & Jackson, P. R. (1982). Social skills and cognitive behavioral approaches to the treatment of problem drinking. Journal of Studies on Alcohol, 43(5): 532-547.
O’Farrell, T.J. (1993). Couples relapse prevention sessions after a behavioral marital therapy couples group program. In T.J. O'Farrell (Ed.), Treating alcohol problems:
Marital and family interventions (pp. 305-326). New York: Guilford Press.
O’Malley, S.S., Jaffe, A.J., Chang, G., Schottenfeld, R.S., Meyer, R.E. & Rounsaville, B. (1992). Naltrexone and coping skills therapy for alcohol dependence: A controlled
study. Archives of General Psychiatry, 49(11): 881-887.
Ouimette, P.C., Finney, J.W. & Moos, R.H. (1997). Twelve-step and cognitive-behavioral treatment for substance abuse: A comparison of treatment effectiveness. Journal
of Consulting and Clinical Psychology, 65(2): 230-240.
Paille, F. M., Guelfi, J. D., Perkins, A. C., Royer, R. J., Steru, L., & Parot, P. (1995). Doubleblind randomized multicentre trial of acamprosate in maintaining abstinence from
alcohol. Alcohol and Alcoholism, 30(2): 239-247.
Persson, J. & Magnusson, P. H. (1989). Early intervention in patients with excessive consumption of alcohol: A controlled study. Alcohol, 6(5): 403.
Petry, N.M. & Martin, B. (2002). Low-cost contingency management for treating cocaine- and opioid-abusing methadone patients. Journal of Consulting and Clinical
Psychology, 70(2): 398-405.
Petry, N.M., Martin, B., Cooney, J.L. & Kranzler, H.R. (2000). Give them prizes, and they will come: Contingency management for treatment of alcohol dependence. Journal
of Consulting and Clinical Psychology, 68(2): 250-257.
Piotrowski, N.A., Tusel, D.J., Sees, K.L., Reilly, P.M., Banys, P., Meek, P., et al. (1999). Contingency contracting with monetary reinforcers for abstinence from multiple
drugs in a methadone program. Experimental and Clinical Psychopharmacology, 7: 399-411.
Pittler, M.H. & Ernst, E. (2003). Kava extract versus placebo for treating anxiety. The Cochrane Database of Systematic Reviews, 1: CD003383.
Quinlivan, R., Hough, R., Crowell, A., Beach, C., Hofstetter, R., & Kenworthy, K. (1995). Service utilization and costs of care for severely mentally ill clients in an intensive
case management program. Psychiatric Services, 46(4): 365-371.
Rawson, R.A., Shoptaw, S.J., Obert, J.L., McCann, M.J., L Hasson, A., Marinelli-Casey, P.J., et al. (1995). An intensive outpatient approach for cocaine abuse treatment:
The matrix model. Journal of Substance Abuse Treatment, 12(2): 117-127.
Rendell, J.M., Gijsman, H.J., Keck, P., Goodwin, G.M., & Geddes, J.R. (2003). Olanzapine alone or in combination for acute mania. The Cochrane Database of Systematic
Reviews, 1: CD004040.
Reynolds, C.F., Frank, E., Perel, J.M., Imber, S.D., Cornes, C., Miller, et al. (2006). Nortriptyliine and interpersonal psychotherapy as maintenance therapies for recurrent
major depression: A randomized controlled trial in patients older than 59 years. New England Journal of Medicine 354: 1130-1138
Reynolds, K. D., Coombs, D. W., Lowe, J. B., Peterson, P. L., & Gayoso, E. (1995). Evaluation of a self-help program to reduce alcohol consumption among pregnant
women. International Journal of the Addictions, 30(4): 427-443.
Richmond, R., Heather, N., Wodak, A., Kehoe, L., & Webster, I. (1995). Controlled evaluation of a general practice-based brief intervention for excessive drinking. Addiction,
90(1): 119-132.
Rohsenow, D. J., Smith, R. E., & Johnson, S. (1985). Stress management training as a prevention program for heavy social drinkers: Cognitive, affect, drinking, and
individual differences. Addictive Behaviors, 10(1): 45-54.
Romelsjö, A., Andersson, L, Barrner, H., Borg, S., Granstrand, C., Hultman, O., et al. (1989). A randomized study of secondary prevention of early stage problem drinkers in
primary health care. British Journal of Addiction, 84(11): 1319-1327.
Sanchez-Craig, M., Davila, R., & Cooper, G. (1996). A self-help approach for high-risk drinking: Effect of an initial assessment. Journal of Consulting and Clinical
Psychology, 64(4): 694-700.
Sanchez-Craig, M., Spivak, K. & Davila, R. (1991). Superior outcome of females over males after brief treatment for reduction of heavy drinking: Replication and report of
therapist effects. British Journal of Addiction, 86(7): 867-876.
Sannibale, C. (1989). A prospective study of treatment outcome with a group of male problem drinkers. Journal of Studies on Alcohol, 50(3): 236-244.
Sass, H., Soyka, M., Mann, K., Zieglgänsberger, W. (1996). Relapse prevention by acamprosate: Results from a placebo-controlled study on alcohol dependence. Archives
of General Psychiatry, 53(8): 673-680.
Schottenfeld, R.S., Pakes, J.R., Oliveto, A., Ziedonis, D., & Kosten, T.R. (1997). Buprenorphine vs. methadone maintenance treatment for concurrent opioid dependence
and cocaine abuse. Archives of General Psychiatry, 54(8):713-20.
Scott, E. & Anderson, P. (1990). Randomized controlled trial of general practitioner intervention in women with excessive alcohol consumption. Drug and Alcohol Review,
10: 313-321.
Shea, M.T., Elkin, I., Imber, S.D., Sotsky, S.M., Watkins, J.T., Collins, J.F., et al. (1992) Course of depressive symptoms over follow-up: Findings from the National Institute
of Mental Health Treatment of Depression Collaborative Research Program. Archives of General Psychiatry, 49:782-787.
Shern, D.L., Tsemberis, S., Anthony, W., Lovell, A.M., Richmond, L., Felton, C.J., et al. (2000). Serving street-dwelling individuals with psychiatric disabilities: Outcomes of a
psychiatric rehabilitation clinical trial. American Journal of Public Health, 90(12): 1873-1878.
Silverman, K., Higgins, S.T., Brooner, R.K., Montoya, I.D., Cone, E.J. & Schuster, C.R. (1996). Sustained cocaine abstinence in methadone maintenance patients through
voucher-based reinforcement therapy. Archives of General Psychiatry, 53(5): 409-415.
Silverman, K., Wong, C.J., Umbricht-Schneiter, A., Montoya, I.D., Schuster, C.R. & Preston, K.L. (1998). Broad beneficial effects of cocaine abstinence reinforcement
among methadone patients. Journal of Consulting and Clinical Psychology, 66(5): 811-824.
Simons, A.D., Murphy, G.E., Levine, J.L., & Wetzel, R.D. (1986). Cognitive therapy and pharmacotherapy for depression: Sustained improvement over one year. Archives of
General Psychiatry, 43(1): 43-48.
Sisson, R.W. & Azrin, N.H. (1986). Family-member involvement to initiate and promote treatment of problem drinkers. Journal of Behavior Therapy and Experimental
Psychiatry, 17(1): 15-21.

20

Exhibit A.5: Citations of Studies Used in the Meta-Analysis, continued
Smith, J.E., Meyers, R.J. & Delaney, H.D. (1998). The community reinforcement approach with homeless alcohol-dependent individuals. Journal of Consulting and Clinical
Psychology, 66: 541-548.
Smith, T.L., Volpe, F.R., Hashima, J.N., & Schukit, M.A. (1999). Impact of a stimulant-focused enhanced program on the outcome of alcohol- and/or stimulant-dependent
men. Alcoholism: Clinical and Experimental Research, 23(11): 1772-1779.
Stein, D.J., Ipser, J.C., & Seedat, S. (2006). Pharmacotherapy for post traumatic stress disorder (PTSD). The Cochrane Database of Systematic Reviews, 1: CD002795.
Stein, D.J., Ipser, J.C., & van Balkom, A.J. (2000). Pharmacotherapy for social anxiety disorder. The Cochrane Database of Systematic Reviews, 4: CD001206.
Stephens, R.S., Roffman, R.A., & Curtin, L. (2000). Comparison of extended versus brief treatments for marijuana use. Journal of Consulting and Clinical Psychology, 68(5):
898–908.
Strain, E.C., Stitzer, M.L., Liebson, I.A., & Bigelow, G.E. (1993). Dose-response effects of methadone in the treatment of opioid dependence. Annals of Internal Medicine,
119: 23-27.
Test, M.A. & Stein, L.I. (1980). Alternative to mental hospital treatment: III. Social cost. Archives of General Psychiatry, 37(4): 409-412.
Test, M.A., Knoedler, W.H., Allness, D.J., et al. (1991). Long-term community care through an assertive continuous treatment team. In C.T. Schultz (Ed.), Advances in
neuropsychiatry and psychopharmacology: schizophrenia research, Vol. 1 (pp.239-246). New York: Raven.
Tharyan, P. & Adams, C.E. (2005). Electroconvulsive therapy for schizophrenia. The Cochrane Database of Systematic Reviews, 2: CD000076.
Thornley, B., Rathbone, J., Adams, C.E., & Awad, G. (2003). Chlorpromazine versus placebo for schizophrenia. The Cochrane Database of Systematic Reviews, 2:
CD000284.
Tomson, Y., Romelsjo, A. & Aberg, H. (1998). Excessive drinking--brief intervention by a primary health care nurse. Scandinavian Journal of Primary Health Care, 16(3):
188-192.
Tuunainen, A., Kripke, D.F., & Endo, T. (2004). Light therapy for non-seasonal depression. The Cochrane Database of Systematic Reviews, 2: CD004050.
Vanichseni, S., Wongsuwan, B., Choopanya, K., & Wongpanich, K. (1991). A controlled trial of methadone maintenance in a population of intravenous drug users in
Bangkok: Implications for prevention of HIV. International Journal of the Addictions, 26(12): 1313-20.
Volpicelli, J.R., Alterman, A.I., Hayashida, M., & OBrien, C.P. (1992). Naltrexone in the treatment of alcohol dependence. Archives of General Psychiatry, 49(11): 876-880.
Wallace, P., Cutler, S. & Haines, A. (1988). Randomised controlled trial of general practitioner intervention in patients with excessive alcohol consumption. British Medical
Journal, 297(6649): 663-668.
Ward, E., King, M., Lloyd, M., Bower, P., Sibbald, B., Farrelly, S., et al. (2000). Randomised controlled trial of non-directive counseling, cognitive-behaviour therapy, and
usual general practitioner care for patients with depression I: Clinical effectiveness. British Medical Journal, 321: 1383-1388.
Weiss, R.D., Griffin, M.L., Greenfield, S.F., Najavits, L.M., Wyner, D., Soto, J.A., et al. (2000). Group therapy for patients with bipolar disorder and substance dependence:
Results of a pilot study. Journal of Clinical Psychiatry, 61(5): 361-367.
White, J., Keenan, M., & Brooks, N. (1992). Stress control: A controlled comparative investigation of large group therapy for generalized anxiety disorder. Behavioural
Psychotherapy, 20: 97-114.
Whitworth, A. B., Fischer, F., Lesch, O. M., Nimmerrichter, A., Oberbauer, H., Platz, T., et al. (1996). Comparison of acamprosate and placebo in long-term treatment of
alcohol dependence. Lancet, 347(9013): 1438-1442.
Wijkstra, J., Lijmer, J., Balk, F., Geddes, J., & Nolen, W.A. (2005). Pharmacological treatment for psychotic depression. The Cochrane Database of Systematic Reviews, 4:
CD004044.
Wilson, D., Tien, G., & Eaves, D. (1995). Increasing the community tenure of mentally disordered offenders: An assertive case management program. International Journal
of Law and Psychiatry, 18: 61-69.
Winters, J., OFarrell, T.J., Fals-Stewart, W., Birchler, G.R. & Kelley, M.L. (2002). Behavioral couples therapy for female substance-abusing patients: Effects on substance
use and relationship adjustment. Journal of Consulting and Clinical Psychology, 70(2): 344-355.
Wood, K., & Anderson, J. (1995). The effect on hospital admissions of psychiatric case management involving general practitioners: preliminary results. Australian and New
Zealand Journal of Psychiatry, 29(2): 223-229.
Woody, G.E., McLellan, A.T., Luborsky, L. & OBrien, C.P. (1995). Psychotherapy in community methadone programs: A validation study. American Journal of Psychiatry,
152(9): 1302-1308.
*Ziguras, S.J. & Stuart, G.W. (2000). A meta-analysis of the effectiveness of mental health case management over 20 years. Psychiatric Services, 51(11): 1410-1421.

References marked with an asterisk (*) were not meta-analyzed but provided references to many studies used in this analysis.

21

Appendix B: Methods and Parameters to Model the
Benefits and Costs of Evidence-Based Treatment
To estimate the benefits and costs of evidence-based
treatment (EBT) for people with alcohol, drug, and mental
illness disorders, we developed an economic model.
Appendix B describes the technical structure of the model as
well as the data used with the model to produce the estimates
for this study.
The basic model takes the following form (each of the
elements in the model is discussed in greater detail in this
Appendix):

(

)

3 Y MTE ∗ E + HP + HC + D + C − PC
ty
ty
ty
ty
ty
ty
ty
∗ Nt
y
+
disrate
(
1
)
t =1 y

B(1): T = ∑∑

B(2): N t = WAPOP ∗ 12MOPREVt ∗ (1 - TX t ) ∗ POTENTIALt
Equation B(1) is the basic model developed to estimate the
total net present value of EBT, notated as T. We estimate
three prototype EBT programs: one for people with alcohol
disorders, one for people with drug disorders, and one for
people with mental health disorders (we account for “comorbidities” in our prevalence estimates, as discussed below).
In the equation, the three prototype EBT programs are notated
with a t.
For each program, we estimate a series of annual cash flows
that run from y to Y, where y represents the years following
participation in an EBT. The subscript y equals 1 during the
year that a person is administered an EBT and ends in Y—the
maximum number of years over which effects are estimated.
The model computes the marginal treatment effect, MTEty, for
each of the three prototype EBTs in each year following
treatment (the computation of MTEty is described later in this
Appendix). As we discuss, we project these treatment effects
to decay over time. The marginal effects are multiplied by the
sum of five sources of benefits estimated in this study. These
are: the value of economic production due to improvements in
disorder-caused morbidity, Ety; the value of household
production due to improvements in disorder-caused morbidity,
HPty; the value of reduced health care and other costs due to
reduced disorder rates, HCty; the value of economic and
household production due to reductions in disorder-caused
mortality, Dty; and the value of avoided disorder-caused crime,
Cty. Each of these factors is described in this Appendix.
Annual program costs, PCty, are subtracted from the annual
benefits. The annual net cash flows are then discounted to
present value with a discount rate, disrate. The present-valued
dollars are thus based in the year in which the initial
investment in an EBT would be made.
A benefit-to-cost ratio, BCt, is computed for each prototype
EBT by rearranging equation B(1):

(

Y MTE ∗ E + HP + HC + D + C
ty
ty
ty
ty
ty
ty
y
(
1
disrate
)
+
B(3): BCt = y
Y
PCty
y
y (1 + disrate)

∑

∑

22

)

Additionally, an internal rate of return can be computed for
each EBT by using Microsoft Excel’s IRR function for the
annual cash flows, CFty, given by:

(

)

B(4): CFty = MTEty ∗ Ety + HPty + HCty + Dty + Cty − PCty
Finally, to calculate the total net benefits for Washington,
equation B(1) multiplies the per-person net present value for
each prototype EBT by the number of people in Washington
estimated to be in need of treatment, Nt. The computation of
variable Nt is given in equation B(2) and is the product of the
total number of people currently living in Washington in the
age groups selected to be eligible for an EBT, WAPOP; times
the 12-month prevalence of the disorder in the population,
12MOPREVt; times one minus the percent of people already
treated with an EBT, TXt; times an assumption about the
percentage of the remaining people in Washington with the
disorder who might realistically be treated, POTENTIALt.
Exhibits B.1, B.2, and B.3 display a list of the parameters used
in our analytical approach; the following description refers to
the information in those Exhibits.

B1. General Model Parameters
The model uses a number of parameters pertinent to all three
evidence-based prototypes estimated in this study. Exhibit
B.1 lists these parameters.
The range of discount rates used in this study is shown on line
1 of Exhibit B.1. The high end of the range is a 7 percent real
discount rate. This discount rate reflects the rate that has
been recommended by the federal Office of Management and
Budget.39 The low end of the range is a 3 percent real
discount rate used by the Congressional Budget Office in a
variety of analyses including its projections of the long-term
financial position of Social Security.40 Our study uses a
medium discount of 5 percent, the difference between the high
and low rates.41
Some of the estimated benefits in this study reflect the effect
of improvements in the Diagnostic and Statistical Manual of
Mental Disorders (DSM) alcohol, drug, and mental illness
disorders on economic outcomes. Key parameters in these
projections are the level of earnings and the long-term
expected rate of real (inflation-adjusted) growth in earnings.
The level of earnings by age is taken from cross-sectional data
from the 2005 March Supplement to the Current Population
Survey (CPS), with data on earnings during 2004. The
earnings are those for people with education levels between
9th grade through some college. The number of non-earners
is included in the estimates so that the average earning level
reflects earnings of all people at each age (earners and non42
earners). The cross-sectional estimates from the CPS are
shown on Exhibit B.2 by age of person.

39

Office of Management and Budget, Circular A-94 (revised 1992).
See Congressional Budget Office report: http://www.cbo.gov/ftpdocs/
72xx/doc7289/06-14-LongTermProjections.pdf
41
For a general discussion of discount rates for applied public benefitcost analyses, see: C. Bazelon, and K. Smetters. (1999). Discounting
inside the Washington D.C. Beltway. Journal of Economic Perspectives,
13(4): 213-28. See also: H. Kohyama. (2006). Selecting discount rates
for budgetary purposes, Briefing Paper No. 29.
http://www.law.harvard.edu/faculty/hjackson/DiscountRates_29.pdf
42
Current Population Survey data downloaded from the US Census
Bureau site with the DataFerrett extraction utility:
http://www.bls.census.gov/cps/cpsmain.htm
40

Exhibit B.1
The Benefits and Costs of Evidence-Based Treatment:
General Model Parameters
Parameter
Line
number
1
2
3
4
5
6
7
8
9
10
11

High
Discount Rate
Real annual rate of growth in earnings
Fringe benefit multiple for earnings
Tax rate for earnings
Real annual rate of growth in health care costs
Year of dollars for the analysis
Year of dollars for the Current Population Survey used in the study
Real cost of capital (used in the crime sub-model)
Hours per week for household production, per person
Dollars per hour for household production
Year of dollars for the household production

.070
.023
.044
-

Medium
.050
.013
1.423
.316
.034
2006
2004
.025
19.5
$10.08
2004

Low
.030
.003
.024
-

Line 2 of Exhibit B.1 shows the long-run expected growth rate
in real earnings. The estimate for the medium case is taken
from the Congressional Budget Office (CBO) analysis of long43
run Social Security. We included the higher rate of growth
and the lower rate of growth in our sensitivity analyses,
described below.

household maintenance. These estimates are quite close to
49
the 21.4 hours per week calculated by Douglass et al. The
average shadow wage rate, shown on line 10 of Exhibit B.1,
for these three household services was taken from United
State Bureau of Labor Statistics data on average wage rates
50
in Washington in 2004 for each service.

Line 3 of Exhibit B.1 shows an estimate for the average fringe
benefit rate for earnings. This estimate is from the
Employment Cost Index as computed by the United States
Bureau of Labor Statistics.44 Line 4 shows the average tax
45
rate applied to earnings.

B2. Program Effectiveness Parameters

Line 5 shows our assumed rate of growth in real health care
costs. The medium case is taken from the current forecast for
2006 to 2015 from the US Department of Health and Human
Services.46 For high and low cases, we assumed one
percentage point above and below the medium rate.
Line 6 of Exhibit B.1 indicates the year chosen for the overall
analysis. All costs are converted to this year’s dollars with the
inflation index shown in Exhibit B.2. The inflation index is
taken from the Washington State Economic and Revenue
Forecast Council, the official forecasting agency for
Washington State government. The index is the chain-weight
47
implicit price deflator for personal consumption expenditures.
Lines 9 through 11 of Exhibit B.1 indicate the estimates used
to monetize the value of household production, a common
procedure in cost-of-illness studies.48 We estimate 19.5 hours
per week for household production. This estimate is based on
an assumed 1.5 hours per day for housekeeping services, 1.0
hours per day for food preparation, and 2.0 hours per week for
43

See Congressional Budget Office data for the June 2006 report, Table
W-5, at: http://www.cbo.gov/ftpdocs/72xx/doc7289/06-14SupplementalData.xls
44
United State Bureau of Labor Statistics, Employment Cost Index, March
14, 2006 release, data for December 2005:
http://www.bls.gov/news.release/ecec.toc.htm
45
Tax Foundation Special Report, April 2006, Table 1, page 4:
http://www.taxfoundation.org/files/sr140.pdf
46
US Department of Health and Human Services, Office of the Actuary in
the Centers for Medicare & Medicaid Services. National Health Care
Expenditures Projections: 2005-2015. http://www.cms.hhs.gov/National
HealthExpendData/downloads/proj2005.pdf
47
Washington State Economic and Revenue Forecast Council:
http://www.erfc.wa.gov/pubs/feb06pub.pdf
48
See, for example, W. Max, D. Rice, H. Sung, and M. Michel. (2004).
Valuing human life: Estimating the present value of lifetime earnings, 2000.
Center for Tobacco Control Research and Education. Economic Studies
and Related Methods. Paper PVLE2000.
http://repositories.cdlib.org/cgi/viewcontent.cgi?article=1049&context=ctcre

The first section of Exhibit B.3 lists the estimates we produced
for the average effectiveness of EBT for persons with serious
alcohol, illicit drug, and mental illness disorders. These
results, shown on lines 1 through 3 of Exhibit B.3, are the
meta-analytic results discussed in Appendix B. Line 1 is the
unadjusted weighted effect size of EBT for each of the three
types of disorders. Line 2 is the associated standard error
from the meta-analysis. Line 3 is the adjusted effect size after
applying the Institute rules, described in Appendix A3, to
account for the methodological quality of the evidence,
outcome measurement relevance, and the degree of
researcher involvement.
Line 4 is an estimated standard error for the Institute-adjusted
effect size. A standard error is computed for this parameter
because it is used in sensitivity analyses (discussed in
Appendix B12). Since we cannot estimate a standard error
directly for the Institute-adjusted effect size, we employ a
simple procedure to calculate a standard error for the Instituteadjusted effect size:
B(5): AdjustedSE =

AdjustedES
⎛ UnadjustedES ⎞
⎟⎟
⎜⎜
⎝ UnadjustedSE ⎠

In this formula, we compute an estimated standard error for
the Institute-adjusted effect size by dividing the Instituteadjusted effect size by the t-statistic for the unadjusted effect
size (from the meta-analyses). This means we are assuming
the same level of statistical significance for our adjusted effect
size as that obtained from the unadjusted meta-analysis as
described in Appendix A.

49

J. Douglass, G. Kenney, and T. Miller. (1990). Which estimates of
household production are best? Journal of Forensic Economics, 4(1): 25-45.
50
US Bureau of Labor Statistics, November 2004 Washington Wage Data
from: http://www.bls.gov/oes/current/oes_wa.htm#b39-0000

23

Exhibit B.2
The Benefits and Costs of Evidence-Based Treatment:
Annual Data Series

1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006

Inflation
Index
0.521
0.567
0.598
0.624
0.648
0.669
0.686
0.709
0.737
0.770
0.805
0.834
0.858
0.878
0.896
0.916
0.935
0.951
0.960
0.976
1.000
1.021
1.035
1.055
1.082
1.113
1.137

Age
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80

Washington
Population, 2006
90,790
90,133
92,505
92,067
91,768
92,829
91,519
90,951
89,859
84,783
83,095
82,259
81,105
83,687
79,971
82,154
88,366
94,869
98,231
90,956
88,589
86,958
89,355
97,011
97,353
98,843
99,616
100,711
102,892
97,464
100,774
98,177
96,511
97,627
92,805
92,303
87,140
84,198
85,219
79,737
81,019
79,625
60,323
60,948
59,924
58,056
50,275
49,947
49,296
48,336
47,086
45,567
43,810
41,846
39,708
37,434
35,059
32,620
30,153
27,690
25,262
22,897
20,617

Average annual
earnings,
workers and nonworkers, United
States
$3,174
$5,741
$7,972
$10,316
$11,527
$14,325
$15,325
$18,032
$18,144
$19,968
$20,505
$22,468
$22,530
$24,514
$23,978
$22,431
$23,354
$25,804
$27,221
$26,220
$26,894
$27,028
$27,636
$27,153
$27,214
$28,534
$28,222
$28,414
$27,974
$27,794
$28,189
$28,038
$27,896
$27,865
$28,098
$25,713
$26,649
$26,356
$23,163
$25,921
$21,941
$22,215
$23,097
$19,166
$17,390
$12,120
$11,068
$8,034
$8,775
$6,869
$7,039
$5,633
$6,577
$6,375
$3,867
$2,838
$2,027
$3,492
$2,285
$1,104
$1,844
$1,601
$976

Total number of
people in
households,
United States
3,809,016
3,464,472
3,659,116
3,612,517
3,794,167
3,749,240
3,888,289
3,844,850
3,609,340
3,684,725
3,659,564
3,788,098
3,651,021
3,629,443
3,701,149
3,974,746
4,336,910
4,124,783
3,904,503
3,856,313
4,028,587
4,007,543
4,565,264
4,329,129
4,390,913
4,310,340
4,500,372
4,679,133
4,199,705
4,509,734
4,237,686
4,189,064
4,363,843
3,964,673
3,889,799
3,521,706
3,710,336
3,574,332
3,450,510
3,543,593
3,377,117
2,792,955
2,814,165
2,640,818
2,718,679
2,320,776
2,269,077
2,391,316
2,086,775
1,987,848
1,845,228
1,833,058
1,668,781
1,697,679
1,683,738
1,593,615
1,642,561
1,622,661
1,544,163
1,700,186
1,432,898
1,295,198
5,383,474

Total number of
people in group
quarters, United
States
730
4,480
6,681
552
284
0
0
0
4,165
0
4,442
4,190
0
7,278
0
0
0
3,056
1,149
4,190
0
4,190
0
8,617
0
6,824
1,036
4,172
4,020
0
9,286
4,561
0
4,328
1,209
3,614
0
7,151
0
0
3,855
0
0
0
1,063
0
5,472
0
0
2,698
3,407
0
2,525
0
1,444
1,444
1,444
0
0
0
0
2,345
7,235

Total number of
people in family
households,
United States
3,596,957
2,987,935
2,872,904
2,773,204
2,779,428
2,706,395
2,816,151
2,723,761
2,711,579
2,803,019
2,912,597
2,955,566
2,935,608
3,055,730
3,081,560
3,359,017
3,633,578
3,473,568
3,335,370
3,247,817
3,408,447
3,494,064
3,871,119
3,761,068
3,788,287
3,678,303
3,852,283
3,966,309
3,531,316
3,787,472
3,487,046
3,469,515
3,557,464
3,291,757
3,146,486
2,838,386
3,057,872
2,913,325
2,775,454
2,808,089
2,735,219
2,245,174
2,192,840
2,084,095
2,145,492
1,825,997
1,803,933
1,846,406
1,638,184
1,541,097
1,419,601
1,368,879
1,193,160
1,288,380
1,206,481
1,143,833
1,136,078
1,058,835
1,043,494
1,218,259
941,825
762,039
3,095,700

Probability of
shifting
household
production costs
upon disability or
death
0.945
0.864
0.787
0.768
0.733
0.722
0.724
0.708
0.752
0.761
0.797
0.781
0.804
0.844
0.833
0.845
0.838
0.843
0.854
0.843
0.846
0.873
0.848
0.871
0.863
0.855
0.856
0.848
0.842
0.840
0.825
0.829
0.815
0.831
0.809
0.807
0.824
0.817
0.804
0.792
0.811
0.804
0.779
0.789
0.789
0.787
0.797
0.772
0.785
0.776
0.771
0.747
0.716
0.759
0.717
0.718
0.692
0.653
0.676
0.717
0.657
0.589
0.576

The inflation index is implicit price deflator for personal consumption expenditures. The Washington population numbers are from the Washington State Office of Financial
Management. The average earnings data are for workers and non-workers and are from the 2005 Current Population Survey from the US Census Bureau. The household data are
from the same CPS.

24

Line number

Exhibit B.3
The Benefits and Costs of Evidence-Based Treatment:
Program-Specific Model Parameters
Evidence-Based Treatment: Adults With
Alcohol, Drug, or Mental Illness Disorders

See text for information about these parameters

Adults with a
serious DSM
alcohol disorder

Adults with a
serious DSM drug
disorder

Adults with a
serious DSM mental
illness disorder

Program Effectiveness Parameters
1
2
3
4
5
6
7
8
9

Unadjusted effect size from the meta analyses (program effect on disordered outcome)
Standard error for the unadjusted effect size from the meta analyses
Adjusted effect size after applying WSIPP* adjustments
Estimated standard error for the WSIPP*-adjusted effect size
Expected annual rate of decay in effect size
Standard error
Expected diminishing returns to effect size with large scale ramp up
(lower expected rate of decay)
(higher expected rate of decay)

-.312
.027
-.247
.021
-.062
.027
.750
1.000
.500

-.451
.044
-.355
.035
-.164
.072
.750
1.000
.500

-.402
.052
-.360
.058
-.176
.089
.750
1.000
.500

39.9
13.4
18
65
$2,300
$500
2005
.000
1.0
75%

36.4
13.4
18
65
$2,300
$500
2005
.000
1.0
75%

40.4
13.4
18
65
$3,596
$782
1992
.000
1.0
75%

15.69%
5.55%
0.26%

2.94%
2.05%
0.16%

6.36%
3.80%
0.22%

11.1%
0.4%
50%
75%
25%
4,145,297
230,087
204,435
102,218

14.7%
0.9%
50%
75%
25%
4,145,297
84,955
72,497
36,248

46.2%
3.5%
50%
75%
25%
4,145,297
157,521
84,746
42,373

80
3
3
2
1
99

80
4
45
11
0
0

80
2
8
2
43
0

1992
2,125,554
333,598
107,360
0.32

2000
2,362,000
69,502
23,544
0.34

1992
2,125,554
135,189
32,381
0.24

1
-0.260
0.061
21,356
76.0%
29,715
-15.6%

1
-0.262
0.059
21,356
76.0%
29,715
-15.7%

1
-0.250
0.038
21,356
76.0%
29,715
-15.0%

$44.1
1998
204,426,000
15,127,524
$4,496
10.0%
43.2%
11.2%
45.6%

$15.7
2002
215,127,000
3,226,905
$6,114
10.0%
59.0%
12.7%
28.3%

$46.2
1992
185,473,000
7,047,974
$13,799
10.0%
79.1%
-7.4%
28.3%

0.9194
-0.0228
-0.0009
0.0000
30

0.6108
-0.0601
0.0023
0.0000
30

0.5861
-0.0177
0.0004
0.0000
30

Program Design Parameters
10
11
12
13
14
15
16
17
18
19

Average age of program participant
Standard deviation of age of program participant
Minimum age of program participant
Maximum age of program participant
Average annual program cost
SD of average program cost
Year of program cost estimate
Annual real rate of escalation in program costs
Average number of years of treatment episode, per average participant
Percent of program costs paid by taxpayer

Prevalence Parameters
20
21
22

Lifetime prevalence of DSM disorder in this population cohort
Current (12-mo) prevalence of DSM disorder in this population cohort
Standard error

Total Potential Population to Be Treated
23
24
25
26
27
28
29
30
31

Proportion of target population already treated with evidence-based program
Standard error
Proportion of the currently unserved target population that might realistically be served
high
low
Total current Washington population (in the age group of those to be treated)
Those currently with the DSM disorder
Market potential: the number not already being treated with evidence-based treatment
Realistic market potential: the number realistically available for evidence-based treatment

Mortality Parameters (age of death for person with disorder)
32
33
34
35
36
37
38
39
40
41
42
43

Maximum Age for Death (Normal life expectancy for control group)
Distribution type for probability density
Probability distribution: Parameter 1
Probability distribution: Parameter 2
Probability distribution: Parameter 3
Probability distribution: Parameter 4
Attributed Death Factor (Of those with disorder, prob death is caused by the disorder)
Year of analysis
Total deaths in year of analysis, United States
Of the deaths that year, the number that had (ever in lifetime) a DSM condition
Deaths due to disorder in the year, United States
Probability of a lifetime disorder AND that the death was due to the disorder

Morbidity Parameters (earnings and household production)
44
45
46
47
48
49
50

Effect size applies to: 1 (employment rate), or 2 (earnings of earners)
Unadjusted ES: Economic outcomes (either employment or earnings) Earnings =f(Disorder)
Standard error
Average earnings (CPS 2004) includes non-earners
Percent with earnings (CPS 2004)
Standard deviation of average earnings (CPS 2004) earners only
Percent change to average earnings, from the disorder

Health Care Costs
51
52
53
54
55
56
57
58
59

Total cost (billions), United States
Year of estimate
Adult population for year of estimate, United States
Current (12-month) number of people with a DSM disorder
Annual cost per current abuser (adjusted to base year for real growth in costs)
Assumed percentage (plus and minus) from the average cost
Percent of costs paid by taxpayer
Percent of costs paid by participant
Percent of costs paid by other private payers

Natural Rate of Recovery Parameters
Constant
60
61
Time
Time^2
62
Time^3
63
64
Cutoff age
* Washington State Institute for Public Policy

25

Exhibit B.3 (Continued)
The Benefits and Costs of Evidence-Based Treatment:
Program-Specific Model Parameters
Line number

Evidence-Based Treatment: Adults With
Alcohol, Drug, or Mental Illness Disorders
Adults with serious Adults with serious Adults with serious
DSM Alcohol
DSM Drug Disorder DSM Mental Illness
Disorder
Disorder

Crime Parameters
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80

Effect Size: Crime outcomes as a function of the disorder, from meta analysis
Standard error
Minimum age for crime distributions
Maximum age for crime distributions, =Y
Maximum age for observed crime parameters, =X
Scaleup =1 total convictions, Scaleup=2 for felony convictions
Scaleup: estimated difference in crime at age X to age Y (=X/Y)
Out of population, total percent with a crime event by age X
Of those with a crime event, the average number of events per person at age X
Of the total population, the average number of events per person at age X
Murder offenses for this population at age X
Sex offenses for this population at age X
Robbery offenses for this population at age X
Aggravated assault offenses for this population at age X
Property offenses for this population at age X
Drug offenses for this population at age X

Line 5 of Exhibit B.3 lists one of several conservative
assumptions we use in this analysis. It displays the annual
rate of decay that we assume for the effect size shown on Line
3. For the most part, the effect size on line 3 reflects the
results from a meta-analysis of individual program evaluation
studies that usually have fairly short-term follow-up periods. In
this benefit-cost analysis, on the other hand, we estimate the
long-run benefits of EBT based on these short-term effect
sizes. It can be argued that these short-term effect sizes will
decay over time; that is, the effect that is observed after one
year may not persist five or ten years into the future. The
purpose of the estimate on line 5 is to provide a way to model
the uncertainty of this potential decay. This assumed rate of
decay is an important factor that determines range of
uncertainty in our overall estimates. We found that effects of
treatment were eroded by half in 11 years for alcohol
disorders, four years for disordered drug use, and 3.5 years
for mental illness.
For each of the three classes of treatment (alcohol, drugs, and
mental illness), we estimate a mean annual rate of decay and
a standard error for the mean. These two parameters are then
used in sensitivity analyses. We do this by using data from
those studies in our analysis where the follow-up period is
noted.51 For each broad treatment type, our regression
analysis uses up to seven different functional forms to examine
how length of the follow-up period influences the observed
effect size. The model with the best adjusted R-square value
(that is, the best fit) is chosen for each class of treatment.

51

Not all studies clearly stated the follow-up period. Our analysis
included 91 studies on alcohol treatment and 40 studies on drug
treatments. We estimated a single rate of decay for treatment of mental
illness. Because follow-up times were often very short for mental illness,
we limited this analysis to those treatments with numerous studies of
varying follow-up times: chlorpromazine, Assertive Community
Treatment, and non-drug therapies for depression and anxiety disorders.
For some mental health treatments, where we relied on Cochrane
reviews, we coded an effect size and follow-up period for each study in
the review. Our analysis of effect size decay included 84 studies on
mental illness.

26

.192
.099
10
80
32
1
66.5%
15.4%
2.65
.41
223
1224
1133
2766
14910
6019

3.140
.000
10
80
80
1
100.0%
100.0%
2.00
2.00
0
0
0
0
0
6019

.392
.046
10
80
32
1
66.5%
15.4%
2.65
.41
223
1224
1133
2766
14910
6019

Lines 7 through 9 on Exhibit B.3 describe another set of
conservative assumptions we employ. The purpose of this
study is to estimate the aggregate benefits and costs of EBT
for a relatively large percentage of people with alcohol, drug,
or mental health disorders in Washington. The effect size that
we estimate on line 3, however, is derived mostly from
individual studies of much smaller populations. Because of
self-selection and diminishing returns, it can be conjectured
that the average treatment effect obtained from these studies
of more serious populations will not be as great if EBT
programs were extended to a wider group of people with
clinical disorders. It can also be argued that, as programs get
larger, it becomes more difficult to maintain quality control and,
therefore, a larger-scale program would yield reduced effects
compared with those obtained from smaller programs. Thus,
the assumptions employed on lines 7 through 9 provide a
means to model this uncertainty. The assumptions are
multiplicative factors that we apply to the adjusted and
decayed effect sizes. For example, the base case assumption
shown on line 7—a factor of .75—means that we assume the
average treatment effect will only be 75 percent as large if the
program were to be implemented on a large scale. In the
sensitivity analyses, we allow this assumption to vary by the
higher and lower assumptions shown on lines 8 and 9.

B3. Program Design Parameters
The second section in Exhibit B.3 lists two of the parameters
we use to describe the generic EBT programs. The first set of
parameters, lines 10 through 13, describes the age groups that
might be eligible for the three prototype programs. These
parameters are used in estimating the total size of the potential
treatment populations as well as in the calculation of the
estimated benefits. Using a normal distribution with a mean
age (line 10) and standard deviation (line 11), and bounding
the distribution by the minimum age (line 12) and maximum
age (line 13), a density distribution P is estimated for the
probability of program participation, such that,
B(6): 1 =

max

∑ Pp ,

p = min

where the distribution P is defined to be normally distributed
with a mean age and its standard deviation.
Lines 14 through 19 list the assumptions we made about the
cost of EBT programs. These include estimates of the
average cost per treatment episode, assumptions regarding
the standard deviation for these average costs, and the extent
to which EBT programs would be financed by tax dollars.
Rather than costing-out each of the individual EBT programs
examined, we assumed that EBT is the norm for those
currently receiving services. Therefore, the observed average
cost per treatment episode is a reasonable approximation of
the average cost per episode of an average EBT program. Of
course, to the extent the current practices do not represent
evidence-based approaches, we may be under-estimating the
cost of EBT programs.
The average costs of EBT for alcohol, drug, and mental health
are derived from two sources. According to one recent report,
the average cost of EBT for alcohol or drug abuse in Washington
State was $2,300 per episode in 2002.52 The report did not
provide separate estimates for alcohol and drug treatment,
therefore, the same figure is used for both program areas.
A similar episode-based cost estimate for treatment of serious
mental illness was not available for Washington State.
Fortunately, the same study that we used to describe health
care and other costs attributable to mental illness also
provided an estimate of mental health treatment costs, which
in 1992 dollars, averaged $3,596 per episode.53 Updated to
current dollars, we assume this to be the cost of EBT for
serious mental health disorders.

B4. Prevalence Parameters
To determine the size of the population in Washington that has a
serious disorder that could be addressed with one of the three
prototype EBT programs, we reviewed the national literature on
the prevalence of the disorders in the general population. There
have been several national studies conducted in the last 20
years to estimate the lifetime and current prevalence of serious
alcohol, drug, and mental illness in the general population.
Lines 20 to 22 show the estimates from our reading of the
national literature. Line 20 shows the estimated lifetime
probability of having one of the disorders. This parameter is
used when we model the mortality effects of the disorders,
described in Appendix B9. For alcohol and drug dependence,
54
we use the lifetime prevalence rates listed in Harwood et al.
The Harwood lifetime rates were taken from their analysis of the
National Longitudinal Alcohol Epidemiologic Survey for adults
ages 18 to 64. Harwood reports lifetime prevalence rates for
males and females; we combine them into an overall average
using 1992 census data on the ratio of males to females in the
18 to 64 age group.
The estimate we used for a lifetime prevalence of serious
mental illness, shown on line 20, was derived in the following
manner. Harwood et al. (2000) provided an estimate of the
55
12-month prevalence of serious mental illness for males and
females at .03 and .046, respectively, for an average rate of
.038. This number accounts for comorbidity, that is, persons
52

T.M. Wickizer, A. Krupski, K. Stark, D. Mancuso, and K. Campbell.
(in press). The Effect of Substance Abuse Treatment on Medicaid
Expenditures among General Assistance Welfare Clients in
Washington State. Milbank Quarterly.
53
Harwood et al., The economic costs of mental illness, page 3-6.
54
Harwood et al., The economic costs of alcohol and drug abuse.
55
Harwood et al., The economic costs of mental illness.

with more than one serious mental illness are counted only
once. We estimated lifetime prevalence by summing lifetime
prevalence rates reported for the National Comorbidity Survey
56
Replication for schizophrenia, bipolar disorders, and serious
forms of major depression and panic disorder. To account for
comorbidity, we then multiplied by the ratio of Harwood’s 12month prevalence to the sum of 12-month prevalences for
each of these disorders. Note that this rate on line 20 is for
severe diagnoses which the Harwood report defines to be
schizophrenia, non-affective psychosis, manic depressive
disorder, severe forms of major depression and panic
disorder.57 In our study, we confine our economic analyses to
these severe forms of mental disorders.
Line 21 of Exhibit B.3 is the estimate we use in this study of the
current (i.e. 12-month) prevalence of each disorder in the general
population. For serious alcohol and drug disorders, we use the
estimates provided in Narrow et al. (2002) which are based on
their interpretation of the clinical significance of findings from the
National Comorbidity Survey and the Epidemiologic Catchment
Area study.58 For serious mental illness disorders, we use the
59
estimate provided in Harwood et al.
We account for the comorbidity between drug and alcohol
dependency with the following calculations. Narrow et al.
report a total disorder rate for any alcohol or drug disorder of
7.6 percent for the 18- to 54-year-old age group. They also
report a 6.5 percent rate of alcohol disorders and a 2.4 percent
rate for other drug use disorders. To account for comorbidity
and avoid double counting people later in our analysis, we
estimate the unique alcohol disorder rate as 5.5 percent (.055
= 0.076*(6.5/(6.5+2.4))) and the unique other drug disorder
rate as 2.05 percent (.0205 = 0.076*(2.4/(6.5+2.4))). We
estimate the size of the standard errors with the number of
subjects in the National Comorbidity Survey (7,599). The
associated standard errors are used in sensitivity analyses.

B5. Total Potential Population to Be Treated
We estimated two additional factors to help focus the analysis
on the size of the population that could take advantage of the
prototype EBT programs. First, we estimate the size of the
disordered population already being treated with EBT
programs in Washington. These estimates are shown on lines
23 and 24 of Exhibit B.3. For people with serious alcohol
disorders and for those with serious illicit drug disorders, we
analyzed the public use data set for the National
Epidemiological Survey on Alcohol and Related Conditions
(NESCAR). Among people indicating alcohol dependence in
the past 12 months, we calculated the average percent and
standard deviation that had been treated professionally for
their alcohol disorder in the past 12 months. We used the
same procedure for those with dependence on illicit drugs.
For people with serious mental illness disorders, we relied on
estimates of treatment rates by Kessler et al. based on the
60
National Comorbidity Survey.
On lines 25 to 27 we also make additional restrictions on the
size of the population that might be treated with EBT
programs. It is never possible to completely saturate a
market, so we provide factors to estimate low, medium, and
56

R. Kessler et al. Prevalence, severity, and comorbidity of 12-month
DSM-IV disorders in the National Comorbidity Survey Replication.
57
Harwood et al., The economic costs of mental illness. Table 4.7.
58
Narrow et al., Revised prevalence estimates of mental disorders.
59
Harwood et al., The economic costs of mental illness. Table 4.7.
60
R. Kessler, P. Berglund, M. Bruce, J. Koch, E. Laska, P. Leaf, R.
Mandersheid, R. Rosenheck, E. Walters, and P. Wang. (2001). The
prevalence and correlates of untreated serious mental illness. Health
Services Research, 36(6): 987-1007.

27

high market penetration rates. These alternative rates are
used in the sensitivity analysis described in Appendix B12.

To compute the earnings effect of an incidence of a DSM
disorder, we begin with the following equation:

The factors described above are used to compute the total
size of the current population in Washington that: (a) has a
serious disorder, (b) is not currently being treated, and (c)
might be realistically treated with a prototype EBT. Line 28 of
Exhibit B.3 reports the size of the total population in
Washington between the minimum and maximum age groups
described on lines 12 and 13. The Washington population
estimate is taken from the Washington State Office of
Financial Management, and the actual population estimates
61
are shown in Exhibit B.2. To this figure, we then applied the
12-month prevalence estimate (from line 21) to produce line
29: the estimated total current number of people in
Washington with the disorder. Line 30 then subtracts the
estimated percentage of the population already being treated
with EBT programs (from line 23). Finally, line 31 applied the
base assumption about the realistic potential (from line 25) to
enroll disordered people in a prototype EBT.

B(8): Ea = EARNINGS a * FRINGE * INFLATION

B6. Morbidity Parameters and Methods
Prior studies of the costs of alcohol, drug, and mental illness
disorders have found that, among people with the disorders,
performance in the economic marketplace is reduced.62 To
provide an independent test of this hypothesis, we conducted
several meta-analyses. We sought to determine if existing
research studies indicate that either an individual’s level of
earnings conditional on employment, or the rate of employment
alone, was significantly related to the presence of having an
alcohol, drug, or mental illness disorder. We reviewed the
literature on the topics and used the meta-analytic methods
described in Appendix A to this report.
Exhibit B.4 summarizes the results of our meta-analyses. We
found that all three disorders are significantly related to the
probability of employment, but not earnings conditional on
employment. The effect sizes for employment from the metaanalyses are shown on line 45 of Exhibit B.3 and the
associated standard errors are listed on line 46. To find the
marginal effect of a disorder on average earnings levels (via
the effect on employment rates), we compute the following:
ES
⎞
⎛ AE
* sin(arcsin( ER ) +
)^2 − AE ⎟
⎜
ER
2
⎠
⎝
,
B(7): EEt =
AE
where EE is the estimated earnings effect for each of the
evidence-based treatments, t, and represents the percentage
change in average earnings; AE is the average earnings of
earners and non-earners taken as a whole (this estimate,
shown on line 47, comes from the CPS; ER is the employment
rate (shown on line 48 of Exhibit B.3, also from the CPS) and
ES is the effect size of the effect of having a disorder on ER
(shown on line 45, from the meta-analysis). Since the arcsine
transformation is used to compute the effect size in the metaanalyses, as described in Appendix A, that effect is reversed
here to return the unit change.

61

Washington State Office of Financial Management,
http://www.ofm.wa.gov/pop/default.asp
See: (a) Harwood, Updating estimates of the economic costs of alcohol
abuse in the United States, from Table 3; (b) Office of National Drug
Control Policy. The economic costs of drug abuse in the United States,
from Table III-1; and (c) Harwood et al., The economic costs of mental
illness, from Table 6.1.
62

28

For each age a, the total earnings of a person Ea is the annual
EARNINGS taken from the Current Population Survey for that
age, shown on Exhibit B.2, times the FRINGE benefit multiple,
shown on line 3 of Exhibit B.1, times the INFLATION
adjustment from Exhibit B.2 to inflate the CPS series
(denominated in 2004 dollars) to the year chosen for this
analysis (2006 dollars).
The annual cash flows of lost earnings associated with having
a disorder of type t is estimated with this process:
P

B(9): $ Ety = ∑ E p + y −1 *(1 + ER) y −1 * EEt * PPtp * −1
p

In this equation, $Ety is the annual cash flow of lost earnings
for a person with disorder type t in year y, where y is the
number of years following participation in an EBT. The
subscript y equals 1 during the year that a person is
administered an EBT.

Exhibit B.4
Meta-Analytic Estimates of Standardized Mean Difference Effect Sizes
Results Before Adjusting Effect Sizes
Number
of Effect
Sizes
Included
in the
Analysis

Fixed Effects
Model
Weighted Mean
Effect Size
& p-value
ES

Random Effects
Model

Homogeneity
Test

p-value

p-value

Weighted Mean
Effect Size
& p-value
ES

p-value

Employment =f(alcohol disorder)

11

-.183

0.000

0.000

-.239

Wages =f(alcohol disorder)

5

.004

0.701

0.124

na

Employment =f(DSM mental illness)

8

-.246

0.000

0.000

-.250

0.000

Wages =f(DSM mental illness)

7

-.140

0.000

0.000

-.213

0.000

Employment =f(drug disorder)

6

-.230

0.000

0.000

-.262

0.000

Wages =f(drug disorder)

1

.000

0.981

na

na

0.000
na

na

Crime =f(Mental Illness)

3

.337

0.000

0.001

.392

0.000

Crime =f(Alcohol Disorder)

3

.176

0.000

0.000

.192

0.053

Studies (complete citation on next page)

Used to estimate

Zuvekas, Cooper, & Buchmueller, 2005
Mullahy and Sindelar, 1993
Mullahy and Sindelar, 1996
Mullahy and Sindelar, 1997
Terza, 2002
Terza, (undated)
Chevrou-Severac and Jeanrenaud, 2002
Feng et al., 2001
Auld, 2002
MacDonald & Shields, 2004
Cook & Peters, 2005
Zuvekas, Cooper, & Buchmueller, 2005
Mullahy and Sindelar, 1993
Zarkin et al., 1998
Kenkel and Ribar, 1994
Bray, (2005)
Harwood et al., 2000
Ettner et al., 1997
Farahati et al., 2003
Savoca, 2000
Alexandre & French, 2001
Kessler et al., 1999
Hamilton et al., 1997
Chatterji et al., 2005
Ettner et al., 1997
Marcotte, 2003
Kessler & Frank, 1997
Frank & Gertler, 1991
Bartel & Taubman, 1986
French & Zarkin, 1998
Stewart et al., 2003
DeSimone, 2002
Buchmueller and Zuvekas, 1998
Zuvekas, Cooper, & Buchmueller, 2005
Terza, (undated)
Alexandre & French, 2004
French, Roebuck, Alexandre, 2001
Zuvekas, Cooper, & Buchmueller, 2005
Hodgins et al., 1996
Tiihonen, 1997
Wallace et al., 2004
Carpenter, 2003
Fergusson and Horwood, 2000
Lipsey et al., 1996

Employment =f(alcohol disorder)
Employment =f(alcohol disorder)
Employment =f(alcohol disorder)
Employment =f(alcohol disorder)
Employment =f(alcohol disorder)
Employment =f(alcohol disorder)
Employment =f(alcohol disorder)
Employment =f(alcohol disorder)
Employment =f(alcohol disorder)
Employment =f(alcohol disorder)
Employment =f(alcohol disorder)
Wages of workers =f(alcohol disorder)
Wages of workers =f(alcohol disorder)
Wages of workers =f(alcohol disorder)
Wages of workers =f(alcohol disorder)
Wages of workers =f(alcohol disorder)
Employment =f(DSM mental illness disorder)
Employment =f(DSM mental illness disorder)
Employment =f(DSM mental illness disorder)
Employment =f(DSM mental illness disorder)
Employment =f(DSM mental illness disorder)
Employment =f(DSM mental illness disorder)
Employment =f(DSM mental illness disorder)
Employment =f(DSM mental illness disorder)
Wages of workers =f(DSM mental illness disorder)
Wages of workers =f(DSM mental illness disorder)
Wages of workers =f(DSM mental illness disorder)
Wages of workers =f(DSM mental illness disorder)
Wages of workers =f(DSM mental illness disorder)
Wages of workers =f(DSM mental illness disorder)
Wages of workers =f(DSM mental illness disorder)
Employment =f(drug disorder)
Employment =f(drug disorder)
Employment =f(drug disorder)
Employment =f(drug disorder)
Employment =f(drug disorder)
Employment =f(drug disorder)
Wages of workers =f(drug disorder)
crime =f(mental illness)
crime =f(mental illness)
crime =f(mental illness)
crime =f(alcohol disorder)
crime =f(alcohol disorder)
crime =f(alcohol disorder)

29

Exhibit B.4a
Citations to Studies in Exhibit B.4
Alexandre, P. & French, T. (2001). Labor supply of poor residents in metropolitan Miami, Florida: The role of depression and the co-morbid effects of substance use. The
Journal of Mental Health Policy and Economics, 4(4): 161-173.
Alexandre, P.K. & French, M.T. (2004). Further evidence on the labor market effects of addiction: Chronic drug use and labor supply in metropolitan Miami. Contemporary
Economic Policy, 22(3): 382-393.
Auld, M.C. (2002). Robust system estimation of causal effects on binary outcomes, with application to effect of alcohol abuse on employment. Calgary, Alberta, Canada:
Department of Economics, University of Calgary. http://econ.ucalgary.ca/research/200301WP.pdf.
Bartel, A. & Taubman, P. (1986). Some economic and demographic consequences of mental illness. Journal of Labor Economics, 4(2): 243-256.
Bray, J. W. (2005). Alcohol use, human capital, and wages. Journal of Labor Economics, 23(2): 279-312.
Buchmueller, T.C. & Zuvekas, S.H. (1998). Drug use, drug abuse, and labour market outcomes. Health Economics, 7(3): 229-45.
Carpenter, C. (2003). Does heavy alcohol use cause crime? Evidence from underage drunk driving laws. Unpublished paper.
Chatterji, P., Alegría, M., Lu, M., & Takeuchi, D. (2005). Psychiatric disorders and labor market outcomes: Evidence from the National Latino and Asian American Study
(Working Paper No. 11893). Washington, DC: National Bureau of Economic Research.
Chevrou-Severac, H. & Jeanrenaud, C. (2002). The impact of alcohol abuse on employment in Switzerland. Neuchatel, Switzerland: IRER.
http://perso.wanadoo.fr/ces/Pages/english/Poster17.pdf.
Cook, P.J. & Peters, B.L. (2005). The myth of the drinker's bonus (Working Paper No. W11902). Washington, DC: National Bureau of Economic Research.
DeSimone, J. (2002). Illegal drug use and employment. Journal of Labor Economics, 20(4): 952-977.
Ettner, S.L., Frank, R.G., & Kessler, R.C. (1997). The impact of psychiatric disorders on labor market outcomes. Industrial & Labor Relations Review, 51(1): 64-81.
Farahati, F., Booth, B., & Wilcox-Gők, V. (2003). Employment effects of comorbid depression and substance use. DeKalb, IL: Northern Illinois University, Department of
Economics.
Feng, W., Zhou, W., Butler, J., Booth, B., & French, M. (2001). The impact of problem drinking on employment. Health Economics, 10(6): 509-521.
Fergusson, D.M. & Horwood, L.J. (2000). Alcohol abuse and crime: A fixed-effects regression analysis. Addiction, 95(10): 1525-1536.
Frank, R. & Gertler, P. (1991). An assessment measurement error bias for estimating the effect of mental distress on income. The Journal of Human Resources, 26(1): 154164.
French, M.T. & Zarkin, G.A. (1998). Mental health, absenteeism and earnings at a large manufacturing worksite. The Journal of Mental Health Policy and Economics, 1(4):
161-172.
French, M.T., Roebuck, M.C., and Alexandre, P.K. (2001). Illicit drug use, employment, and labor force participation. Southern Economic Journal, 68(2): 349-368.
Hamilton, V.H., Merrigan, P., & Dufresne, E. (1997). Down and out: estimating the relationship between mental health and unemployment. Health Economics, 6(4): 397-406.
Harwood, H. (2000). The economic cost of mental illness, 1992. Fairfax, VA: The Lewin Group.
Hodgins, S., Mednick, S.A., Brennan, P.A., Schulsinger, F., & Engberg, M. (1996). Mental disorder and crime. Evidence from a Danish birth cohort. Archives of General
Psychiatry, 53(6): 489-96.
Kenkel, D.S. & Ribar, D.C. (1994). Alcohol consumption and young adults' socioeconomic status. Brookings papers on economic activity: Microeconomics: 119-161.
Kessler, R.C. & Frank, R. (1997). The impact of psychiatric disorders on work loss days. Psychological Medicine, 27(4): 861-873.
Kessler, R.C., Barber, C., Birnbaum, H.G., Frank, R.G., Greenberg, P.E., Rose, R.M., et al. (1999). Depression in the workplace: Effects on short-term disability. Health
Affairs, 18(5): 163-171.
Lipsey, M.W., Wilson, D.B., Cohen, M.A., & Derzon, J.H. (1996). Is there a causal relationship between alcohol use and violence? A synthesis of evidence. In M. Galanter
(Ed.), Recent Developments in Alcoholism, Volume 13: Alcoholism and Violence (pp. 245-282). New York: Plenum Press.
MacDonald, Z. & Shields, M. (2004). Does problem drinking affect employment? Evidence from England. Health Economics, 13 (2): 139-155.
Marcotte, D.E. & Wilcox-Gők, V. (2003). Estimating earnings losses due to mental illness: A quantile regression approach. The Journal of Mental Health Policy and
Economics, 6(3): 123-134.
Mullahy, J. & Sindelar, J.L. (1993) Gender differences in labor market effects of alcoholism. The American Economic Review, 81(2): 161-165.
Mullahy, J. & Sindelar, J.L. (1996). Employment, unemployment, and problem drinking. Journal of Health Economics, 15(4): 409-34.
Mullahy, J. & Sindelar, J.L. (1997). Women and work: Tipplers and teetotalers. Health Economics, 6(5): 533-537.
Savoca, E. & Rosenheck, R. (2000). The civilian labor market experiences of Vietnam-era veterans: The influence of psychiatric disorders. The Journal of Mental Health
Policy and Economics, 3(4): 199-207.
Stewart, W.F., Ricci, J.A., Chee, E., Hahn, S.R., & Morganstein, D. (2003). Cost of lost productive work time among US workers with depression. Journal of the American
Medical Association, 289(23): 3135- 3144.
Terza, J.V. (2002). Alcohol abuse and employment: A second look. Journal of Applied Econometrics, 17(4): 393-404.
Terza, J.V. (n.d.). Assessing the impact of substance abuse on employment status. Charleston, SC: Medical University of South Carolina, Department of Health
Administration and Policy, Center for Health Economic and Policy Studies. http://people.musc.edu/~terza/empfina2.pdf.
Tiihonen, J., Isohanni, M., Räsänen, P., Koiranen, M., & Moring, J. (1997). Specific major mental disorders and criminality: A 26-year prospective study of the 1966 northern
Finland birth cohort. American Journal Psychiatry, 154(6): 840-845.
Wallace, C., Mullen, P.E., & Burgess, P. (2004). Criminal offending in schizophrenia over a 25-year period marked by deinstitutionalization and increasing prevalence of
comorbid substance use disorder. American Journal of Psychiatry, 161(4): 716-727.
Zarkin, G.A., French, M.T., Mroz, T., & Bray, J.W. (1998). Alcohol use and wages: New results from the National Household Survey on Drug Abuse. Journal of Health
Economics, 17(1): 53-68.
Zuvekas, S., Cooper, P.F., Buchmueller, T.C. (2005). Health behaviors and labor market status: The impact of substance abuse (Working Paper No. 05013). Rockville, MD:
US Department of Health and Human Services, Agency for Healthcare Research and Quality. http://gold.ahrq.gov.

30

The earliest age that a person might participate in an EBT is
notated as p and runs to the maximum possible age P (values
for p and P are shown on lines 12 and 13 of Exhibit B.3,
respectively). The annual cash flows in each year following
investment is the weighted sum of the product of the adjusted
earnings E in year y for the age of the program participant p,
times the annual real growth rate in earnings ER, times the
estimated earnings effect EEt, times the probability of program
participation PP at age p. This procedure produces a series of
expected annual cash flows representing lost earnings
following investment and weighted by the probability of
program participation for the ages of the people assumed to
enter the EBT.

B7. Lost Household Production Methods
As described above, in addition to the value of reduced or lost
performance in the commercial economy, many studies of
morbidity and mortality costs include estimates of the reduced
or lost value of household production.63 We adopt that
approach in this study.
To compute the household production effect for the incidence
of the DSM disorders, we begin with the following equation:
B(10): H a = HOURS * $ HOUR * 52 * Pr SHIFTa * INFLATION
For each age a, the annual value of household production Ha is
the HOURS per week for household tasks (line 9 from Exhibit
B.1, times the weighted average dollars per hour $HOUR for
household tasks (line 10), times 52 weeks per year, times the
probability that household tasks get shifted to someone else
PrSHIFT for a person who is age a (from Exhibit B.2), times the
INFLATION adjustment to bring the hourly wage (denominated
in 2004 dollars) to the year chosen for this analysis (2006
dollars).
Not all of the value of lost household production will be shifted
to others if a person dies or is disabled as a result of having an
alcohol, drug, or mental health disorder. Some people live
alone and no one else is required to assume the household
production if the person becomes disabled or dies as a result
of the disorder. We provide an estimate for this with the
variable PrSHIFTa, used in the previous equation. This variable
provides an estimate of the probability that a person at age a
will not be living alone and, if he or she becomes disordered,
that the value of his or her household production will be shifted
to someone else. We estimate this probability with national
data from the same 2005 Current Population Survey (with data
for 2004) described above.64 The results of this estimation are
shown in Exhibit B.2 and are computed with this equation:
B(11): Pr SHIFTa =

FHH a
( HH a − GQa )

The probability of shifting household production PrSHIFTa in the
event of a disorder is given by the total number of people in
households with family members FHHa divided by the total
number of people in households HHa (less those living in group
quarters GQa). Values for all three variables come from the CPS.
The annual cash flows of lost household production associated
with having a disorder of type t is estimated with this process:

63

Max et al., Valuing human life.
Current Population Survey data downloaded from the US Census Bureau
site with the DataFerrett extraction utility:
http://www.bls.census.gov/cps/cpsmain.htm

P

B(12): $ HPty = ∑ H p + y −1 *(1 + ER) y −1 * EEt * PPtp * −1
p

In this equation, $HPty is the annual cash flow of shifted
household production in year y, where y is the number of years
following participation in an EBT. The subscript y equals 1
during the year that a person is administered an EBT. The
earliest age that a person might participate in an EBT is notated
as p and runs to the maximum possible age P (values for p and P
are shown on lines 12 and 13 of Exhibit B.3, respectively). The
annual cash flows in each year following investment is the sum
of the product of household production H in year y for the age of
the program participant p, times the annual real growth rate in
earnings ER, times the estimated earnings effect EE, times the
probability of program participation PP at age p. This procedure
produces a series of expected annual cash flows representing
shifted household production following investment and weighted
by the probability of program participation for the ages of the
people assumed to enter the EBT.

B8. Health Care and Other Costs
An additional set of costs of alcohol, drug, and mental health
disorders covers the effect the disorders have on health care
costs. We show our assumptions and estimates for this on
lines 51 through 59 in Exhibit B.3. We start with the national
estimates provided by Harwood in his several reports on the
costs of alcohol, drug, and mental health disorders. These
amount to $44 billion for alcohol disorders in 1998, $15.7 billion
for drug disorders in 2002, and $46.2 billion for serious mental
illness in 1992.65 On line 54, we show the adult (age 18 and
over) population for the relevant years from the US Census
Bureau as reported in the Statistical Abstract of the United
States. On line 55, we multiply the total adult population
estimates by the same 12-month prevalence percentages
reported in the Harwood studies (.074 for alcohol, .015 for
drug, and .038 for serious mental illness). The average costs
are then computed and shown on line 55; we report on line 56
the plus and minus percentage change we use in sensitivity
analyses for the average health care cost figure. Finally, on
lines 57 though 59 we report the Harwood percentages for the
amount of health care costs incurred by taxpayers,
participants, and other private payers.
The annual cash flows of health care costs associated with
having a disorder of type t is estimated with this process:
P

B(13): $ HCty = ∑ HCCOSTt * (1 + HR) y −1 ∗ PPtp
p

In this equation, $HCty is the annual cash flow of health care
costs in year y, where y is the number of years following
participation in an EBT. The subscript y equals 1 during the year
that a person is administered an EBT. Before entering this
equation, the HCCOST estimate is already denominated in the
dollars for the year chosen for this analysis, including the real
rate of escalation in health care costs from the year of the
underlying Harwood study to the base year chosen for this
analysis (2006 dollars). The earliest age that a person might
participate in an EBT is notated as p and runs to the maximum
possible age P (values for p and P are shown on lines 12 and 13
of Exhibit B.3, respectively). The annual cash flows in each year
following investment is the sum of the product of average per
capita health care costs HCCOST, times the annual real growth
rate in health care costs HR, times the probability of program

64

65

See footnote 6.

31

participation PP at age p. This procedure produces a series of
expected annual cash flows representing health care costs
following investment and weighted by the probability of program
participation for the ages of people assumed to enter the EBT.

B9. Mortality Parameters and Methods
If the prevalence of alcohol, drug, or mental health disorders is
reduced with EBT, then one form of benefits will be that people
live longer and, as a result, are more productive in the
marketplace. All cost-of-illness studies estimate these
mortality-related effects. The mortality methods we employed
in this study required three pieces of information. The first is
shown on line 32 on Exhibit B.3: the normal life expectancy for
the average adult today. We estimated this parameter from the
Center for Disease Control for the average life expectancy of a
40-year-old, which corresponds roughly to the average age of
a person in our prototype programs.66
For people who die of a disorder, we estimated the probability
of death by age of death. We used data from the Washington
State Vital Statistics dataset. For alcohol and drug related
deaths, we counted the age of all deaths in Washington with
ICD-10 death codes where a certain percentage of the deaths
had been attributed to the disorder. For alcohol related deaths,
we used the attribution factors for the individual diagnoses
67
contained in Max et al. For drug related deaths, we used the
68
attribution factors contained in Harwood et al. For suicide
deaths, we used all deaths in Washington coded as a suicide.
Using these counts of actual Washington deaths, we then
estimated a probability density distribution for each disorder
(alcohol, drug, and suicide). Lines 33 through 37 contain the
parameters of these equations. We found that for alcohol
related deaths, a Beta distribution best fit the actual
Washington data; for drug related deaths, a Normal distribution
fit the data; and for suicides (mental health deaths), a Weibull
distribution was best. For alcohol and drug deaths, we
estimated the distributions with Washington data for 2004; for
suicides we used Washington data for 2003 and 2004 to
increase the sample size.
For each disorder, this process produces:
B(14): DDa ,
where DDa is the probability density distribution of a person
with an alcohol or drug disorder or a suicide at age a, and the
distributions are defined by a Beta, Normal, or Weibull,
respectively.
Not everyone who has an alcohol, drug, or mental illness
disorder dies of the disorder. Lines 38 through 43 of Exhibit B.3
list the parameters we used to estimate the probability that a
person with a history of a disorder dies of the disorder. For the
United States, Harwood estimated the total number of deaths in
1992 (for alcohol), 2000 (for drugs), and 1992 (for suicides) that
were caused by having an alcohol, drug, or mental disorder,
respectively. These values are shown on line 42, while line 40
shows the total number of deaths in the United States (for people
15 or older) during those years. Line 41 is the product of line 40
66

D. Hoyert, H. Kung, and B. Smith. (2005). Expectation of life by age,
race, and sex: United States, final 2002 and preliminary 2003. US
Department of Health and Human Services, Centers for Disease Control
and Prevention, National Vital Statistics Report, 53(15), Table 6.
http://www.cdc.gov/nchs/data/nvsr/nvsr53/nvsr53_15.pdf
67
Max et al., Valuing human life, Table 2.
68
Office of National Drug Control Policy. (2004).The economic costs of drug
abuse in the United States, Table B-10. http://www.whitehousedrugpolicy.
gov/publications/economic_costs/economic_costs.pdf

32

and line 20, the lifetime prevalence of each disorder. This
provides an estimate of the number of people who died in the
relevant year who had a disorder sometime in their lives. Line
43 is computed as line 42 divided by line 41; it is the attributed
death factor, ADF, for each disorder.
The annual cash flows of lost earnings and household
production associated with having a death caused by having a
disorder of type t is estimated with this process:
B(15):
P

$ Dty = ∑ [ E p + y −1 + H p + y −1] * (1 + ER) r −1 * DDtp + y −1 K
p

L * ADFt * Ptp

In this equation, $Dty is the cash flow of lost earnings E and
household production H in year y, where y is the number of years
following participation in an EBT. The subscript y equals 1 in the
year that a person is administered an EBT, and runs to M—the
maximum follow-up period for estimating cash flows. The earliest
age that a person might participate in an EBT is notated as p and
runs to the maximum possible age P (values for p and P are
shown on lines 12 and 13 of Exhibit B.3, respectively). The
annual cash flows in each year following investment is computed
as the weighted sum of the product of the adjusted earnings E by
year y for the age of the program participant p, plus the
household production H by year y for the age of the program
participant p, times the real growth rate in earnings ER, times the
probability of a death occurring, DD, by year y for the age of the
program participant, times the attributed death factor ADF for the
disorder, times the probability of program participation PP for a
participant of age p. This procedure produces a series of
expected annual cash flows representing lost earnings and lost
household production following investment and weighted by the
probability death and of program participation for the ages of the
people assumed to enter the EBT.

B10. Crime Parameters
The effect that alcohol, drug, and mental health disorders have
on crime is estimated in a two-step process. First, we use
meta-analyses to examine the existing research literature on
the linkage between each of these disorders and crime.
Second, if the meta-analyses reveal a statistically significant
relationship, we then use the Institute’s cost-of-crime model to
estimate the effects that the increased levels of crime have on
taxpayers (who fund the criminal justice system) and crime
victims (who suffer out-of-pocket costs and pain and suffering
costs when they are victimized). Then, to the degree that an
evidence-based treatment reduces the incidence of a disorder,
the estimated costs of crime are also reduced via this linkage.
In Exhibit B.4 we list the results of the meta-analyses we
performed on the linkage between the disorders and crime.
We only found a few studies where the research design was
robust. These few studies did provide some evidence of a
statistically significant relationship between alcohol disorder
and crime, and between mental illness and crime. We were
unable to locate studies establishing a relationship between
drug disorders and crime; this is a result consistent with other
inquires into this topic.69 Nonetheless, in Washington State the
consumption of these substances is illegal and, therefore,
69

See, for example: H. White, & D. Gorman. (2000). Dynamics of the
drug crime relationship. In G. Lafree (Ed.), Criminal Justice 2000:
Volume 1: The nature of crime: continuity and change (NCJ 182408,
pp. 151-218). Washington, DC: US Department of Justice, Office of
Justice Programs, National Institute of Justice.
http://www.ncjrs.org/criminal_justice2000/vol_1/02d.pdf.

these drug crimes can result in a criminal justice system
response including arrest, prosecution, and a full range of
sentencing outcomes. These effects are modeled.
The Institute’s model of the costs of crime has been described in
detail elsewhere; the interested reader can find a full description
of the routines used to calculate costs in the earlier reports.70

B11. Marginal Treatment Effect
The estimated benefits of treatment are determined by the
marginal effectiveness, over time, of EBT. The following
equation is used to estimate the marginal treatment effect MTE
for a person in an EBT treating a disorder of type t:
( ESty )
B(16): MTEty = Nty − sin(arcsin( Nty ) +
)^ 2 ,
2
where
B(17): ESty = ESt ∗ (1 + decayratet ) y −1 ∗ scaleupt ,
and where
B(18): Nty = NRta + NRtb1 * y + NRtb2 * y 2 + NRtb3 * y 3 .
For each of the three prototype programs t, we estimate the
marginal treatment effect with the parameters in these equations.
The variable Nty is the “natural rate of recovery” from a disorder
without treatment in year y for treatment type t, where y is a year
following participation in an EBT. The subscript y equals 1
during the year that a person is administered an EBT.
We estimated years from onset to “natural recovery” using data
from the 2001–2002 National Epidemiologic Survey on Alcohol
and Related Conditions (NESARC).71 The NESARC is a
longitudinal survey conducted by the federal National Institute
on Alcohol Abuse and Alcoholism. The 2001–2002 NESARC
is the first wave of the survey, with a sample of 43,093
respondents representative of the US adult population 18 years
of age and older. We performed separate analyses for
respondents who reported ever having the following conditions:
alcohol dependence, substance dependence, major
depression, dysthymia, mania or hypomania, panic disorders
and agoraphobia (anxiety), social phobia, specific phobia, and
generalized anxiety.
We analyzed the NESARC data using the generalized leastsquares estimation method that modeled the elapse (in years)
between the onset of a condition and the year in which the last
episode of symptoms was reported. To simulate “natural
72
recovery,” we estimated the elapsed time only for respondents
who reported no treatment since onset. Each estimation model
includes the following covariates: age at the interview, age at
the onset of the condition, sex, and high school diploma status.
In addition, the model on alcohol dependence includes the
covariates of ever having substance dependence and ever
having a DSM-IV diagnosis of mental illness; the model on
substance dependence includes the covariates of ever having
alcohol dependence and also ever having a DSM-IV diagnosis
of mental illness; and the models for mental illness conditions
each include the covariates of ever having alcohol dependence

and substance dependence. The analyses were performed
using the SAS procedure of SURVEYREG. SURVEYREG is
specially designed for regression analyses involving sample
survey data. The procedure allows for adjustments for complex
sample designs, including sample stratification, clustering, and
73
unequal weights. The parameters shown on lines 60 through
63 in Exhibit B.3 are the parameters for a third degree
polynomial for each prototype; for use in the simulation model,
these are linear representations of the logistic models estimated
with SAS.
The determination of the effect size that is used for each year,
ESty is computed with the short-run effect size, ESt, for each
prototype evidence-based treatment, discussed elsewhere in
this Appendix. These effect sizes were almost always obtained
from studies with quite short follow-up periods, usually around
a year. To account for the possibility that these short-run effect
sizes might decay over the long run, we estimated decay rates,
decayratet, for each prototype treatment. We describe how we
obtained estimates for the decay rate in Appendix B2. In
addition, also as described in Appendix B2, we multiplied the
effect sizes by a factor, scaleupt, that is designed to reflect
reduction in effect sizes that are likely to occur when smallscale programs are expanded significantly.

B12. Sensitivity Analysis
The model as described in this Appendix produces a unique
result given the set of inputs listed. As we describe, however,
there is a significant amount of uncertainty around many of the
inputs. For most inputs to the model, we determine the range
of uncertainty with the standard errors or standard deviations
from relevant statistics of the underlying data for each
parameter. For a few other parameters, we hypothesized low
and high ranges to place bounds on our estimates of
uncertainty.
After we specified ranges of uncertainty on each of the inputs,
we then used a simulation approach to determine how
sensitive the final result is to these levels of uncertainty. To
conduct the simulation, we used Palisade Corporation’s
@RISK® simulation software. Using a Monte Carlo approach
to the simulation, the software randomly draws from the userdesignated input variables after a particular type of probability
distribution and its parameters have been specified for the
input. We ran a Monte Carlo simulation for 10,000 cases.
Exhibit B.5 shows input variables along with the specified
probability distributions that we used in the simulation.

70

See footnote 5.
http://niaaa.census.gov/
The term “recovery” refers to situations in which the last episode of
symptoms had occurred no later than a year prior to the interview. It should
be noted that this term is not used in the strict meaning as “cured” because
in some situations the absence of symptoms before the interview could just
be the “breathing” period between episodes.

71
72

73

®

SAS Institute Inc. 2004. SAS OnlineDoc 9.1.2. Cary, NC: SAS
Institute Inc.

33

Exhibit B.5
The Benefits and Costs of Evidence-Based Treatment:
Model Parameters Varied in the Monte Carlo Simulations
Probability
Distribution Type
in @RISK

®

See text for information about these parameters

Evidence-Based Treatment: Adults With
Alcohol, Drug, or Mental Illness Disorders
Adults with a
serious DSM
alcohol disorder

Adults with a
serious DSM drug
disorder

Adults with a
serious DSM mental
illness disorder

1. Program Effectiveness Parameters
Adjusted effect size after applying WSIPP* adjustments
Estimated standard error for the WSIPP-adjusted effect size
Expected annual rate of decay in effect size
Standard error
Expected diminishing returns to effect size with large scale ramp up
(lower expected rate of decay)
(higher expected rate of decay)

Normal

-.247
.021
-.062
.027
.750
1.000
.500

-.355
.035
-.164
.072
.750
1.000
.500

-.360
.058
-.176
.089
.750
1.000
.500

Normal

$2,300
$500

$2,300
$500

$3,596
$782

Normal

5.55%
0.26%

2.05%
0.16%

3.80%
0.22%

Normal

11.1%
0.4%
50%
75%
25%

14.7%
0.9%
50%
75%
25%

46.2%
3.5%
50%
75%
25%

Normal

-0.260
0.061

-0.262
0.059

-0.250
0.038

Triangular

$4,496
10.0%

$6,114
10.0%

$13,799
10.0%

Normal
Triangular

2. Program Design Parameters
Average annual program cost
Standard deviation of average program cost

3. Prevalence Parameters
Current (12-mo) prevalence of DSM disorder in this population cohort
Standard error

4. Potential Population to be Treated
Proportion of target population already treated with evidence-based program
Standard error
Proportion of the currently unserved target population that might realistically be served
high
low

Triangular

5. Morbidity Parameters (earnings and household production)
Employment outcomes =f(Disorder)
Standard error

6. Health Care Costs
Annual cost/ current abuser (adjusted to base year for real growth in costs)
Assumed percentage (plus and minus) from the average cost

7. General Model Parameters
Discount Rate
Real annual rate of growth in earnings
Real annual rate of growth in health care costs
* Washington State Institute for Public Policy

34

High
Discrete (equal %)
Triangular
Triangular

.070
.023
.044

Medium
.050
.013
.034

Low
.030
.003
.024

For further information, contact Steve Aos at
(360) 586-2740; saos@wsipp.wa.gov

Document No. 06-06-3901
Washington State
Institute for
Public Policy
The Washington State Legislature created the Washington State Institute for Public Policy in 1983. A Board of Directors—representing the legislature,
the governor, and public universities—governs the Institute and guides the development of all activities. The Institute’s mission is to carry out practical
research, at legislative direction, on issues of importance to Washington State.

 

 

BCI - 90 Day Campaign - 1 for 1 Match
Advertise Here 4th Ad
The Habeas Citebook Ineffective Counsel Side