Psychology of Addictive Behaviors Copyright 2001 by the Educational Publishing Foundation

2001, Vol. 15, No. 2, 83-88 0893-164X/01/S5.00 DOI: 10.1037//0893-164X.15.2.83

Manual-Guided Cognitive-Behavioral Therapy Training: A Promising

Method for Disseminating Empirically Supported Substance Abuse
Treatments to the Practice Community

Jon Morgenstern Thomas J. Morgan and Barbara S. McCrady

Mount Sinai School of Medicine Rutgers—The State University of New Jersey

Daniel S. Keller Kathleen M. Carroll

New York University Yale University School of Medicine
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

A gap exists between empirically supported substance abuse treatments and those used in community
settings. This study examined the feasibility of training substance abuse counselors to deliver cognitive-
behavioral treatment (CUT) using treatment manuals. Participants were 29 counselors. Counselors were
randomly assigned to receive CBT training or to a control group. Counselor attitudes were assessed pre-
and posttraining. In addition, CBT therapy sessions were videotaped and rated for adherence and
skillfulness. CBT counselors reported high levels of satisfaction with the training, intention to use CBT
interventions, and confidence in their ability to do so. Ratings indicated that 90% of counselors were
judged as having attained at least adequate levels of CBT skillfulness. Findings demonstrate the
feasibility of using psychotherapy technology tools as a means of disseminating science-based treatments
to the substance abuse practice community.

Although demonstrated effective interventions have been devel- increase the accessibility of research findings to clinicians because
oped to treat a number of mental disorders, there continues to be they describe procedures to implement treatments at a high level of
a disjunction between treatments that are empirically supported technical specificity, accelerate the learning of new techniques,
and those used in practice settings. The gap between research and and facilitate instruction to therapists of different theoretical ori-
practice may be especially wide in substance abuse, because sub- entations. In addition, the use of manuals to actually deliver
stance abuse clinicians and scientists differ markedly in their treatment may enhance efficacy and provide a means of quality
training, professional identifications, and treatment philosophies. control for therapist performance, similar to that achieved in re-
Despite long-standing concerns, to date disappointingly little search contexts.
progress has been made in disseminating empirically supported The potential use of manuals as the primary medium for dis-
treatments (ESTs) to substance abuse practitioners (e.g., Gordis, seminating ESTs has sparked considerable debate and calls for
1991). further study (e.g., Addis, 1997). One critical issue concerns the
Recent changes in health care policy and the development of feasibility of training clinicians to deliver ESTs competently and to
treatment standardization procedures may provide a fresh impetus
incorporate them into their practice routines. Community providers
and potential new solutions to address this problem. Specifically,
are typically less well trained than research clinicians and differ in
evolving criteria for third-party reimbursement of services are
theoretical orientation. Questions arise as to whether providers can
likely to be based on evidence of effectiveness and cost (Barlow,
learn to competently deliver ESTs using manuals and whether
1996). Thus, clinicians will have a new and powerful incentive for
providers will replace favored treatment strategies with ESTs
changing treatment practices. In addition, treatment manuals may
offer an ideal tool for dissemination efforts. Treatment manuals following training.
Therapist feasibility issues are particularly salient when consid-
ering the dissemination of ESTs to substance abuse practitioners in
the United States. Substance abuse counselors provide the majority
Jon Morgenstern, Department of Psychiatry, Mount Sinai School of of care in the current system. Counselors have markedly less
Medicine; Thomas J. Morgan and Barbara S. McCrady, Center of Alcohol formal education, and less clinical training, than either therapists
Studies, Rutgers—The State University of New Jersey; Daniel S. Keller, used in clinical trials or than their counterparts in mental health
Department of Psychiatry, New York University; Kathleen M. Carroll, treatment. For example, a substantial proportion of counselors do
Department of Psychiatry, Yale University School of Medicine. not have master's degrees, and many have not completed 4 years
Preparation of this article was supported by Grant AA08747 from the
of college (Institute of Medicine, 1997). In addition, interventions
National Institute on Alcohol Abuse and Alcoholism
Correspondence concerning this article should be addressed to Jon developed in research settings have been predominantly cognitive-
Morgenstern, Mount Sinai School of Medicine, Department of Psychiatry, behavioral in orientation. However, most counselors espouse a
Box 1230, One Gustave L. Levy Place, New York, New York 10029. 12-step approach to treating substance use problems (Wallace,
Electronic mail may be sent to [email protected]. 1996). Cognitive-behavioral therapy (CBT) and 12-step treatment
83
 84 MORGENSTERN, MORGAN, McCRADY, KELLER, AND CARROLL

approaches differ substantially, especially with regard to their school. Most counselors had extensive substance abuse treatment experi-
underlying theory. ence (M = 8.7 years, SD = 6.8). Thirty-eight percent of the counselors
Lower levels of education and clinical training, and a strong reported being in recovery.
allegiance to a conflicting treatment model, raise serious questions
about counselor ability to master the delivery of CBT for substance Procedure
abuse and their response to protocol-based training methods, as
Participant recruitment. The selection of a setting for the study was
well as their willingness to embrace the use of CBT techniques initiated by a review of substance abuse treatment programs located in
following training. Only one prior study has reported on attempts central New Jersey. Programs were considered if they provided outpatient
to disseminate protocol-based substance abuse treatments to com- substance abuse treatment using a traditional chemical dependency care
munity practitioners (Sobell, 1996). Sobell (1996) reported on model, were licensed by the state, evidenced financial stability, and were
clinician response but did not evaluate clinicians' ability to deliver recognized as accepted members of the provider community in New Jersey.
CBT following training. Program adherence to a traditional chemical dependency care model was
The primary aim of this study was to examine the feasibility of assessed based either on our knowledge of the program or through an
interview with the program's clinical director. Programs were considered
training front-line substance abuse providers to deliver CBT using
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

as accepted members of the provider community if they belonged to one of

treatment manuals. We addressed three specific issues. First, we
This document is copyrighted by the American Psychological Association or one of its allied publishers.

several county or state substance abuse provider organizations. Seven

were interested in examining counselors' subjective response to programs met these criteria and were contacted. All expressed interest in
the training. Several responses were assessed. These included participating in the study. Two programs were selected on the basis of their
satisfaction with learning CBT, satisfaction with the use of large and representative front-line clinical staffs and their record of fiscal
manual-based training methods, the perceived clinical utility of and clinical stability. These latter factors were important in the selection
CBT, and appraised self-efficacy in delivering CBT following process because the future fiscal viability of many outpatient programs was
training. Second, we were interested in exploring the relation in question because of the introduction of managed-care constraints in New
between counselors' beliefs about treatment and the nature of Jersey at the time.
addictive disorders and the training experience. Specifically, we Counselors who agreed to participate were administered informed con-
sent and completed questionnaires, which are described below. Counselors
assessed whether counselors' allegiance to the 12-step approach
were then assigned, using urn randomization procedures (Stout, Wirtz,
posed an obstacle to learning CBT and whether the training served
Carbonari, & Del Boca, 1994), to two groups balanced on the following six
to modify counselors' beliefs in 12-step and social learning theory factors: treatment beliefs, clinical experience, ethnicity, gender, education,
models. Third, we evaluated counselors' ability to deliver CBT and employment status at the program. Two thirds of the counselors were
following training. We evaluated performance by assessing adher- assigned to a CBT training group, and one third were assigned to a control
ence and skillfulness in delivering CBT based on an "expert" group. CBT training consisted of 35 hr of didactic classroom instruction
standard established in a rigorously implemented research study of over a 2-week period followed by clinical case training and intensive
CBT (Project MATCH Research Group, 1997). supervision. The control group also received training designed to minimize
counselors' feelings of being deprived of a valuable learning experience
and to avoid a Hawthorne-like effect on counselor motivation. The control
Method group received 8 hr of training in traditional substance abuse counseling
and were offered the opportunity to receive training in CBT at the end of
Sample and Setting the study. CBT training is described below.
Counselors assigned to the CBT group were administered an extensive
Participants were 29 front-line substance abuse counselors drawn from quantitative and qualitative evaluation survey at the end of the didactic
the clinical staffs of two outpatient chemical dependency treatment pro- training and again following the clinical case training. All treatment ses-
grams located in central New Jersey. The programs espoused a traditional sions were videotaped, and sessions selected at the end of training were
treatment model with interventions focused on reducing denial, educating rated for adherence and skillfulness. In addition, all counselors were
clients about the disease of addiction, facilitating affiliation with 12-step re-administered questionnaires assessing beliefs at the end of training.
self-help groups, and maintaining abstinence. Each employed master's- Responses to questionnaires and evaluation forms were treated as confi-
and less than master's-level trained personnel as front-line clinical staff. dential. Counselors in the CBT condition were asked not to share training
Counselors were eligible to participate in the study if they were currently information with those in the control condition, and this was monitored
providing substance abuse treatment to clients, had a minimum of 1 year of throughout the study.
prior treatment experience, and did not have prior formal training in CBT CBT training. Cognitive Behavioral Coping Skills Training (CBCST;
for substance abuse. Thirty-eight counselors representing the entire clinical Kadden et al., 1992) was selected as the CBT intervention. Protocol-based
staffs of each program's adult treatment division were approached and methods used to train therapists in research studies were adapted to train
agreed to participate. Two counselors did not meet eligibility criteria: One the counselors. To increase the applicability of training results, a struc-
had received prior CBT training, and the other did not meet the minimum tured, time-limited curriculum, similar in format to a continuing-education
treatment experience requirement. Four counselors were not included be- course, was developed. Thus, although training was intensive, all counsel-
cause of scheduling conflicts, and 3 left the programs prior to the end of ors were trained concurrent with the conduct of their regular counseling
training. Twenty counselors were trained in CBT, and 9 counselors served duties.
as a control group. The mean age of the sample was 41.5 years (SD = Thirty-five hr of didactic training were provided. Didactic training
11.4), and 65% (n = 19) were women. The ethnic composition of the contained theoretical and experiential elements, including discussions of
sample was 72% (n = 21) Caucasian, 21% (n = 6) African American, and the similarities and differences between CBT and 12-step models, the role
7% (n = 2) Hispanic. As is typical of substance abuse program staffs, of therapeutic alliance in the delivery of protocol-driven treatments, and
counselors had quite varied educational backgrounds. About 45% (n = 13) extensive role plays for each of the CBCST treatment sessions. CBT
had at least a master's degree in either psychology, counseling, social counselors then treated at least three and, if possible (depending on time
work, or nursing (1 counselor had a doctoral degree), and 55% (n = 16) constraints), four clients in 12-session individual treatment using the
either a bachelor's degree, an associate's degree, or had graduated high CBCST manual during the training period. Twenty-six percent of the
 MANUAL-GUIDED CBT 85

clients treated completed all 12 sessions. Sessions were videotaped and informal skills assessment, (f) skill teaching, (g) in-session skill exercise,
viewed by supervisors who provided session-by-session feedback. CBT and (h) assignment of practice exercise. The ITRS was constructed to
counselors received 1 hr of individual and 1 hr of group supervision per assess adherence to and skillfulness in delivering these eight elements. The
week. Supervisors were five doctoral-level clinical psychologists with scale contains 18 items: 1 item to assess extent of delivery and 1 to assess
extensive experience treating and supervising others in CBT for substance skillfulness of delivery for each element, as well as 2 summary items for
abuse. Three study authors (Jon Morgenstem, Thomas J. Morgan, and the entire session. Extent of delivery was rated on a 5-point Likert scale
Daniel S. Keller) were supervisors. Counselors received about 100 hr of (anchors: 1 = not at all, 5 = extensively), as was skillfulness (anchors: 1 =
didactic and clinical training over a 5-month period. very poor, 5 = excellent). We computed intraclass correlations (ICCs) to
Although the structure of the training was similar to that used in Project establish item reliabilities using ratings of 20 randomly selected tapes rated
MATCH, we substantially augmented and modified training materials to by all raters. Median ICCs and ranges for the various scales were as
address the challenges raised in training front-line counselors. In addition, follows: MTRS Extent, Man = .85 (.67-.96); MTRS Skillfulness, Man =
we revised many sections of the Project MATCH manual to simplify and .80 (.6S-.96); ITRS Extent, Man = .87 (.6S-.92); ITRS, Man = .87
streamline delivery. (.43-.94).
Rating CBT adherence and skillfulness. Clinical supervisors and Evaluation survey. We assessed counselors' responses to training us-
trained raters assessed counselor performance at the end of training on the ing an extensive quantitative and qualitative survey. Quantitative responses
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

basis of session videotapes. Methods and standards used in Project were measured using Likert-scaled items. Responses in five domains were
This document is copyrighted by the American Psychological Association or one of its allied publishers.

MATCH to assess therapist delivery were adapted for this study (Carroll et assessed: (a) overall satisfaction with training, (b) satisfaction with training
al., 1998). methods, (c) perceived clinical utility of CBT, (d) appraised self-efficacy in
Specifically, in Project MATCH supervisors rated the CBT performance delivering CBCST, and (e) ideological conflict experienced in delivering
of therapists using relevant items from the MATCH Videotape Rating CBT. Satisfaction with training was assessed with several items. Items
Scale (MTRS; Carroll et al., 1998). In that study, therapists who scored were highly correlated (average r — .77) and had similar response distri-
below 3 on a 5-poinl scale were judged less than adequate. In this study, butions. Quantitative responses are presented for two items: "How does
supervisors rated sessions using CBT items drawn from the MTRS. In this training experience compare to others?" and "Would you recommend
addition, supervisors provided an overall score for counselor skillfulness this training to a colleague?" We assessed response to training methods
based on their session ratings for the last case. Overall performance was using one item: "Compared to other training methods you've encountered,
rated on a 5-point Likert scale with anchors of 1 (very poor) to 5 (excel- how would you rate the combined use of treatment manuals, videotaping
lent). We attempted to calibrate supervisors' ratings of counselors to those sessions, and supervision on these videotapes?" We assessed the perceived
used by supervisors in Project MATCH. First, supervisors met regularly to clinical utility of CBT using the following two items: (a) "How often do
discuss counselor performance and view session videotapes. These discus- you plan to use CBT with your clients in group and individual treatment?"
sions were led by Daniel S. Keller, who was a CBT trainer and clinical and (b) "To what extent have the patients you treated with CBT benefited
supervisor in Project MATCH. In addition, Daniel S. Keller and Kathleen from the treatment?" Appraised serf-efficacy in delivering CBCST was
M. Carroll, who are very familiar with CBT therapists' performances in assessed with one item: "How confident are you that you can deliver CBT
Project MATCH, rated representative counselor videotapes and indicated effectively?" Conflict experienced in delivering CBT was assessed with the
how these compared to the performance of Project MATCH therapists. following item: "To what extent does [the] CBT you have been taught
We also trained raters to assess CBT adherence and competence cali- conflict with your convictions about what constitutes effective treatment
brated to standards used in Project MATCH. Raters were four advanced for substance abusers?" Counselors responded to this item under two
doctoral clinical psychology students with experience in CBT for substance hypothetical conditions: (a) when CBT is delivered as one component of a
abuse. Training raters involved several steps. First, videotapes for 8 coun- traditional program and (b) when CBT is delivered as a stand-alone
selors and 14 Project MATCH CBT therapists were selected to represent a intervention. A summary of qualitative responses to items is also reported.
wide range of therapist performance. These videotapes were then rated by Qualitative data were analyzed by identifying themes among the responses
the original raters used in Project MATCH. Videotapes were rated for CBT and ranking themes by frequency of occurrence. This analysis was con-
adherence and skillfulness using the MTRS and a scale developed specif- ducted by Jon Morgenstem.
ically for this study (see the following paragraph for description). The Assessment of beliefs about the nature of alcoholism and substance
purpose of this step was to calibrate ratings of counselor CBT performance abuse treatment. Two scales were administered to assess counselor be-
to those used in rating Project MATCH therapists. Next, study raters liefs. The Understanding of Alcoholism Scale (UAS; Moyers & Miller,
received 20 hr of didactic training and then rated at least 10 tapes that were 1993) is a 50-item self-report measure designed to assess beliefs about the
evaluated with regard to consensus ratings provided by the Project etiology and appropriate treatment of alcoholism. The UAS has two sub-
MATCH raters. scales: the Disease Model Beliefs subscale, which reflects adherence to the
disease model of alcoholism; and the Psychosocial Beliefs subscale, which
Measures reflects the belief that alcoholism in influenced by cultural experiences,
familial experience, or both. Item are presented on a 5-point Likert scale
Videotape rating scales. Adherence to protocol and skillfulness were ranging from weakest agreement (1) to strongest agreement (5). The
assessed with the MTRS (Carroll et al., 1998) and the Project IMPACT (an Treatment Processes Questionnaire (TPQ; Morgenstern & McCrady, 1992)
acronym used to identify this study; it stands for Improving Addiction is a 35-item self-report measure designed to assess clinicians' beliefs about
Counseling Through Technology Transfer) Tape Rating Scale (ITRS). The the therapeutic value of different processes for treating substance abuse.
MTRS assesses the extent of use of active ingredients of treatments The scale includes 10 Disease Model and 17 Behavioral Treatment Pro-
delivered in Project MATCH, including CBCST. In this study, raters cesses. Processes are represented on a 7-point Likert scale with anchors of
assessed only the eight items directly related to CBT. In addition, items +3 as "essential"; -3 as "detrimental"; and the midpoint, 0, as "no effect."
were added to assess skillfulness of delivery of these eight CBT ingredi-
ents. As indicated above, we accentuated the structured aspects of CBT
during counselor training, including providing an ideal prescribed sequence
Results
for the delivery of protocol elements. This invariant session structure Evaluation Survey
contained the following eight elements: (a) assessment of the client's
concerns and agenda, (b) addressing substance use or cravings since last The top part of Table 1 presents results of counselors' evalua-
session, (c) review of practice exercise, (d) delivery of session rationale, (e) tion of satisfaction with the CBT training as a whole as well as
 86 MORGENSTERN, MORGAN, McCRADY, KELLER, AND CARROLL

Table 1 ing corrected misperceptions that CBT is dry, mechanical, or

Results of Counselor Training Evaluation Survey incompatible with a 12-step approach; CBT is not compatible with
an experiential or psychodynamic approach.
Response Satisfaction with manualized training methods was also high but
Survey topic 1 2 3 4 5 somewhat less so than for the training experience as a whole.
Qualitative comments were: supervision via the use of session
Training satisfaction videotapes provides a unique opportunity to view one's perfor-
COMPARED" 0 5 11 32 53 mance objectively and far surpasses reliance on memory to convey
RECOMMEND* 0 5 15 5 75
MANUAL 0 5 5 10 65 15 session events during supervision; the manual enhances learning
Perceived utility of CBT and performance by providing structure and a reference to prepare
USE CBTd 0 0 5 40 55 for and review one's performance; manuals are too dry, uninter-
BENEFIT6 0 0 32 53 16 esting, and restrictive.
CONFIDENT' 0 0 20 60 20 In the middle section of Table 1 are presented counselors'
Experienced conflict
evaluations of the clinical utility and appraised confidence in
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

delivering CBT
AS COMPONENT8 delivering CBT. Counselors endorsed high ratings for the clinical
This document is copyrighted by the American Psychological Association or one of its allied publishers.

0 0 5 25 70
STAND ALONE" 5 20 20 20 35 utility of CBT, with 95% indicating that they would use CBT
interventions often or very often with patients, and rated high
Note. N = 20. Numbers represent percentages of counselors responding.
levels of benefit for clients whom they had treated with CBT.
Response formats for all items were 5-point Likert scales. Anchors for
items differed. In some cases responses add to more than 100% because of Counselors also expressed confidence that they could treat patients
rounding. CBT = cognitive-behavioral therapy. effectively with CBT while receiving routine clinical supervision.
" Represents responses to the question "How does this training experience
compare with others you have had?" Anchors were 1 = below average,
2 = average, 3 = above average, 4 = very good, 5 = one of the Relation of Counselor Beliefs and CBT Training
best. h Represents responses to the question "Would you recommend this
training experience to a colleague?" Anchors were 1 = no, definitely not; In the bottom section of Table 1 are presented counselors'
2 = no, I don't think so; 3 = yes, I think so; 4 = yes, with some responses regarding whether treating patients with CBT conflicted
reservations; 5 = yes, enthusiastically. c Represents responses to the with their convictions about what constitutes effective substance
question "Compared to other training methods you have encountered, how
abuse treatment. This question was posed under two hypothetical
would you rate the use of manuals, videotaping sessions, and supervision
of these videotapes?" Anchors were 1 = worse than others, 2 = about the conditions: when CBT is delivered as part of (a) a comprehensive
same as others, 3 = somewhat better than others, 4 = better than others, treatment program or as (b) a stand-alone treatment. When CBT is
5 = far better than others. d Represents responses to the question "How delivered as one component of a comprehensive program (as it was
often would you use CBT interventions with patients either in group or during this training), 70% of counselors experienced no conflict.
individual treatment?" Anchors were 1 = never, 2 = rarely, 3 = some-
times, 4 = often, 5 = very often. c Represents responses to the question
However, responses differed when CBT is delivered as the only
"To what extent have the patients you treated benefited from the CBT they treatment: About half of the counselors indicated this would rep-
received?" Anchors were 1 = not at all, 2 = minimally, 3 = moderately, resent a moderate to extreme conflict, and the remainder indicated
4 = considerably, 5 = extensively. f Represents responses to the question either slight or no conflict. Qualitative comments were: eclectic
"How confident are you that you can deliver CBT effectively with only one approaches work best for clients; CBT alone misses many ele-
hour of group supervision per week?" Anchors were 1 = very low
confidence, 2 = low confidence, 3 = moderate confidence, 4 = high ments needed for recovery, including developing sober peer sup-
confidence, 5 = very high confidence. E Represents responses to the ports, spirituality, and Alcoholics Anonymous affiliation; CBT
question "To what extent does CBT conflict with your convictions about delivered via a manual is too rigid and didactic to effectively
what constitutes effective treatment for substance abusers, when CBT is address patient needs without additional treatment.
delivered as one component of a comprehensive treatment program?"
Next we examined whether the CBT training influenced coun-
Anchors were 1 = extreme conflict, 2 = considerable conflict, 3 =
moderate conflict, 4 = slight conflict, 5 = no conflict. h Represents selors' beliefs about the nature of alcoholism and the therapeutic
responses to the question "To what extent does CBT conflict with your value of disease model and behavioral treatment processes. We
convictions about what constitutes effective treatment for substance abus- conducted a repeated measures analysis of variance to determine if
ers, when CBT is delivered alone? Anchors are the same as for AS counselors in the CBT group increased social learning theory
COMPONENT.
beliefs and decreased disease model beliefs following the training.
Group (CBT vs. control counselors), time (pre- and posttraining),
with the use of manuals and videotape supervision as a training and Group X Time interactions were not significant for disease
method. Counselors reported high levels of satisfaction with the model and behavioral treatment processes and social learning
training: Over 50% endorsed the most positive anchor for the two theory beliefs about alcoholism. There was a significant Group x
training satisfaction items (Compare, Recommend), and only 5% Time interaction for the Disease Model Beliefs subscale of the
gave a negative evaluation. Qualitative comments regarding over- UAS, F(l, 27) = 7.2, p < .01. Disease model beliefs decreased in
all satisfaction (presented in the order of frequency of occurrence) the CBT group but increased in the control group. Further exam-
were: CBT is an effective treatment, therefore, it is valuable to ination suggested that, prior to training, counselors endorsed sim-
learn; the training broadened the counselor's repertoire of clinical ilar levels of disease model and social learning theory beliefs. For
skills; traditional treatment doesn't work for everyone, therefore, it example, the mean counselor rating of disease model treatment
is important to know an alternative approach; CBT is appealing processes was 1.84 (SD = 0.72), and the mean rating of behavioral
because, unlike other approaches, it provides a systematic, step- treatment processes was 1.77 (SD = 0.59), suggesting that neither
by-step, concrete approach to address patient problems; the train- model was strongly favored over the other.
 MANUAL-GUIDED CBT 87

Counselor Adherence and Skillfulness in Delivering CBT interventions provided a good fit for counselors. Third, as adapted
for this study, the Project MATCH manual proved to be an
In Table 2 are presented ratings of counselor performance at the excellent training device, providing counselors with rapid access to
end of CBT training. Supervisors rated 90% of the counselors (n = a new set of therapeutic skills they deemed valuable. In part,
18) as at least adequate. Raters' independent ratings of Session 2 satisfaction with the CBT protocol training may stem from the
videotapes yielded a similar percentage of counselors who were limited skill counselors have in specific therapeutic techniques.
judged as at least adequate based on the standards applied to For example, counselors typically approach treatment with the
therapist performance in Project MATCH. A sum score of the goal of getting clients to actively cope with situational risks or
eight CBT items assessed by the MTRS was highly correlated with problematic thinking, but they lack specific techniques to accom-
the two summary items of the ITRS (adherence: r = .57, p < .01; plish these goals. The manual provided them with these
skillfulness: r = .71, p < .01). Therefore, there appeared to be techniques.
consistency across observers and rating scales. Some authors (e.g., Strupp & Anderson, 1997) have warned that
manuals may stifle the clinical artistry of therapists and limit their
Discussion autonomy. We anticipated that this would be a major source of
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

resistance. However, our experience was that manuals appeared to

Overall, the results indicate that counselors responded well to
This document is copyrighted by the American Psychological Association or one of its allied publishers.

improve the clinical work of many counselors by providing suffi-

the CBT content and manualized-based format of the training,
cient structure and specificity to facilitate a sustained, productive
expressing high levels of satisfaction with the experience and
confidence in their ability to effectively use the techniques. Very therapeutic focus.
Overall, the majority of counselors were able to learn to deliver
few dissemination studies have been conducted with substance
manualized CBT competently. Several caveats should be consid-
abuse counselors; therefore, formal hypotheses about training out-
ered in evaluating these performance results. Rated skillfulness
comes were not formulated. However, counselor response was
better than might be anticipated, given the expectation that adher- refers to the delivery of protocol-driven CBT and does not refer to
less standardized modes of CBT treatment. In addition, ratings of
ence to a conflicting treatment model, modest therapy skill levels,
and general clinician resistance to using standardized protocols skillfulness refer primarily to delivery of the core sessions from the
would pose major obstacles to training the counselors. Project MATCH CBT manual. These sessions address alcoholic-
Counselors' qualitative responses and our own observations of specific coping skills. The Project MATCH manual also contains
the training provide some insights into why these factors proved a series of elective skills sessions that address general coping
less problematic than anticipated. First, although most counselors skills. Counselors had greater difficulty learning to deliver these
espoused a 12-step treatment orientation, there was little evidence sessions. Finally, despite the general success, 10%-20% of coun-
of dogmatism or closed-mindedness in their approach to learning selors expressed minimal satisfaction with the training or were not
and using other treatment techniques. Rather, counselors acknowl- judged as adequate. Counselors not judged as adequate were ones
edged the limits of current treatments and were actively searching who expressed interest in the training but could not master the
for new skills that could improve client outcomes. CBT techniques.
Second, there are very significant differences between CBT and
the 12-step approach at the level of theory. However, at the level Study Limitations
of technique—the level addressed by treatment manuals—both
approaches share elements that are compatible (McCrady, 1994). Several study limitations should be noted. The study reports on
For example, both treatments are active and directive, and both intensive training of 20 front-line counselors. Several study pro-
place a primary focus on abstinence and make substantial use of cedures—selection of representative community programs and
didactic materials. Therefore, the style and content of the CBT random assignment of counselors—enhance the generalizability of
findings; nevertheless, the sample size is small, and findings
require replication with a larger sample. In addition, training cases
Table 2 were selected from clients who might benefit from CBT based on
Ratings of Counselor Adherence and Skillfullness in Delivering the judgment of the counseling staff. Generalization of counselor
Cognitive-Behavioral Therapy performance to other clients and to conditions that more fully
mirror routine clinical practice is not warranted. Finally, this study
Response* reports only on the feasibility of training counselors to deliver
Type of rating 1 2 3 4 5 science-based treatments, not on clinical outcomes. In subsequent
studies we hope to report on how the trained counselors performed
b
Supervisors' rating of skillfulness 5 5 45 45 0 under typical clinical conditions as well as to compare the clinical
Raters' ratings of adherence" 0 5 30 55 10 outcomes of CBT provided by counselors versus treatment as
Raters' ratings of skillfulness0 0 10 50 40 0 usual.
Note. N = 20.
* Anchors for skillfulness ratings were: 1 = very poor, 1 = poor, 3 = Conclusions
adequate, 4 = good, 5 = excellent; anchors for adherence ratings were 1 =
not at all, 2 = a little, 3 = somewhat, 4 = considerably, 5 = exten- Overall, the study's findings demonstrate the feasibility of using
sively. b These are supervisors' ratings of cognitive-behavioral therapy
skillfiilness of counselors' last training case. c These are raters' ratings of psychotherapy technology tools—-manuals, videotape monitoring,
the two summary items of the IMPACT Tape Rating Scale for the second and supervision—as a means of disseminating science-based treat-
session of each counselors' final training case. ments to the substance abuse practice community. Several issues
 MORGENSTERN, MORGAN, McCRADY, KELLER, AND CARROLL

require consideration in planning future efforts to capitalize on D., Litt, M., & Hester, R. (1992). Cognitive-behavioral coping skills
manual-guided training as a dissemination device. Materials de- therapy manual: A clinical research guide for therapists treating indi-
signed to train expert therapists in clinical trials are inadequate to viduals with alcohol abuse and dependence (NIAAA Project MATCH
train front-line counselors. Further work is needed to develop Monograph, Vol. 3, DHHS Publication No. ADM 92-1895). Washing-
better training methods. In addition, work is needed to adapt ton, DC: U.S. Government Printing Office.
McCrady, B. S. (1994). Alcoholics Anonymous and behavior therapy: Can
protocols designed for controlled-trials research to clinical practice
habits be treated as diseases? Can diseases be treated as habits? Journal
settings. For example, greater implementation flexibility is needed
of Consulting and Clinical Psychology, 62, 1159-1166.
in clinical settings to handle such issues as when to end treatment, Morgenstem, J., & McCrady, B. (1992). Curative factors in alcohol and
what to do with clients who are not responding to the protocol, or drug treatment: Behavioral and disease model perspectives. British Jour-
the advisability of combining or sequencing treatments. Finally, nal of Addiction, 87, 615-626.
further study is needed to determine the optimum amount and Moyers, T. B., & Miller, W. R. (1993). Therapists' conceptualizations of
types of training experiences counselors need to learn to incorpo- alcoholism: Measurement and implications for treatment decisions. Psy-
rate new techniques into routine practice. chology of Addictive Behaviors, 7, 238—245.
Project MATCH Research Group. (1997). Matching alcoholism treatments
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

to client heterogeneity: Project MATCH posttreatment drinking out-

This document is copyrighted by the American Psychological Association or one of its allied publishers.

References comes. Journal of Studies on Alcohol, 58, 7-29.

Addis, M. (1997). Evaluating the treatment manual as a means of dissem- Sobell, L. C. (1996). Bridging the gap between scientists and practitioners:
inating empirically validated psychotherapies. Clinical Psychology, 4, The challenge before us. Behavior Therapy, 27(3), 297-320.
1-11. Stout, R. L., Wirtz, P. W., Carbonari, J. P., & Del Boca, F. K. (1994).
Barlow, D. H. (1996). Health care policy, psychotherapy research, and the Ensuring balanced distribution of prognostic factors in treatment out-
future of psychotherapy. American Psychologist, 51, 1050-1058. come research. Journal of Studies on Alcohol, 55(Suppl. 12), 70-75.
Carroll, K. M., Connors, G. J., Cooney, N. L., DiClemente, C. C., Dono- Strupp, H., & Anderson, T. (1997). On the limitations of therapy manuals.
van, D. M., Kadden, R. R., Longabaugh, R. L., Rounsaville, B. J., Wirtz, Clinical Psychology, 4, 76-82.
P. W., & Zweben, A. (1998). Internal validity of Project MATCH Wallace, J. (1996). Theory of 12-step-oriented treatment. In F. Rotgers,
treatments: Discriminability and integrity. Journal of Consulting and D. S. Keller, & J. Morgenstem (Eds.), Treating substance abuse: Theory
Clinical Psychology, 66, 290-303. and technique (pp. 13-37). New York: Guilford Press.
Gordis, E. (1991). Linking research with practice: Common bonds, com-
mon progress. Alcohol Health & Research World, 15, 173-174.
Institute of Medicine. (1997). Managing managed care: Quality improve- Received May 7, 1999
ment in behavioral health. Washington, DC: National Academy Press. Revision received May 9, 2000
Kadden, R., Carroll, K. M., Donovan, D., Cooney, N., Monti, P., Abranms, Accepted May 10, 2000