Journal of Substance Abuse Treatment 47 (2014) 307–313Cont.docx

Journal of Substance Abuse Treatment 47 (2014) 307–313
Contents lists available at ScienceDirect
Journal of Substance Abuse Treatment
Regular articles
Predicting substance-abuse treatment providers' communication
with
clients about medication assisted treatment: A test of the
theories of
reasoned action and planned behavior☆
Anthony J. Roberto, Ph.D. a,⁎, Michael S. Shafer, Ph.D. b,
Jennifer Marmo, Ph.D. c
a Hugh Downs School of Human Communication at Arizona
State University
b School of Social Work and Center for Applied Behavioral
Health Policy at Arizona State University
c Department of Education, Arizona State University
a b s t r a c ta r t i c l e i n f o
☆ This paper was made possible by Cooperative Agree
from the Department of Health and Human Services,
Health Services Administration. The opinions expressed
those of the authors and no endorsement of the HHS or
⁎ Corresponding author. Tel.: +1 11 480 9654 111.
E-mail address: [email protected] (A.J. Rober
http://dx.doi.org/10.1016/j.jsat.2014.06.002

0740-5472/© 2014 Elsevier Inc. All rights reserved.
Article history:
Received 13 August 2013
Received in revised form 3 June 2014
Accepted 8 June 2014
Keywords:
Medicated assisted treatment (MAT)
Substance-abuse treatment providers
Theory of reasoned action
Theory of planned behavior
The purpose of this investigation is to determine if the theory of
reasoned action (TRA) and theory of planned
behavior (TPB) can retrospectively predict whether substance-
abuse treatment providers encourage their
clients to use medicated-assisted treatment (MAT) as part of
their treatment plan. Two-hundred and ten
substance-abuse treatment providers completed a survey
measuring attitudes, subjective norms, perceived
behavioral control, intentions, and behavior. Results indicate
that substance-abuse treatment providers have
very positive attitudes, neutral subjective norms, somewhat
positive perceived behavioral control, somewhat
positive intentions toward recommending MAT as part of their
clients' treatment plan, and were somewhat
likely to engage in the actual behavior. Further, the data fit both
the TRA and TPB, but with the TPB model
having better fit and predictive power for this target audience
and behavior. The theoretical and practical
implications for the developing messages for substance-abuse
treatment providers and other health-care
professionals who provide treatment to patients with substance
use disorders are discussed.
ment Number 1UR1TI024242
Substance Abuse and Menta
in this manuscript are strictly

SAMHSA is to be inferred.
to).
© 2014 Elsevier Inc. All rights reserved.
Great strides have been made in the past decade in the
efficacious
application of pharmacological intervention in the
detoxification,
treatment, and long-term sobriety of patients experiencing
alcohol
and illicit drug abuse. Medication-Assisted Treatment (MAT) is
a form of
pharmacotherapy and refers to the treatment for a substance use
disorder that includes a pharmacologic intervention as part of a
comprehensive substance abuse treatment plan.
Pharmacotherapeutic
interventions have been demonstrated efficacious in the
treatment of
opioid abuse (Knudsen, Ducharme, & Roman, 2007; Weiss et
al., 2011),
alcohol dependence (Chandreakekaran, Sivaprekash, &
Chitraleka,
2001), and cocaine dependence (Carroll et al., 2000). In spite of
the
growing evidence base, adoption and widespread
implementation of
MAT has lagged, hampered by a combination of structural,
financial, and
workforce related issues (Knudsen et al., 2007).
In contrast to other chronic health conditions, treatment of
substance use disorders remains largely a disease treated by
counselors, social workers and therapists through a network of
community based, non-medically-based treatment agencies.
Among
surveyed substance abuse treatment facilities, only one-third

report
l
provision of MAT (National Survey of Substance Abuse
Treatment
Services, 2008), while the vast majority of primary care
physicians
report little knowledge of, or attendance to, the treatment of
substance use disorders among their patients (Mark et al.;
2003).
Confounding this situation are long held social beliefs and
attitudes
regarding the use of medication to treat substance use disorders,
with
such beliefs often present among a sizeable group of the
professionals
serving as addiction providers who are themselves in recovery
(Institute of Medicine, 1995, 1997). As evidence of the efficacy
of
MAT continues to accumulate (Friedmann & Schwartz, 2012),
so does
the research related to providers' and clients' attitudes beliefs,
and
behaviors, regarding MAT (Forman, Bovassdo, & Woody, 2001;
Reickmann, Daley, Fuller, Thomas, & McCarty, 2007). In
general,
these studies report rather powerful social normative influences
mediating what might best be described as neutral to negative
attitudes toward MAT.
Little research exists that explores effective strategies for
impacting
these attitudes and the corresponding behavioral intentions that
providers might have about discussing MAT with their clients.
Evidence-based targeted communications and information for
pro-
viders are needed to facilitate improved openness to MAT

efficacy, along
with their own professional efficacy in promoting and
integrating MAT
as part of the treatment and recovery services they provide to
their
patients. Given the potentially important role previous research
seems
to assign to attitudes, norms, and efficacy in this area, the
theories of
http://crossmark.crossref.org/dialog/?doi=10.1016/j.jsat.2014.06
.002&domain=pdf
mailto:[email protected]
http://www.sciencedirect.com/science/journal/07405472
308 A.J. Roberto et al. / Journal of Substance Abuse Treatment
47 (2014) 307–313
reasoned action and planned behavior were selected to guide
this
inquiry. A discussion of each of these theories follows.
1. The theory of reasoned action and the theory of
planned behavior
According to the theory of reasoned action (TRA; Ajzen &
Fishbein, 1980; Fishbein & Ajzen, 1975), the best predictor of a
person's behavior is their intention to perform or not perform
the
behavior, and the best predictors of intention are a person's
attitude
toward the behavior (i.e., do they feel positively or negatively
toward
the behavior) and subjective norms (i.e., how they think

significant
others think they should behave). The theory of planned
behavior
(TPB; Ajzen, 1985) adds a direct link from perceived behavioral
control
(i.e., how much influence the person has over the behavior) to
both
intention and behavior. Notably, the TPB “was made necessary
by the
original model's limitations in dealing with behaviors over
which
people have incomplete volitional control” (Ajzen, p. 181).
Thus,
Ajzen predicts there should be less difference between the TRA
and
TPB when the behavior in question is under volitional control.
Many
factors affect whether someone perceives a behavior under their
volitional control, such as time, money, skills, cooperation of
others,
etc. A visual representation of the TPB is included in Fig. 1.
Meta-
analyses by Albarracin, Johnson, Fishbein, and Muellerleile
(2001)
and Downs and Hausenblas (2005) offer consistent support for
the
ability of these theories to predict behavior.
While the TRA and TPB are typically used to predict how likely
an
individual is to engage in a given healthy behavior themselves,
research
also suggests that they can be used to explain recommendations
made
to patients by medical practitioners (Millstein, 1996; Perkins et
al.,

2007; Roberto, Goodall, West, & Mahan, 2010; Taylor,
Montano, &
Koepsell, 1994; Walker, Grimshaw, & Armstrong, 2001). For
example,
Millstein (1996) found that both the TRA and TPB accurately
predicted
primary care physicians' intentions and behavior to provide STI
education to adolescents. However, it should be noted that most
of
these studies took place more than a decade ago, focused on
physicians,
and did not include any sort of behavioral measure (i.e., the
majority
focused on intentions rather than actual behavior). Further, the
question
remains if the TPB is generalizable to other health professionals
such as
substance-abuse treatmentproviders. So, it seemsthere is still a
need for
more current research in this area using different participants,
an
additional topic, and a behavioral measure.
Among other things, Reickmann et al. (2007) used the TRA to
predict substance abuse treatment counselor's intentions to tell
their
patients to use each of four different types of MAT (methadone,
buprenorphine, clonidine, and ibogaine). Results indicate that
attitudes and norms explained between 40 and 71% in intentions
in
Attitudes:
Positive or negative
evaluation of the
behavior.
Beha

What
Perceived Behavioral
Control: Perceived
ease or difficulty of
adopting behavior.
Subjective Norms:
What you think others
think you should do.
Fig. 1. The theory of reasoned action (Ajzen & Fishbein, 1980;
Fishbein & Ajzen, 1975) and th
reasoned action. The entire figure with shaded box shows the
theory of planned behavior.
these instances. Similarly, Kelly, Deane, and Lovett (2012)
looked at
whether the TPB accurately predicted if residential substance
abuse
workers would make an effort to employ evidence-based
practices
(EBP) into their treatment of clients. In this study, EPB were
defined
as, “an approach which integrates the most appropriate clinical
information and scientific evidence, with a view to improving
psychological interventions and therapeutic relationships, and
pro-
ducing the best treatment outcomes for clients” (p. 662). Results
indicate that attitude, norms, and perceived behavioral control
explained 41% of the variance in intentions to use EBP.
Notably,
neither of these studies included a behavioral measure.
In sum, though previous applications of the TRA and TPB in the
health arena have focused primarily on predicting whether
individuals engage in healthy behaviors, work by Perkins et al.

(2007) suggests that they should provide a solid theoretical
framework for health professionals behavior in general, and
Millstein (1996), Reickmann et al. (2007), and Kelly,
Thompson,
and Waters (2006) suggest they might also predict health
professionals communication with patients in particular. Thus,
the
following research questions and hypothesis are advanced:
RQ1: What are substance-abuse treatment providers' attitudes,
subjective norms, perceived behavioral control, intentions, and
behavior regarding recommending medication-assisted treatment
as
part of their clients' treatment plan?
H1A-B: The (A) TRA and the (B) TPB will accurately predict
whether or
not substance-abuse treatment providers encouraged their
clients to
use medication-assisted treatment as part of their treatment
plan.
RQ2: Does the TPB add to the predictive power of the TRA for
this
target audience and behavior?
2. Method
2.1. Response rate and research participants
2.1.1. Response rate
A link to the survey was sent via email to all 510 individuals
who
were (1) subscribers to an e-newsletter distributed by the
Addiction
Technology Transfer Center(s) (ATTC), and (2) who identified
them-
selves as serving in a clinical/direct service role in the

provision of
substance abuse treatment as counselors, clinical supervisors, or
peer
recovery specialists. Twenty-eight of these surveys were
returned as
undeliverable. Response rate was calculated as the number of
surveys
returned (n = 210) divided by the number of surveys that were
sent
out and not returned asundeliverable (n = 510 − 28 = 482).
Thus, the
final response rate is 43.57%.
vioral Intention:
you plan to do.
Behavior:
What you actually do.
e theory of planned behavior (Azjen, 1991). Note: Non-shaded
boxes show the theory of
309A.J. Roberto et al. / Journal of Substance Abuse Treatment
47 (2014) 307–313
2.1.2. Research participants
Participants were 210 substance-abuse treatment providers
reporting an average age of 48 (range = 26 to 76; SD = 11.11)
and 14 years of substance abuse treatment experience (M =
13.84,
SD = 9.37). In the 30 days immediately preceding the survey,
these
respondents reported seeing a median of 39 clients (M = 64.34;
SD = 104.90). Additional descriptive statistics are provided in
Table 1. Taken together, these descriptive statistics suggest that

respondents had regular, frequent, and intensive interaction with
substance abusing clients. Finally, using the first digit from the
ZIP
code from the agency for which the participants worked, it was
possible to determine that participants from all 10 of the U.S.
Postal
Service's general regions of the country completed a survey
(range = 5 to 18% per region, M = 10.9% per region).
2.2. Instrumentation
All TRA and TPB measures were developed using procedures
outlined by Ajzen and Fishbein (1980) and Madden, Ellen, and
Ajzen
(1992); and are similar to items developed by Reickmann et al.
(2007)
and Kelly et al. (2006). Participants were provided with
instructions
and a definition of MAT adapted from SAMHSA (2010) before
being
Table 1
Participant demographics.
Variable %
Sex
Male 35.8
Female 64.2
Ethnicity
Hispanic or Latino/a 9.0
European-American 81.9
African-American 7.8
Native American 5.4
Asian 0.5
Other 4.2

In recovery
Yes 46.0
No 54.0
Location of work
Outpatient treatment facility 63.7
Residential treatment facility 19.6
Correction/criminal justice program 15.2
Hospital/medical facility program 7.4
Other 20.6
Core work functions
Assessing clients 79.4
Developing treatment plans 74.0
Providing individual counseling 77.5
Providing group counseling 66.7
Provide case management 63.2
Medication-assisted treatment offered
Yes, MAT is provided on-site 30.9
Yes, but in partnership with a physician/group 16.2
No 50.5
MAT organizational support level
Very unsupportive 12.3
Unsupportive 10.9
Neutral 26.7
Supportive 29.2
Very supportive 20.8
Workshops/training about use of MAT to treat substance abuse
Yes 88.2
No 11.8
Self-rating knowledge level of MAT

Very low 2.0
Low 20.1
Moderate 37.7
High 27.5
Very high 12.7
Interest in participation in training using MAT
Yes 79.9
No 17.2
prompted to complete a series of forced-choice questions. The
definition read, “This survey asks questions about medication-
assisted
treatment (sometimes referred to as MAT). For the purposes of
this
survey, medication-assisted treatment is defined as the use of
medications such as suboxone, clonidine, and methadone in
combi-
nation with counseling and behavioral therapies to provide
treatment
of substance-use disorders.”
Behavior was assessed with two questions. First, participants
were
asked, “Do you ever talk to your clients about using medication-
assisted
treatment as part of their treatment plan?” Response categories
were
“no” and “yes”. Those who answered “no” were coded as not
engaging in
the behavior (i.e., engaging in the behavior “0% of the time”).
Those who
answered “yes” were asked the following contingency question,
“In the
past 6 months, approximately what percentage of your clients
have
you spoken to about using medication-assisted treatment as part

of
their treatment plan?” Response categories for this five-
pointitem were,
“1–20% of clients, 21–40% of clients, 41–60% of clients, 61–
80% of clients,
and 81–100% of clients.” Intandem, these twoitemswere
combined into
a single six-point behavioral measure ranging from “0% of
clients” to
“81–100% of clients.”
Behavioral intention [e.g., “I (intend to/plan to) encourage my
clients
to use medication-assisted treatment as part of their treatment
plans in
the future.”] and subjective norms [e.g., “Most colleagues who
are
important to me (think that I should/want me to) encourage my
clients
to use medication-assisted treatment as part of their treatment
plan.”]
were each assessed with two items. Perceived behavioral control
was
assessed using three items (e.g., “I am able to effectively
encourage my
clients to use medication-assisted treatment as part of their
treatment
plan.”/“I am capable of effectively encourage my clients to use
medication-assisted treatment as part of their treatment
plan.”/“It is
easy for me to effectively encourage my clients to use
medication-
assisted treatment as part of their treatment plan.”). Response
categories for these three sets of items ranged from 1 (“strongly
disagree”) to 5 (“strongly agree”). Finally, attitude was assessed
by

asking, “To me, encouraging my clients to use medication-
assisted
treatment as part of their treatment plan is:” followed by three
five-
point semantic differential items (i.e., “bad–good,” “harmful–
helpful,”
and “useless–useful”). Alphas for the multi-item measures
ranged from
.82 to .93 (note: individual alphas and mean item scores for all
TRA and
TPB measures are included in Table 2).
The design of this study and the development of the instrument
were preceded by a qualitative study involving focus groups of
clients
receiving medication assisted treatment (Malvini-Redden,
Tracy, &
Shafer, 2013). Key concepts that emerged from that study,
elucidating
clients' perspectives on the value and challenges of using MAT,
provided general constructs for this study. Instrumentation
followed
an iterative process and included review by a national panel of
colleagues from the ATTC network and a small pilot study with
substance abuse counselors (n = 5). These counselors completed
a
draft version of the instrument and provided verbal feedback
with
regard to the clarity and comprehensiveness of the items and
response options. Based upon their feedback a number of
revisions
were made to the instrument before the final version was imple-
mented for this study.
2.3. Procedures

Dillman, Smyth, and Christian's (2009) tailored-design method
was used to guide all data collection procedures, and data
collection
was conducted by the Institute for Social Science Research
(ISSR), an
organization dedicated to providing a variety of research and
data
collection services for the sponsoring University and the
surrounding
community. A consent form and link to the online survey were
distributed by email. Each participant was contacted by e-mail
up to
three times over a 4-week week period to encourage survey
completion (though once a participant completed the survey,
they
Table 2
Reliability, descriptive statistics, and zero-order correlations for
all measured variables.
α M SD 1 2 3 4 5
1. Attitude toward encouraging
(3 items)
.93 4.05 .81 –
2. Social norms toward encouraging
(2 items)
.90 2.92 1.04 .52⁎ –
3. Perceived behavioral control
toward encouraging (3 items)

.82 3.64 .88 .47⁎ .52⁎ –
4. Behavioral intention to encourage
(2 items)
.88 3.50 .97 .74⁎ .66⁎ .61⁎ –
5. Behavior (1 composite item) NA 1.72 1.53 .39⁎ .38⁎ .40⁎
.42⁎ –
Notes. All variables measured on a 5-point scale, except
behavior which was measured
on a 6-point scale. Correlations based on one-tailed probability
estimates.
⁎ p b .001.
47 (2014) 307–313
did not receive subsequent mailings). Individuals who
completed the
survey before the end date received a $10 gift card to
Amazon.com.
The research procedures were reviewed and approved as exempt
status by the sponsoring University's Office of Research
Integrity
and Assurances.
3. Results
3.1. Data analytic plan
Structural equation modeling (SEM) was used to test the
hypoth-
esized relationship between the TRA and TPB variables. A path
analysis

was conducted using EQS 6.1 software (Bentler, 1995). The
data were
normal (Mardia's PK = 0.13) allowing for the maximum
likelihood
estimation method to be used. Model fit was considered
acceptable
upon meeting the following conditions: (a) a non-significant
chi-square
(Jöreskog & Sörbom, 1993)—a perfect connection between
theory and
the study data would yield a χ2 of zero (Bollen, 1989), (b) a
comparative
fit index (CFI) greater than .95 (Bentler, 1995), (c) root mean-
square-
error of approximation (RMSEA) less than .10 (Bentler &
Bonnett, 1980),
and (d) standardized root mean square residual (SRMR) less
than .05
(Bentler & Bonnett, 1980). R2 is examined for each dependent
construct
to assess predictive power.
3.2. Descriptive statistics
Table 2 provides the alphas, means, and standard deviations for
all
measured variables, as well as the zero-order correlations
among all
measured variables. In answer to research question 1 (and
shown in
Table 2), substance-abuse treatment providers had very positive
attitudes, neutral subjective norms, somewhat positive
perceived
behavioral control, and somewhat positive intentions toward
recom-
mending MAT as part of their clients' treatment plan, but tended

to
engage in the actual behavior less than 20% of the time.
Correlations
demonstrate predicted theoretical relationships at the univariate
level. Consistent with the TRA and TPB, attitudes, social
norms, and
perceived behavioral control were positively and significantly
related
to behavioral intentions. In addition, as predicted by the TRA
and TPB,
Attitudes:
evaluation of the
behavior.
Behavio
What yo
Subjective Norms:
.55
.37
.51
Fig. 2. Path model for TRA (H1A). χ
2(6) = 101.28, p b .001
both behavioral intentions and perceived behavioral control
signifi-
cantly correlated in the expected direction with behavior.
3.3. Measurement model

Hypothesis 1 stated that (A) the TRA and (B) the TPB would
accurately predict whether substance-abuse treatment providers
en-
courage their clients to use medication-assisted treatment as
part of their
treatment plan. Specifically, attitudes, subjective norms,
behavioral
intentions, and behavior were analyzed as measured variables
for the
TRA (see Fig. 1). Twenty participants were excluded from the
analysis
due to missing data for at least one variable, leaving 184
participants to
test whether the model had adequate fit. Table 2 shows the
significant
and substantial positive relationships between intentions and
behavior,
attitudes and intentions, and subjective norms and intentions.
To test the
hypothesis regarding the overall fit of the TRA, these
correlations were
then used to compute the path coefficients in the hypothesized
TRA path
model. All three predicted paths were of sufficient size and
achieved
standard levels of statistical significance. The path coefficient
between
attitude and behavioral intention was substantial, β = .55,
p (.45 ≤ β ≤ .65) = .95. The coefficient between social norms
and
behavioral intention was moderate and statistically significant,
β = .37,
p (.27 ≤ β ≤ .47) = .95. The coefficient between behavioral
intentions and behaviors was moderate and significant, r = .40,
p (.30 ≤ r ≤ .50) = .95. The overall model fit, however, was only
adequate, χ2(6) = 101.28, p b .001, CFI = .74, SRMR = .26, and

RMSEA = .30CI = .24–.34. Fig. 2 displays the structural model
parameters,
as well as the amount of explained variance in intentions (R2 =
64.6%)
and behavior (R2 = 15.8%).
The TRA and the TPB share all of the same variables with only
the
addition of perceived behavioral control in the TPB. The
inclusion of
perceived behavioral control in the model brings with it two
additional
paths: one from perceived behavioral control to behavioral
intent and
another from perceived behavioral control to behavior (see Fig.
1). Both
paths were predicted to be positive such that perceived
behavioral
control should increase both behavioral intent and behavior.
To test the overall fit of the TPB, the correlations from Table 2
were
used to compute the path coefficients in the hypothesized TPB
path
model. All of the TPB predictions that overlapped with the TRA
were
again supported. The path coefficient between attitude and
behavioral
intention, β = .48, p (.29 ≤ β ≤ .67) = .95, the coefficient
between
social norms and behavioral intention, β = .29, p (.10 ≤ β ≤ .48)
= .95,
and the coefficient between behavioral intentions and behaviors,
r = .28, p (.19 ≤ r ≤ .47) = .95, were each moderate to
substantial.
The addition of perceived behavioral control, however, did

significantly
change the overall fit of the model: perceived behavioral control
did have
a moderate effect on behavioral intent, β = .23, p (.04 ≤ β ≤ .42)
= .95
and behavior, β = .20, p (.01 ≤ β ≤ .39) = .95. Moreover, the
hypothesized TPB model suggested excellent fit, χ2(2) = 4.88,
p = .09, CFI = .99, SRMR = .03, and RMSEA = .09CI = .00–.19.
Fig. 3
displays the structural model parameters, as well as the amount
of
explained variance in intentions (R2 = 68.2%) and behavior
(R2 = 18.3%). These findings suggest the addition of the direct
path
between perceived behavioral control and intentions, and
perceived
ral Intention:
u plan to do.
Behavior:
.40
, CFI = .74, SRMR = .26, and RMSEA = .30CI = .24–.34.
mailto:[email protected]
Attitudes:
evaluation of the
behavior.
Behavioral Intention:
What you plan to do.

Perceived Behavioral
Control: Perceived
ease or difficulty of
adopting behavior.
Behavior:
Subjective Norms:
.51
.52
.48
.29
.23
.28
.20
.47
Fig. 3. Path model for TPB (H1B). χ
2(2) = 4.88, p = .09, CFI = .99, SRMR = .03, and RMSEA =
.09CI = .00–.19.
47 (2014) 307–313
behavioral control and behavior. Thus, in support of H1, tests of

both
models supported the prediction of whether substance-abuse
treatment
providers encouraged their clients to use medication-assisted
treatment
as part of their treatment plan; however, the TPB suggested a
stronger fit.
In response to RQ2, two analyses were conducted. First, when
poor
model fit exists, respecification can occur. Both the Wald test
(WT)
and Lagrange multiplier test (LMT) were analyzed. The WT
deter-
mines if parameters should be dropped from the model, whereas
the
LMT frees parameters. Regardless of which test is utilized,
Loehlin
(1992) emphasized that caution must be made when
respecification
of a model occurs; modifications should only occur if they are
theoretically defensible or consistent with a substantial body of
literature. For this study, the WT revealed no options to
improve the
model by dropping parameters nor is that option theoretically-
sound.
The LMT, however, suggested that the changes to be made to
the
model in order to achieve excellent fit were the inclusion of
perceived
behavioral control to intention, the covarying of attitudes,
subjective
norms, and perceived behavioral control, and finally the
direction
correlation of perceived behavioral control with behavior. This
suggested final model was identical to that of TPB. Second, a

step-
wise regression that in the first step regressed intentions on the
core
TRA variables and in the second step perceived behavioral
control
showed a small, but significant increase in the predictive power
of the
model including perceived behavioral control, R2 change =
.038,
p b .001. Thus, in answer to RQ2 the TPB adds a small but
significant
amount of predictive power for this target audience and
behavior.1
1 We were also interested in determining if there were any
differences between
substance-abuse treatment providers who self-identified as a
person in recovery and
those who did not on both the means of the TPB variables, and
in the fit of the final TPB
model. The substance abuse treatment workforce has
historically consisted of
individuals in recovery and the recent emphasis on Recovery
Oriented Systems of
Care (ROSC) places a growing emphasis on the incorporation of
people in recovery
within this workforce. These individuals, many of whom were
treated before the
emergence of MAT as an evidence based practice, could be
expected to have negatively
biased perspectives regarding MAT. A series of independent-
sample t tests revealed
significant difference between these two groups on two of the
TRA/TPB variables.
Specifically, those who were in recovery (M = 2.71, SD = 1.06)
perceived signifi-
cantly lower norms to encourage clients to use MAT as part of

their treatment plan
than those who were not in recovery (M = 3.06, SD = 1.01), t
(184) = -2.29,
p b .05). Further, those who were in recovery (M = 3.31 SD =
1.04) reported significantly
lower intentions to encourage clients to use MAT as part of
their treatment plan than those
who were not in recovery (M = 3.06, SD = 1.01), t (188) = -
2.55, p b .05). Given these
two differences, we also ran the final TPB model separately for
those in and not in recovery.
This did not substantially change the fit of the model.
Specifically, the fit indices for the final
TPB model (as indicated in the main text and in Fig. 3) were,
χ2(2) = 4.88, p = .09,
CFI = .99, SRMR = .03, and RMSEA = .09CI = .00-.19 (with
68.2% of the variance in
intention and 18.3% of the variance in behavior explained by
the model). Whereas the fit
indices for those in recovery were, χ2(3) = 6.12, p = .13, CFI =
.98, SRMR = .04, and
RMSEA = .11CI = .00-.24 (with 66% of the variance in
intentions and 18% of the variance in
behavior explained by the model), and the fit indices for those
not in recovery were
χ2(3) = 5.69, p = .10, CFI = .99, SRMR = .05, and RMSEA =
.10CI = .00-.22 (with
72.6% of the variance in intentions and behavior 14.5% of the
variance in behavior
explained by the model).
4. Discussion
The main goal of this study was to see if the TRA and TPB
accurately
predicted whether substance-abuse treatment providers
encouraged

their clients to use MAT as part of their treatment plan. A
survey
measuring all TRA and TPB variables was sent to 510
substance-
abuse treatment providers, and 210 (43.6%) of these providers
completed this survey. Results indicated that the data fit both
models for this target audience and behavior, and that the
TPB added to the explanatory power on encouraging clients to
use the MAT.
One important outcome of the present investigation is that it
provides a list of factors that influence substance-abuse
treatment
providers' recommendations about the use of MAT. Since
substance-
abuse treatment providers' recommendations likely influence the
decisions their clients make, they play a particularly important
role in
the use and success of MAT. The current investigation
identifies
important concepts to integrate into health communication and
community-based interventions targeting substance-abuse
treatment
providers. These results suggest that interventions targeting
substance-
abuse treatment providers would be effective if organized along
the
constructs of the TPB. For example, an intervention might
attempt to
reinforce already existing positive attitudes toward MAT, or
increase
subjective norms and perceived behavioral control that currently
hover
around neutral to be more positive. Given that each substance-
abuse
treatment provider works with a large number of clients,

interventions
targeting providers could have a much greater impact on more
individuals than those just targeting individual clients.
These results confirm and extend the findings of Reickmann et
al.
(2007) and Kelly et al. (2006). Consistent with both sets of
findings,
these results confirm the applicability of TRA as a conceptual
model
for explaining counselor's attitudes and intentions, and linking
the
influence that social norms have upon both. Extending these
results,
these findings also support the small but important influence
that
counselor's perceived behavioral control plays in their
intentions,
suggesting counselors might see this behavior as somewhat but
not
completely under their control. The current study also included
a
behavioral measure while neither Rieckmann et al. nor Kelly et
al. did
not. It is worth noting that while 80% of the participants
reported
previous training about MAT, an equivalent proportion also
indicated
a desire for additional training. As such, these findings
underscore the
importance of providing substance abuse providers with
accurate
information and skill building opportunities to enhance their
effectiveness in counseling clients to consider the use of MAT
in
addition to information about the physiological properties of

MAT.
While lacking direct evidence, these results could reflect
providers'
unease in their personal effectiveness to promote and to induce
their
clients to make use of MAT. Skill building opportunities for
providers
that focus on the use of motivational interviewing, and other
strategies that address clients' ambivalence in using MAT could
47 (2014) 307–313
provide critical influence in facilitating the broader adoption
and
implementation of this evidence-based practice.
4.1. Strengths and limitations
A key strength of this study is that it is theory-based and
extends the
scope of the TRA and TPB to a topic (i.e., substance-abuse
prevention)
and target audience (i.e. substance-abuse treatment providers).
Second,
our survey was designed using procedures outlined by Ajzen
and
Fishbein (1980) and Madden et al. (1992), which, in tandem
with the
high alphas obtained in the present study, both provides a high
level of
confidence in our measures and allows our results to more
accurately be
compared to other studies. Also for example, the data were
collected

using Dillman et al.'s (2009) tailored design method, which was
specifically developed to reduce nonresponse error (e.g., by
increasing
participants' motivation to respond) and measurement error
(e.g., by
helping respondents provide more complete, accurate, and
precise
answers). Further, path analysis was conducted using well-
established
procedures (Bentler, 1995; Bentler & Bonnett, 1980; Bollen,
1989;
Jöreskog & Sörbom, 1993), and with a satisfactory sample size
for this
type of analysis. Third, we had a relatively large national
sample,
especially given that substance-abuse treatment providers'
communi-
cative behaviors were being studied. Finally, as noted above,
these
results have important practical implications.
As with any investigations, some potential limitations must also
be
acknowledged. The main limitation is that the intention–
behavior link
was measured retrospectively (i.e., both were measured at the
same
time as opposed to measuring intentions at one time and
behavior at
some later time). While not ideal, this is a common and
accepted
practice in TRA and TPB research, especially when using health
care
providers as research participants (Albarracin et al., 2001;
Perkins
et al., 2007). However, now that the TRA and TPB have been

shown
to be relevant to this topic and target audience, future research
should
be conducted where the measure of intentions precedes the
measure
of behavior.
A second limitation is that this study did not include antecedent
measures attitudes, subjective norms, or perceived behavioral
control.
For example, these theories suggest that (1) behavioral beliefs
and
outcome evaluations should predict attitudes, (2) normative
beliefs
and motivation to comply should predict norms, and (3) self-
efficacy
and controllability should predict perceived behavioral control.
Since
only a smaller proportion of TRA and TPB research includes
measures
of these antecedent variables, there is no doubt that
understanding
the factors that underlie these three variables would provide
valuable
information for both theoretical and practical reasons.
Of course, the TRA and TPB also have limitations of their own.
For
example, the TRA is designed to explain behaviors that are
under a
person's volitional control. Though the TPB was designed to
address
this issue to some extent by adding perceived behavioral
control,
other variables likely also play a role, either directly or
indirectly, in

such decisions. To illustrate, the behavior ecological models
(Hovell,
Wahlgren, & Gehrman, 2002) include other variables that might
influence behavior at numerous levels. The TRA and TPB take
into
account many key variables at the individual (such as attitudes
and
skills) and interpersonal (such as norms) levels, but neither
explicitly includes variables that might affect behavior at the
organization, community, or public policy levels. Another
limitation
is that both theories assume humans are rational decision
makers,
and will only be effective to the extent that this is true for the
behavior under investigation.
5. Conclusions
The results of this study provide important new information to
facilitate the adoption of MAT and extend our knowledge about
implementing evidence-based practices in substance abuse
treatment
settings. Our results suggest that developing theory-based
interventions
using TRA or TPB should be effective in targeting substance
abuse
treatment providers' communications with their clients about
innova-
tive, evidence-based treatment strategies, such as MAT. Future
studies
designed to change substance-abuse treatment providers'
behavior can
built upon these findings by testing the relative influence that
knowledge dissemination and skill building strategies in
combination
with promotional and communication strategies has upon

provider's
behavior and behavioral intentions regarding their client
communica-
tions on MAT or other evidence-based treatment innovations.
References
Ajzen, I. (1985). From intentions to actions: A theory of
planned behavior. In J. Kuhl, &
J. Beckman (Eds.), Action control: From cognition to behavior
(pp. 11–39). Berlin:
Springer-Verlag.
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and
predicting social behavior.
Englewood Cliffs, NJ: Prentice-Hall.
Albarracin, D., Johnson, B. T., Fishbein, M., & Muellerleile, P.
A. (2001). Theories of
reasoned action and planned behavior as models of condom use:
A meta-analysis.
Psychological Bulletin, 127, 142–161.
Bentler, P. M. (1995). EQS structural equations program
manual. Encino, CA: Multivariate
Software, Inc.
Bentler, P. M., & Bonnett, D. G. (1980). Significance tests and
goodness of fit in the
analysis of covariance structures. Psychological Bulletin, 88,
586–606.
Bollen, K. A. (1989). Structural equations with latent variables.
New York: Wiley.
Carroll, K. M., Nich, C., Ball, S. A., McCance, E., Franforter,
T. L., & Rounsaville, B. J. (2000).

One year follow-up of disulfiram and psychotherapy for
cocaine-alcohol uses:
Sustained effects of treatment. Addiction, 95, 1335–1349.
Chandreakekaran, R., Sivaprekash, B., & Chitraleka, V. (2001).
Five years of alcohol de-
addiction services in tertiary care general hospital. Indian
Journal of Psychiatry, 43, 58–60.
Dillman, D. A., Smyth, J. D., & Christian, L. M. (2009).
Internet, mail, and mixed-mode
surveys: The tailored design method (3rd ed.). New York:
Wiley.
Downs, D. S., & Hausenblas, H. A. (2005). The theories of
reasoned action and planned
behavior applied to exercise: A meta-analytic update. Journal of
Physical Activity
and Health, 2, 76–97.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and
behavior. Reading, MA:
Addison-Wesley.
Forman, R. F., Bovassdo, G., & Woody, G. (2001). Staff beliefs
about addiction treatment.
Journal of Substance Abuse Treatment, 21, 1–9.
Friedmann, P. D., & Schwartz, R. P. (2012). Just call it
“treatment”. Addiction Science &
Clinical Practice, 7, 10.
Hovell, M. F., Wahlgren, D. R., & Gehrman, C. A. (2002). The
behavioral ecological
model: Integrating public health and behavioral science. In R. J.

DiClemente, R. A.
Crosby, & M. C. Kegler (Eds.), Emerging theories in health
promotion practice and
research: Strategies for improving public health (pp. 347–385).
San Francisco, CA:
Jossey-Bass.
Institute of Medicine (1995). The development of medications
for the treatment of opiate
and cocaine addictions: Issues for the government and private
sector. Washington, DC:
National Academy Press.
Institute of Medicine (1997). Dispelling the myths about
addiction: Strategies to increase
understanding and strengthen research. Washington, DC:
National Academy Press.
Jöreskog, K. C., & Sörbom, D. (1993). LISREL 8: Structural
equation modeling with the
SIMPLIS command language. Hillsdale, NJ: Lawrence Erlbaum
Associates, Inc.
Kelly, K. S., Thompson, M. F., & Waters, R. D. (2006).
Improving the way we die: A
coorientation study assessing agreement/disagreement in the
organization–public
relationship of hospices and physicians. Journal of Health
Communication, 11,
607–627.
Kelly, P. J., Deane, F. P., & Lovett, M. (2012). Using the theory
of planned behavior to
examine residential substance abuse workers’ intention to use
evidence-based
practices. Psychology of Addictive Behaviours, 26, 661–664.

Knudsen, H. K., Ducharme, L. J., & Roman, P. M. (2007). The
adoption of medications in
substance abuse treatment: Associations with organizational
characteristics and
technology clusters. Drug & Alcohol Dependence, 16, 164–174.
Loehlin, J. C. (1992). Latent variable models: An introduction
to factor, path, and structural
analysis. Hillsdale, NJ: Lawrence Erlbaum Associates.
Madden, T. J., Ellen, P. S., & Ajzen, I. (1992). A comparison of
the theory of planned
behavior and the theory of reasoned action. Personality and
Social Psychology
Bulletin, 18, 3–9.
Malvini-Redden, S., Tracy, S. J., & Shafer, M. S. (2013). A
metaphor analysis of recovering
substance abusers' sense making of medication-assisted
treatment. Qualitative
Health Research, 23, 951–962.
Mark, T. L., Kranzler, H. R., Song, X., Bransberger, P., Poole,
V. H., & Crosse, S. (2003). Physicians'
opinions about medications to treatment alcoholism. Addiction,
98, 617–626.
Millstein, S. G. (1996). Utility of the theories of reasoned
action and planned behavior for
predicting physician behavior. A prospective analysis. Health
Psychology, 15, 398–402.
National Survey of Substance Abuse Treatment Services (N-
SSATS) (2008). Data on
substance abuse treatment facilities. DASIS Series: S-49, HHS

Publication No. (SMA)
09-4451, Rockville, MD.
Perkins, M. B., Jensen, P. S., Jaccard, J., Gollwitzer, P.,
Oettingen, G., Pappadopulos, E.,
et al. (2007). Applying theory-driven approaches to
understanding and modifying
clinicians' behavior: What do we know? Psychiatric Services,
58, 342–348.
http://refhub.elsevier.com/S0740-5472(14)00095-6/rf0005

47 (2014) 307–313
Reickmann, T., Daley, M., Fuller, B. E., Thomas, C. P., &
McCarty, D. (2007). Client and
counselor attitudes toward the use of medications for treatment
of opioid
dependence. Journal of Substance Abuse Treatment, 32, 207–
215.
Roberto, A. J., Goodall, C. E., West, P., & Mahan, J. D. (2010).
Persuading physicians to test
their patients' level of kidney functioning: The effects of
framing and point of view.
Health Communication, 25, 107–118.
Substance Abuse & Mental Health Services Administration
(SAMHSA) (2010). About
medicated-assisted treatment. Retrieved July 26, 2014 from.
http://www.dpt.
samhsa.gov/patients/mat.aspx
Taylor, V. M., Montano, D. E., & Koepsell, T. (1994). Use of
screening mammography by
general internists. Cancer Detection and Prevention, 18, 455–
462.
Walker, A. E., Grimshaw, J. M., & Armstrong, E. M. (2001).
Salient beliefs and intentions
to prescribe antibotics for patients with a sore throat. British
Journal of Health

Psychology, 6, 347–360.
Weiss, R. D., Potter, J. S., Fiellin, D. A., Byrne, M., Connery,
H. S., Dickenson, W., et al.
(2011). Adjunctive counseling during brief and extended
buprenorphine–naloxone
treatment for prescription opioid dependence: A 2-phase
randomized controlled
trial. Archives of General Psychiatry, 68, 1238–1246.
http://www.dpt.samhsa.gov/patients/mat.aspx
http://www.dpt.samhsa.gov/patients/mat.aspx
http://refhub.elsevier.com/S0740-5472(14)00095-
6/rf0150Predicting substance-abuse treatment providers'
communication with clients about medication assisted
treatment: A test of t...1. The theory of reasoned action and the
theory of planned behavior2. Method2.1. Response rate and
research participants2.1.1. Response rate2.1.2. Research
participants2.2. Instrumentation2.3. Procedures3. Results3.1.
Data analytic plan3.2. Descriptive statistics3.3. Measurement
model4. Discussion4.1. Strengths and limitations5.
ConclusionsReferences

Journal of Substance Abuse Treatment 47 (2014) 307–313Cont.docx

Journal of Substance Abuse Treatment 47 (2014) 307–313Cont.docx

Recommended

Recommended

More Related Content

Similar to Journal of Substance Abuse Treatment 47 (2014) 307–313Cont.docx

Similar to Journal of Substance Abuse Treatment 47 (2014) 307–313Cont.docx (20)

More from priestmanmable

More from priestmanmable (20)

Recently uploaded

Recently uploaded (20)

Journal of Substance Abuse Treatment 47 (2014) 307–313Cont.docx