Provider General Attitudes versus Specific Perceptions of Evidence-Based Psychotherapies for PTSD

Abstract

This study examined the role of attitudes toward evidence-based psychotherapies (EBPs) in predicting use of Prolonged Exposure (PE) and Cognitive Processing Therapy (CPT), two EBPs for posttraumatic stress disorder (PTSD) among PTSD treatment providers within the Department of Veterans Affairs. Providers’ general attitudes toward EBPs, as well as their specific perceptions of PE and CPT, were examined as potential predictors of utilization. One hundred fifty-nine providers from 38 Department of Veterans Affairs’ residential PTSD programs across the United States completed an online survey that included the predictors listed as well as self-reported use of PE on an individual basis and CPT on an individual and on a group basis.. Although general attitudes toward EBPs were related to use of individually administered CPT, they were not related to use of PE or group-administered CPT. For each of the three treatments, however, specific positive perceptions were related to use. In examination of other training, skill, and delivery-related variables, general attitudes appear more in line with perceptions and delivery of CPT than PE. Perhaps this is because of the unique exposure component of PE. Assessing provider perceptions of specific EBPs may help providers in guiding their own practice as well as aid treatment developers, trainers and administrators to more effectively tailor dissemination and implementation efforts.

Posttraumatic stress disorder (PTSD) is one of the most common mental health diagnoses among veterans in the United States (Institute of Medicine, 2014). In order to promote access to high-quality PTSD treatment, the U.S. Department of Veterans Affairs (VA), beginning in 2006, instituted national initiatives to provide their mental health workforce training and consultation in two evidence-based psychotherapies (EBPs) for PTSD (Karlin et al., 2010): Prolonged Exposure (PE; Foa, Hembree, & Rothbaum, 2007) and Cognitive Processing Therapy (CPT; Resick, Monson & Chard, 2016). Despite the efficacy of PE and CPT, their implementation within the VA has been variable (Sayer et al., 2017). Better understanding provider adoption of these EBPs across the VA health care spectrum of services is important, not only to identify predictors of quality of care, but to ensure successful investment of time and financial resources.

One way to study providers’ adoption of such treatments is to apply theory-driven models. Decades of research, consistent with the theory of reasoned action and planned behavior as well as other behavioral science theories, indicate that specific, as opposed to general, attitudes are stronger predictors of specific behavior (Ajzen & Fishbein, 2005). For example, Oskamp (1991) demonstrated that attitudes about recycling predicted recycling behavior, but general attitudes about the environment did not influence recycling patterns. Similar findings have been demonstrated for a variety of health care issues, such as birth control pill use (Davidson & Jaccard, 1979) and condom use (Albarracin, Johnson, Fishbien, & Muellerleile, 2001).

An important area of research focus has been providers’ general attitudes toward EBPs (e.g., Jensen-Doss, Hawley, Lopez, & Osterberg, 2009; Stumpf, Higa-McMillian, & Chorpita, 2009). In particular, the Evidence-Based Practice Attitudes Scale (EBPAS; Aarons, 2004) was developed explicitly to assess providers’ general attitudes toward adopting EBPs. The subscales on the original EBPAS measure: (1) the extent to which a provider would adopt a new treatment because it was intuitively appealing, (2) likelihood of adopting EBT given requirements by an agency, supervisor or states to do so, (3) the extent to which a provider was generally open to trying new practices, and (4) the extent to which a provider perceives research-based interventions are not clinically useful and less important than clinical experience. Scores on the EBPAS predict self-reported use of EBPs in a variety of mental health settings (e.g., Lim, Nakamura, Higa-McMillan, Shimabukuro, & Slavin, 2012) across countries and cultures (Aarons, 2004; Aarons, McDonald, Sheehan, & Walrath-Greene, 2007; Rye et al., 2017).

There are preliminary data, however, to indicate that these general attitudes towards EBPs may operate differently from perceptions of specific EBPs. For example, in a study of community mental health providers that controlled for general attitudes towards EBPs (using the EBPAS), providers’ perceptions of EBPs’ appealing features were related to the degree to which they use it (Reding, Chorpita, Lau, & Innes-Gomberg, 2014). General attitudes toward EBPs were unrelated to self-reported EBP use, indicating that specific attitudes and beliefs may be an important component of provider perceptions.

One model for more comprehensively measuring providers’ specific perceptions of an intervention is Rogers’ (2003) seminal Diffusion of Innovations. Rogers proposed six main attributes important for adoption of a new product: relative advantage (is the innovation perceived as better in terms of efficacy, satisfaction, convenience and prestige?); compatibility (is the innovation consistent with adopters’ existing values, experiences and needs?); complexity (is the innovation easy to understand and use?); trialability (can the innovation be used on a limited or trial basis?); observability (are the innovations’ results visible to others?); and potential for reinvention (can the innovation be adapted, refined or modified to suit the needs of adopters?). Although hundreds of empirical studies have been conducted on the relationship between providers’ perceptions of innovations and their subsequent adoption, relatively few of these studies have been in in the mental health service domain or comparative across similar mental health settings (Greenhalgh et al., 2004), and none to our knowledge has applied this framework to EBPs for PTSD.

Findings from a recent, systematic review of 32 publications from 19 studies on predictors of implementation of PE and CPT in the VA health care system show how important provider beliefs about a treatment were in its adoption (Rosen et al., 2016). One of the most consistent predictors was provider beliefs about the effectiveness of these treatments. Providers’ use of CPT and PE was associated with the extent to which they believed that these two psychotherapies are effective, especially relative to other treatments. However, belief in effectiveness (e.g., relative advantage) is only one part of a provider’s set of beliefs about a treatment and there are other components likely contributing to adoption.

PE and CPT are both trauma-focused therapies, in which the patient engages with the trauma memory and related thoughts. There is a substantial literature showing that some providers are reluctant to use trauma-focused treatments and (rightly or wrongly) believe certain patients may not be ready for those treatments. For example, in a sample of over a thousand mental health providers enrolled in the VA national PE training program, many expressed concerns that PE might increase patient distress (Ruzek et al., 2014). Similarly, in a study of mental health care staff who were providing treatment to soldiers during a combat deployment in Afghanistan, exposure therapy techniques were the least frequently utilized psychotherapy practices (Penix, Adler, Kim, Wilk, & Hoge, 2016). In addition, trauma experts from Europe also reported that they do not often use exposure techniques (van Minnen, Hendriks, & Olff, 2010). These providers feared the use of exposure-based techniques would have a negative effect on patients’ symptom exacerbation and dropout, and would not be suitable for patients who had suffered multiple traumas in childhood. However, data from four well-controlled studies of PE and CPT in those with chronic PTSD indicate that reliable PTSD worsening was nonexistent, and the rate of reliable worsening of depression was low (Jayawickreme et al., 2014). Similarly, data from two randomized controlled trials found that a minority of women undergoing PE and CPT experienced PTSD exacerbations during treatment, and despite this, all had clinically significant improvement by end of treatment (Larsen, Wiltsey Stirman, Smith, & Resick, 2016). This suggests that providers should continue with these EBPs despite patient distress in some/most circumstances.

Specific perceptions are multi-faceted, and like general attitudes, likely contribute to the use of EBPs for PTSD. Understanding these general and specific attitudes can assist in addressing provider’s reluctance to provide these treatments, as well as enhancing attitudes associated with greater use, thereby enhancing dissemination efforts. In the current study, we had the opportunity to examine providers’ specific views of PE and CPT and compare these views to their general attitudes towards EBPs. Our data source was a national, longitudinal investigation of the implementation of PE and CPT in 38 VA residential treatment programs for PTSD. In the early stages of this five-year study, we developed a self-report measure to capture providers’ perceived characteristics of specific treatments (the Perceived Characteristics of Intervention Scale: PCIS; Cook et al., 2012; Cook, Thompson, & Schnurr, 2015) based on Rogers’ Diffusion of Innovations model. Using the second yearly wave of data collection, we found that the PCIS correlated with provider self-reported intentions to use PE and CPT (Cook, Thompson et al., 2015), as well as self-reported actual use of these treatments (Cook, Dinnen, et al., 2015). The current study builds on the prior work by comparing the EBPAS with the PCIS in order to examine how general versus specific attitudes relate to use of EBPs. This additional data collection also afforded the opportunity to cross-validate the prior findings on the PCIS with a newer wave of data.

There are differential rates of PE and CPT in VA residential programs (Cook, Dinnen, Thompson, Simiola, & Schnurr, 2014). Higher rates of CPT implementation in such settings may be related to the flexibility within the protocol to deliver the treatment in group or individual formats, and with the option of removing the trauma narrative without a reduction in efficacy. Indeed, the most commonly cited barrier to implementing PE across these programs was insufficient time and dedicated resources (Cook, Dinnen, Simiola et al., 2014). For example, numerous providers indicated they did not have the flexibility in their schedule to block 90-minute individual sessions at least weekly if not more. Given these factors, we chose to assess both individual and group formats for the use of CPT.

Based on the literature reviewed above, we hypothesized that more favorable general attitudes and specific perceptions would be related to higher self-reported use of PE and CPT, and that there would be a unique effect of specific perceptions even after accounting for general attitudes. In addition, we examined providers’ responses to the EBPAS and the PCIS and their relation to a number of other items that could also be related to use, including PE and CPT self-rated training and skill level, number of patients treated and sessions delivered, as well as perceived outcomes, ability, and sustainability of PE and CPT. We hypothesized that more favorable general attitudes and specific PE and CPT perceptions would be related to higher training and skill level of PE and CPT, more patients treated, more sessions delivered, higher belief in good outcomes, higher ability to effectively deliver aspects of the treatments, and perceived sustainability of PE and CPT.

Method

Participants and Procedure

The data from this investigation come from a larger, national, mixed-method, longitudinal investigation examining the implementation of PE and CPT in VA residential PTSD treatment programs (Cook, Dinnen, Thompson, Simiola, & Schnurr, 2014). All PTSD programs that report patient outcomes to the VA’s Northeast Evaluation Center were invited to participate. Of the 214 providers identified as working in the residential programs, 159 participants from 38 programs responded to the survey portion of the study and provided sufficient data for the purpose of this study. All participants were asked to log on to an encrypted Internet site and give their consent before participation. This study was deemed exempt from formal review by the Yale Human Research Protection Program due to low perceived risk, and was approved by the VA Connecticut Health Care System Institutional Review Board.

Measures

Measure of general attitudes toward evidence-based practice

The original EBPAS is a 15-item self-report measure used to assess mental health providers’ attitudes toward evidence-based practice (Aarons, 2004). It has four theoretically derived subscales: appeal (the extent to which a provider would adopt a new treatment because it was intuitively appealing), requirements (likelihood of adopting EBP given requirements by an agency, supervisor or states to do so), openness (the extent to which a provider was generally open to trying new practices), and divergence the extent to which a provider perceives research-based interventions are not clinically useful and less important than clinical experience. All items have a 5-point response scale, ranging from 0, not at all, to 4, to a very great extent. The overall Cronbach’s alpha reliability for the total EBPAS score is good (α = .77), and subscale alphas ranged from .90 to .59.

Measures of specific perceptions

The PCIS is an 18-item self-report measure of provider perceptions of specific treatments (Cook, Thompson, & Schnurr, 2015). It captures Rogers’ (1962) six perceived characteristics of innovation and four additional constructs added by Greenhalgh et al. (2004). The 10 perceived characteristics of innovation are: relative advantage, compatibility, complexity, trialability, observability, potential for reinvention, risk, task issues, nature of knowledge, and augmentation-technical support. The definitions and items that comprise each of these subscales as well as their descriptive statistics and reliabilities from a previous wave of data collection can be found in table 1 in Cook, Thompson et al. (2015).

In a prior wave of data collection (year two), it was clear that most perceptions of these two EBPs, with the exception of risk, were best characterized as a single dimension of positive views toward the specific treatments (Cook, Thompson, et al., 2015). We label this in subsequent tables below as positive perceptions or view. The PCIS risk scale, although distinct psychometrically from positive view and thus not included in the positive perception scale, also has adequate reliability (Cook et al., 2012). The scores for these scales were similar in this administration to those reported earlier and the alpha indicators of reliability were all over .85.

Utilization of EBP

Global self-report of EBP use was assessed using a series of three self-report items assessing: (1) use of PE administered on an individual basis; (2) use of CPT administered individually; and (3) use of CPT administered on a group basis. Each item assessed the proportion of patients to whom the treatment was utilized. For example, the item assessing PE read, “How often do you conduct PE on an individual basis?” and response options were on a 6-point scale, ranging from with less than 10% (1) to with over 90% of clients (6). “Not applicable” was also included as an option, but these providers were not included in these analyses, as they were not even rare users of the treatment.

Additional EBP delivery variables

For both PE and CPT, the survey included items on highest level of training, skill level, number of patients treated in past six months, number of session delivered, expected positive outcomes in modality, confidence in effectively delivering modality, perceived sustainability, and institutional support. More specifically, providers were asked to indicate the highest level of training they had achieved. The responses were: have not attended any training nor read the manual; have read the manual on their own or participated in informal training; had attended the multi-day VA training or similar interactive workshop; completed VA case consultation or received similar expert case supervision (within or outside of VA); achieved VA certification; became a VA consultant; or became a VA trainer. Skill level in both treatments was rated from 1 (no skill) to 7 (expert level skill). The number of patients treated with each treatment over the past three months was listed as well as the number of sessions of PE/CPT delivered. Providers were asked to rate their confidence in their ability to effectively deliver core elements of PE/CPT ranging from 1 (not confident) to 7 (completely confident). They rated the likelihood their program would continue to use PE/CPT in the future and to what extent (increase use, decrease use, or remain the same). Finally, providers were asked if they believed the program had taken steps to ensure the sustainability or continued use of these treatments.

Statistical Analysis

To account for possible shared variance among providers assessed at the same VA program, utilization of each of the three modalities of treatment examined were modeled using multilevel modeling, with providers nested within sites. Thus, the following procedure was repeated once for each of the three outcomes: PE utilization, CPT utilization with individual patients, and CPT utilization with group patients.

First, a block of control variables (number of beds, average length of stay for patients, and profession of provider) was entered, and used to estimate a multivariate model predicting the outcome. These variables have been found to be associated with utilization in some research (e.g., Everson et al., 2016), although not in previous research using this dataset (e.g., Cook et al., 2015). Next, the general attitudes (the four subscales of the EBPAS) were entered as a block. Finally, the specific perceptions (PCIS positive view and PCIS perceived risk) were entered, and a final model was estimated. To give an approximate estimate of the explanatory power of the blocks of variables, a pseudo-R² estimate of variance explained was calculated, using the −2 log likelihood method.

Results

The descriptive information about the analysis sample is presented in Table 1. As can be seen, the average number of beds was approximately 20 and the average length of stay was nearly 43 days. Almost half of the participants were psychologists and more than a third were social workers; combined, these respondents represented more than 80% of the sample. The means for the EBPAS scales were closely in line with prior research using these scales, with the exception of Divergence, which was lower than has been found in prior research (e.g., Aarons et al., 2007; Lim et al., 2012). The means for the measures of specific attitudes were very similar to those found in year two of this sample (Cook, Thompson, & Schnurr, 2015), suggesting that attrition and new recruitment had not changed the character of the sample.

Table 1.

Descriptive Information on the Provider Analysis Sample (N = 159)

	% (n) or M (SD)
Variable	Sample	Psychologists	Social Workers
Control
Number of beds	19.92 (11.94)	22.59 (12.80)*	18.17 (11.30)
Average length of stay	42.93 (22.33)	42.18 (27.08)	42.30 (20.18)
Type: Psychiatrist	7.5% (12)
Type: Psychologist	44.0% (70)
Type: Social worker	37.7% (60)
Type: Other	10.7% (17)
EBPAS: General attitudes
Requirement	3.66 (1.13)	3.67 (1.17)	3.73 (1.10)
Appeal	4.13 (0.74)	4.19 (0.56)	4.17 (0.80)
Openness	3.92 (0.76)	3.99 (0.70)	3.95 (0.78)
Divergence	1.81 (0.66)	1.73 (0.56)	1.87 (0.72)
PCIS: Specific attitudes
Positive view: PE	71.74 (13.03)	74.26 (14.41)	69.90 (11.74)
Risk: PE	5.74 (1.94)	5.58 (1.53)	5.76 (1.91)
Positive view: CPT	79.63 (14.38)	80.68 (14.51)	78.70 (15.06)
Risk: CPT	4.85 (1.97)	4.68 (1.80)	5.11 (2.15)

Note. EBPAS = Evidence-Based Practice Attitudes Scale; PCIS = Perceived Characteristics of Intervention; PE = Prolonged Exposure; CPT = Cognitive Processing Therapy.

^*p < .05

PE and CPT were the standard of care in the majority of the 38 residential programs. More specifically, 19 of the 38 (50%) residential programs offered both PE and CPT. Twelve sites (31%) primarily offered CPT without the option to engage in PE and three (8%) programs provided only PE without the option to engage in CPT. Four (11%) programs did not offer either PE or CPT and instead the core of programming was Acceptance and Commitment Therapy (n = 2), war-zone focused group psychotherapy (n = 1), or cognitive restructuring (n = 1).

The bivariate correlations among the EBPAS total score, PCIS (positive view and risk) for both PE and CPT and EBP use are presented in Table 2. There were high correlations among the various specific perceptions and general attitudes. More central to the hypotheses, the bivariate correlations between attitudes and outcome variables varied, depending on the outcome examined. Consistent with our hypothesis that more favorable general attitudes and specific perceptions would be related to higher self-reported use, more positive specific attitudes were correlated with more use of treatment for each of the three modalities (PE, CPT-Individual, and CPT-Group). On the other hand, positive general attitudes correlated with higher levels of use of CPT-Individual only.

Table 2.

Correlation Matrix of EBPAS, PCIS and EBP Use

	1	2	3	4	5	6	7	8
1. EBPAS Total	1	.17*	−.20*	.44*	−.37*	.07	.30*	.13
2. PE PCIS Positive View		1	−.52*	.30*	−.29*	.34*	.20*	−.13
3. PE PCIS Risk			1	−.16	.56*	−.19*	−.20*	−.06
4. CPT PCIS Positive View				1	−.38*	−.06	.49*	.38*
5. CPT PCIS Risk					1	−.07	−.26*	−.07
6. PE Utilization						1	.13	−.23*
7. CPT-Individual Utilization							1	.24*
8. CPT-Group Utilization								1

Note. EBPAS = Evidence-Based Practice Attitudes Scale; PE = Prolonged Exposure; CPT = Cognitive Processing Therapy;

^*p < .05

The bivariate correlations among the EBPAS total score, the PCIS (positive view), PE and CPT self-rated training and skill level, number of patients treated and sessions delivered, as well as perceived outcomes, ability, and sustainability of PE and CPT are shown in Table 3. There were significant correlations for almost all of these additional PE and CPT training, skill, and use-related variables with the PCIS positive view. However, the EBPAS total scored was only significantly correlated with one PE-related variable (i.e., expected positive outcomes in PE). The EBPAS total score was significantly correlated with six of the CPT-related variables.

Table 3.

Correlations among EBPAS and PCIS and EBP Training, Skill, and Confidence

EBP-Related Variable	PE		CPT
	PCIS	EBPAS	PCIS	EBPAS
	Positive View		Positive View
Training	.49*	.10	.47*	.37*
Self-rated skill	.55*	.01	.61*	.33*
Number of patients treated with modality/3 months	.15	−.01	.41*	.19*
Number of sessions per patient of modality	.24*	−.03	.38*	.11
Expected positive outcomes in modality	.61*	.17*	.76*	.48*
Confidence in effectively delivering modality	.52*	.03	.56*	.21*
Perceived sustainability	.37*	.04	.33*	.12
Institutional support	.64*	.12	.65*	.35*

Note. EBPAS = Evidence-Based Practice Attitudes Scale; PCIS = Perceived Characteristics of Intervention; PE = Prolonged Exposure; CPT = Cognitive Processing Therapy;

^*p <.05

These hypotheses were more directly tested in the multilevel modeling. The final models predicting utilization of each of the three EBPs are presented in Table 4. There was substantial support for an effect of specific attitudes; more positive specific perceptions were associated with more use of each of the three modalities of treatment. This was true even after accounting for general attitudes. On the other hand, there were no unique effects of general attitudes. To confirm that the lack of unique effects of the EBPAS was not due to an effect limited to one or more particular subscale of the EBPAS, these analyses were re-run using the EBPAS subscales (i.e., requirement, appeal, openness, and divergence). This approach resulted in no improved specification of the general attitudes portion of the model. These results are available from the first author.

Table 4.

Mixed Models Implementation Outcomes

	PE			CPT-Individual			CPT-Group
Fixed Effect	B	SE	T	B	SE	T	B	SE	t
Intercept	−0.59	1.30	−0.45	−1.27	1.56	−0.78	1.10	2.21	0.50
Control
Number of beds	−0.01	.01	−0.81	0.02	.02	1.07	0.00	.03	−0.17
Average length of stay	0.01	.01	0.14	0.00	.01	−0.44	0.00	.01	−0.33
Type: Psychiatrist	0.83	.52	1.60	−1.60	.78	−2.06*	−1.32	.99	−1.33
Type: Psychologist	0.47	.43	1.11	−0.94	.65	−1.45	−1.53	.83	−1.84
Type: Social worker	0.14	.42	0.32	−0.58	.65	−0.90	−0.93	.83	−1.12
EBPAS
Total Score	0.03	.05	0.73	0.06	.07	0.79	−0.01	.09	−0.16
PCIS
Positive View	0.03	.01	2.63*	0.06	.01	4.22*	0.05	.02	2.72*
Risk	−0.01	.06	−0.15	−0.07	.08	−0.85	0.03	.11	0.30

Note. Pseudo R² based on −2 restricted log. EBPAS = Evidence-Based Practice Attitudes Scale; PCIS = Perceived Characteristics of Intervention; PE = Prolonged Exposure; CPT = Cognitive Processing Therapy;

^*p <.05

Discussion

In a nationwide sample of VA PTSD residential treatment providers, specific perceptions of PE and CPT were better than general attitudes toward EBPs in predicting self-reported use of these treatments. Specific positive perceptions of PE, CPT-individual, and CPT-group were associated with greater use, whereas general positive attitudes were associated only with CPT-individual use. The lack of relationship between positive general attitudes did not appear to be due to the multivariate attenuation of effects because general attitudes were unrelated to use of PE and CPT-group even when no other variables were taken into account (i.e., bivariate correlations).

There are conceptual differences between the EBPAS and PCIS that might impact the observed findings. The first is the target of those attitudes (EBPs in general as compared to a specific EBP). The second is the types of attitudes assessed by Aarons’ list of EBPAS attributes as compared to Rogers’ list of attributes. The third is that the findings may reflect attitudes about treatments being trauma-focused, not just evidence-based, an important distinction for PTSD psychotherapies.

First, Ajzen and Fishbein (1977) demonstrated the lack of predictive value for individuals’ general attitudes on specific actions. In their review of 102 studies, of which 54 assessed general attitudes to predict specific behaviors, nearly half had non-significant results. A later meta-analysis of eight studies (Kraus, 1995) found a stronger correlation between specific attitudes than global attitudes and the target behavior. Thus, our findings may not be related to the predictability of Rogers’ list of attributes as opposed to Aarons’ list of attributes and instead may reflect established findings in the social science literature on attitudes and behavior change. Second, our study indicated that Rogers’ perceptions of innovations’ characteristics (i.e., relative advantage, compatibility, complexity, trialability, observability, potential for reinvention, risk, task issues, nature of knowledge, and augmentation-technical support) are uniquely associated with self-reported delivery of trauma-focused EBPs over and above the more general attitudes towards evidence-based practice (i.e., requirements, appeal, openness and divergence). This is consistent with findings from the Reding et al. (2014) study of community mental health providers. In that investigation, attitudes toward specific EBPs were related to use of the treatments above and beyond that of general attitudes toward EBPs. However their measure of specific perceptions came from two subscales of a longer version of the EBPAS, appeals and limitations. The EBPAS subscales of appeals and limitations are not overlapping with Rogers’ perceived characteristics of innovations and the items on the PCIS also do not overlap.

Third, although there are notable challenges in translating EBPs into routine clinical practice, there may be additional obstacles for trauma-focused EBPs for PTSD. Providers’ concerns about what trauma-focused treatments entail and their applicability to complicated patients may negatively impact use. For example, in a previous wave of data collection, we found that there were three broad patient factors that dissuaded providers’ use of these EBPs (Cook, Dinnen, Simiola, Thompson, & Schnurr, 2014): the presence of severe psychiatric comorbidities, cognitive limitations, and level of motivation or “readiness” to engage in trauma-focused treatment. The correlations in Table 2 show that the EBPAS scores are more in line with perceptions of CPT than perceptions of PE. Relatedly, as shown in Table 3, general attitudes towards EBPs are more in line with CPT-related training, skill, and competence than PE. Thus our findings may be due less to the superior performance of PCIS and more to a disconnect between attitudes toward EBPs and attitudes towards trauma-focused treatments involving exposure.

It is also possible that structural or organizational factors are more predictive of the use of PE and CPT-group, as opposed or in addition to provider attitudes. For example, programs might not have adequate staffing or time to provide 90-minute individual sessions to patients, or factors such as rolling admissions might make it difficult to organize group treatments. It is also possible that these links are largely driven by perceived fit, or what Rogers refers to as compatibility. On the other hand, perceived fit is so strongly related to other facets of perceptions of the interventions that it is difficult to disentangle them (Cook, Thompson, & Schnurr, 2015). Providers report that it is structurally more difficult and time intensive to deliver PE on an individual basis on VA PTSD residential treatment programs (Cook, Dinnen, Simiola et al., 2014).

The positive relationship between providers’ positive perceptions about a particular psychotherapy and its use has implications for organized care settings. The PCIS could be used to engage in systematic program development and evaluation. For example, psychologists who are clinic or peer leaders could ask practitioners in their settings to complete the PCIS shortly before training in a particular EBP to facilitate active engagement and planning of implementation. The information garnered could be used to aid in transforming existing programs. Conceivably, general negative perceptions as to an EBP’s characteristics, including relative advantage, compatibility, complexity, may indicate that additional supports are needed to implement this particular EBP, and what those additional precursor or supplemental supports needed may be. A clinic-wide assessment using the PCIS may indicate the need or suggest efforts to expand and improve services and for development of new or quality improvement programs. Additionally, those with high positive views of a particular EBP may be assigned to serve as a champion of the treatment in their setting, thereby helping to champion a supportive campaign to persuade resistant providers.

Since provider perceptions can be a precursor to the decision to try or regularly use a treatment, the PCIS also could be used for training providers. The PCIS could be part of a careful, planned pre-implementation assessment of staff that could assist and enhance the training and support process. For example, administering the PCIS to training participants could be helpful in assessing whether perceptions are interfering with buy-in. In particular, knowing that providers do not view a certain treatment favorably, do not see it as compatible with their existing practices or service structure, or view it as potentially provoking symptom exacerbation or dropout from treatment in patients, can allow trainers and consultants to address these issues up front. Indeed, future research might evaluate the sensitivity of the PCIS to training to within-provider shifts in their perceptions after training, after use with patients, and after supervision/consultation.

Our findings may have implications to the delivery of psychological services, including patient retention and dropout from therapy, provider burnout, and intent to leave a job. Research indicates that clinical mandates, difficulty carving time into staff schedules, time, and resources are related to provider use of EBPs for PTSD (Cook, Dinnen, Thompson, et al., 2015). It is likely that ongoing administrative pressures, performance measures, work demands, and provision of trauma-focused EBPs may also be a source of stress as providers learn new ways of conducting therapy as well as find time to implement them. Although VA PTSD outpatient providers reported significant satisfaction associated with the use of PE and CPT (Finley et al., 2011), these same providers also reported high levels of burnout (Garcia et al., 2014). It is certainly possible that if providers feel forced to deliver a treatment they perceive negatively, it could impact their job performance. The PCIS might be used to support providers in using EBPs they perceive most positively, and referring patients who want other EBPs to those providers for whom that is their forte. In addition, providers could engage in a private, detailed, reflective self-assessment prior to training and throughout the process of trying to implement an EBP. Perhaps this level of reflection and support might increase job satisfaction. However, this is speculation because we did not measure burnout or related indices. Understanding how providers’ attitudes towards specific EBPs relate to clinical and administrative matters would be an important next step.

The findings on the PCIS thus far should also be interpreted within the context in which it has been studied. The VA is a uniquely resourced organization guided by directives and mandates that help to standardize care across sites. Although the residential treatment programs in this study differed (e.g., number of patients, number of staff, length of stay), they all shared common assessment, treatment, and reporting requirements. Such uniformity is less likely to be found in community-based care. In VA, the mandate that all veterans receiving treatment for PTSD be offered PE or CPT when clinically indicated (U.S. Department of Veterans Affairs, 2008) might have influenced provider views towards PE and CPT. Findings might be different in an investigation with non-VA mental health providers from smaller, less well-resourced or systematized organizations, community centers, university clinics, or private practice whose use of EBPs was encouraged as opposed to mandated. Another consideration is that the PCIS was administered across multiple time points and the EBPAS was only administered at one time point. Perhaps general attitudes toward EBPs would have been more influential in the early stages of the VA PE and CPT training initiative. Also, this study relied on self-reported EBP use rather than observer-assessed delivery or fidelity. Additionally, findings are not necessarily generalizable to outpatient treatment or mental health diagnoses other than PTSD.

Due to concerns of staff burden, our study included the original 15-item version of the EBPAS. However, the EBPAS has multiple iterations (Aarons, Cafri, Lugo, & Sawitzky, 2012; Rye et al., 2017), with additional subscales. Although there appears to be no substantive overlap in these additional subscales with the PCIS, it is possible that using the revised version of the EBPAS would have warranted different results. Comparison of the PCIS and the shorter and pragmatic EBPAS-36 (Rye et al., 2017) seems a worthy avenue for future investigations.

Overall, the current study findings suggest that provider perceptions of PE and CPT are contributing factors in their delivery in VA PTSD residential treatment programs. This is important, given VA’s mandate to provide access to high-quality, evidence-based mental health care. Replication of findings in outpatient settings may help to guide strategies for increasing rates of provider uptake of EBPs. By pinpointing factors related to EBP implementation, providers may be able to increase the likelihood of successful adoption, scalability, and sustained use of EBP in organized health care settings.