Theoretical models that include attitudes have been proposed to explain and improve the dissemination process. For example, Rogers (1995) notes that studies of diffusion process span technologies such as use of the steel axe, agricultural innovation, teaching innovation, and medical/health innovation and disciplines including anthropology, sociology, economics, medicine, and marketing. Attitudes toward innovation can be a precursor to the decision of whether or not to try a new practice and the affective component of attitudes can impact decision processes regarding innovation (Candel & Pennings, 1999; Frambach & Schillewaert, 2002; Rogers, 1995). Still, little is known about behavioral health service provider attitudes toward adoption of EBPs or even how best to measure such attitudes. Indeed, service provider attitudes toward organizational change in community practice have been studied, but constrained samples have limited the generalizability of such studies (Addis, 2002). For example, Lehman, Greener, and Simpson (2002) examined staff attributes in regard to organizational change in substance abuse treatment settings. Most studies in mental health contexts have examined doctoral level licensed psychologists’ attitudes regarding use of treatment manuals and research-based information (Addis & Krasnow, 2000; Morrow-Bradley & Elliott, 1986; Prochaska & Norcross, 1983). There has been study of highly educated providers’ (i.e., doctoral level) concerns regarding manual-based treatments in traditional clinical settings (Addis, Wade, & Hatgis, 1999). However, in public mental health services, the majority of providers do not have doctoral level training (Aarons, Woodbridge, & Carmazzi, 2003) and attitudes of these providers have not been well studied.

Evidence suggests that the DI of innovation such as EBP must take into account the complexity inherent in real-world service settings (Fraser & Greenhalgh, 2001; Hasenfeld, 1992; Henggeler & Schoenwald, 2002; Jankowicz, 2000; Simpson, 2002). For example, in regard to regulatory concerns, service providers often work in programs that are subject to federal, state, and county policies and regulations. In regard to contracting, programs may have to compete for contracts and service provision is often subject to the terms of such contracts. Services also take place within organizational contexts that vary in regard to the quality of leadership and supervision, organizational norms and expectations, and climate (Glisson, 2002). Common methods of social service technology transfer (e.g., treatment manuals, off-site training sessions) often fail to take into account such complexity and thus may lack effectiveness (Addis, 2002; Backer, David, & Soucy, 1995; Backer, Liberman, & Kuehnel, 1986; Henggeler & Schoenwald, 2002; Strupp & Anderson, 1997). Thus, it is necessary to understand and consider attitudes toward adoption of EBPs of providers who are embedded within the complex organizational context of mental health service systems (e.g., Burns et al., 1999; Garland, Kruse, & Aarons, 2003; Glisson, 1992, 2002; Hoagwood et al., 2001).

The extant literature suggests at least four potentially important domains of provider attitudes toward adoption of EBPs. First, the intuitive appeal of innovation is important to consider in organizational change. This notion is supported by studies of persuasion processes and provider efficacy (Cialdini, Bator, & Guadagno, 1999; Tormala & Petty, 2002; Watkins, 2001). For example, studies have shown that providers are more at ease with information derived from colleagues in contrast to research articles or books (Cohen, Sargent, & Sechrest, 1986; Morrow-Bradley & Elliott, 1986) and attitudes toward adoption of EBPs will likely be influenced by the appeal of an EBP including the information source (Frambach & Schillewaert, 2002).

Second, requirements to provide services in a specified way based on organizational policies or funding exigencies may or may not be followed by service providers. For example, there is variability in the degree to which providers adopt and comply with new practices even when “required” by supervisors or agency mandates (Garland et al., 2003). Although some providers may be more or less compliant with required changes, individual and organizational variability can affect the degree to which innovations are adopted and sustained in practice (Glisson, 2002). Compliance with requirements differs from openness (i.e., willingness to try new experiences or consider new ways of doing things; McCrae & Costa, 2003) in that it denotes how employees respond to organizational rules and regulations. For example, an employee may be high on the characteristic of openness, but may also resist authority.

Third, openness to change in general has been identified as an important component of workplace climate that can impact innovation in mental health service programs (Anderson & West, 1998). Individual differences in openness are related to both organizational characteristics and job performance (Barrick & Mount, 1991). Business and organizational literatures have shown that openness to innovation may be important in developing the characteristics of “learning organizations” that are more responsive and adaptive to internal and environmental contingencies (Anderson & West, 1998; Birleson, 1999; Fiol & Lyles, 1985; Garvin, 1993).

Finally, a divergence may occur when there is a perceived difference between current and new practices. For example, mandated use of evidence-based assessment protocols are often perceived as incongruent or unneeded in clinical practice (Garland et al., 2003). Even where systems are in place to make the use of an EBP relatively seamless, there may be skepticism in regard to the use of such practices when perceived by providers to come from the culture of research and evaluation or when imposed by mandate. Similar “process resistance” has been documented in business sector studies (Garvin, 1993).

Thus, these four domains, intuitive Appeal, attitudes toward organizational Requirements, Openness to innovation, and perceived Divergence of research-based innovation, are likely to be important in understanding the process of DI of EBPs, but no measures are currently available to assess these constructs. Further, it is likely that these domains represent measurably distinct aspects of attitudes toward adoption of EBPs. For example, general openness to innovation is likely to be more akin to an attitudinal disposition rather than being contingent upon requirements in the workplace. The attitude of general openness is expected to differ from the attitude of appeal that is conditional upon the intuitive positive perception of an EBP. However, it is likely that perceived divergence of current practice with EBPs would be inversely associated with more favorable attitudes such as openness and appeal. It is expected that these domains can be identified and associations between domains examined.

Provider attitudes toward innovation and change are likely to interact with both individual differences (e.g., professional experience, training) and contextual factors such as organizational structure and organization type (Anderson & West, 1998; Birleson, 1999; Damanpour, 1991; Glisson, 2002). Studies support the contention that DI efforts of EBPs should take into account the education, training, and experience of service providers in order to facilitate the DI process (Ball et al., 2002; Strosahl, 1998). First, educational attainment has been found to be positively associated with endorsement of evidence-based treatment services and adoption of innovation (Loy, 1968; Ogborne, Wild, Braun, & Newton-Taylor, 1998). Second, a natural transition in the training of most clinical and case management professionals occurs during an internship or practicum experience. There is evidence that those still completing their education (e.g., interns) and transitioning into professional roles may be more flexible in regard to learning new interventions. For example, Ogborne et al. (1998) found that certified counselors were more likely than noncertified counselors to adhere to traditional conceptions of the causes and treatment of addictive disorders. Interns in specialty mental health clinics report more positive attitudes to using evidence-based assessment protocols (Garland et al., 2003). Interns providing services represent providers whose training is still in progress and may be less influenced by a long history of practice. As such, it is likely that interns would be more open to adoption of EBPs relative to providers who have been practicing for more protracted periods. Third, primary discipline in which a service worker is trained may also affect perceptions and use of empirical data or practices. In some cases specialized training may actually limit acquisition of new skills (Pithouse & Scourfield, 2002), however, specialized training that spans professional disciplines has the potential to positively affect breadth of practice (Amodeo, 2000). Still, there is variability by discipline in the emphasis on research and combining practice and research and this is becoming more important with the increasing demand to document evidence of effectiveness in practice (Thyer & Polk, 1997; Turnbull & Dietz-Uhler, 1995).

Contextual variation such as program type, organizational structure, and the presence of written policies regarding recommended practice may be important in understanding adherence—or lack thereof—to practice change (Glisson, 2002; Strupp & Anderson, 1997). First, it may be that the type of services to be delivered or program type (e.g., outpatient, residential) may be related to adoption of innovation and there is evidence that organizational innovativeness varies by type of organization (Damanpour, 1991). For example, there is variability in the mission, consumer population, and service staff of different types of mental health service programs. Second, in regard to organizational structure, organizations with high levels of bureaucracy and red tape may be less flexible in responding to change or promoting internal change relative to more flexible organizations (Frambach & Schillewaert, 2002). Baldassare and colleagues (2000) reported that local governments are often perceived as unresponsive, that there is a need for more responsive services, and recommend expanding the use of government contracting with private sector agencies and nonprofits in the provision of public services. Prager (1986) also found that social workers employed in more bureaucratic agencies were less flexible and made more restrictive long-term care decisions compared to social workers employed in agencies with flatter managerial structures. Finally, in working with mental health programs we have observed that some organizations have written policies specifying the use of specific interventions for specific disorders. Such practice policies can be assessed by program manager reports of whether or not a program has written policies specifying the use of particular interventions for a given mental health problem or disorder. Formalized policies may acquaint service providers with new technologies and demonstrate organizational support for matching treatments to disorders. Thus, as noted in the discussion above, it is important to understand the association of provider attitudes toward adoption of EBPs in relation to both individual difference and organizational/contextual factors.

Eighty percent of respondents were full-time employees and primary disciplines included marriage and family therapy (33.9%), social work (32.3%), psychology (22.4%), psychiatry (1.6%), and “other” (9.9%; e.g., criminology, drug rehabilitation, education, public health). Interns were less prevalent in the service system (24.9%) relative to fully employed staff (75.1%), and interns represented disciplines of marriage and family therapy (46.8%), social work (24.7%), psychology (20.8%), psychiatry (1.3%), and “other” (6.5%).

Participant programs were publicly funded child/adolescent mental health programs providing outpatient treatment (52.9%), day treatment (23.5%), case management (11.8%), wraparound services (7.8%), and inpatient treatment (3.9%). Most programs were contracted with the County to provide services (83.7%) in contrast to operating under County administration structure (16.3%). Fewer programs reported having written policies regarding interventions for specific disorders (14.3%) relative to those without such policies (85.7%). There was substantial variability among programs in regard to the number of unduplicated clients served per year (M = 257.6; SD = 452.8) and the number of clinical and/or case management service staff employed (M = 6.6; range = 1–31).

Factor analyses were conducted next. The EFA suggested a four-factor solution in accordance with examination of the scree plot, simple structure criteria, item-total correlations, and Chronbach’s alpha analysis of internal consistency reliability. Fifteen of the original eighteen items were retained and the EFA model accounted for 63% of the variance in the data. Table 1 shows overall means, standard deviations, item-total correlations, eigenvalues, internal consistency reliabilities, and item loadings for each of the scales. Cronbach’s alphas ranged from .90 to .59 with an overall scale alpha of .77. The factors represented four subscales of attitudes toward adoption of EBPs in keeping with hypothesized dimensions. Appeal (four items; α = .80) is the extent to which the provider would adopt a new practice if it is intuitively appealing, makes sense, could be used correctly, or is being used by colleagues who are happy with it. Requirements (three items; α = .90) is the extent to which the provider would adopt a new practice if it is required by an agency, supervisor, or state. Openness (four items; α = .78) is the extent to which the provider is generally open to trying new interventions and would be willing to try or use new types of therapy. Divergence (four items; α = .59) is the extent to which the provider perceives research-based interventions as not clinically useful and less important than clinical experience. Item analyses showed that the reliability coefficient for the Divergence scale would not have been improved by removing items from the scale. Although the internal consistency reliability for the Divergence scale was not optimal (i.e., <.60), such attitudes have been reported as an important construct in previous studies (Garland et al., 2003) and so the subscale was retained.

Next, a CFA was conducted using data from the other randomly selected half of the sample and specifying the factor structure identified in the EFA. As shown in Fig. 1, CFA items were constrained to load only on the primary factor indicated in the previous analysis, thus providing a highly stringent test of the factor structure. As in the EFA, factor intercorrelations were allowed. As shown in Fig. 1, CFA factor loadings confirmed the EFA-based a priori factor structure and the model demonstrated good fit (χ²(84) = 144.92, CFI = .93, TLI = .92, RMSEA = .067, SRMR = .077) further supporting the EBPAS factor structure. Factor intercorrelations ranged from r = .03 to r = .50. Figure 1 shows that Appeal had a strong positive correlation with Openness, a moderate positive correlation with Requirements, and a moderate negative correlation with Divergence. The Requirements scale was moderately negatively correlated with Divergence. Openness had no significant correlation with Divergence or Requirements.

The following sections examine each of the EBPAS subscale and total scores in relation to provider and organizational characteristics. Sample size for regression models varied slightly because of missing responses. Table 2 shows means and standard deviations for each of the EBPAS scales by each of the predictor variables used in the regression analyses below.

Five regression models were conducted examining the association of independent variables with each of the five EBPAS scales. When ICCs for a dependent variable are negligible across organizational units and cluster size is small, there is little reason to conduct multilevel analysis (Kreft & de Leeuw, 1998; Snijders & Bosker, 1999). For the present study, ICCs and design effects were small for all dependent variables including Appeal (ICC = .099), Openness (ICC = .071), Requirements (ICC = .016), Divergence (ICC = .043), and the EBPAS total score (ICC = .101). Despite the small ICCs and design effects, multilevel regression analyses were attempted. When Level 1 predictors were added, ICCs were further reduced for Appeal (ICC = .063), Openness (ICC = .065), Divergence (ICC = .027), and the EBPAS total score (ICC = .09). Because of the negligible cross-organization variability and lack of residual variance, models that included both Level 1 and Level 2 predictors did not converge except for the Appeal scale. This indicates that remaining cross-organization variability of EBPAS scores was accounted for by predictor variables in the model. Thus, because of negligible clustering effects leading to nonconvergence of multilevel models, standard multiple linear regression analyses were conducted to examine the association of individual difference and contextual variables with EBPAS scale scores. The following analyses are reported for each EBPAS subscale and for the EBPAS total score.

As hypothesized, scores on the Appeal scale were positively associated with higher educational attainment (β = .106, SE β = .042, p < .05) and intern status (β = .169, SE β = .098, p < .05), indicating that those with higher educational attainment and interns endorsed more positive attitudes toward adoption of EBPs given their intuitive appeal. In regard to organizational variables, providers working in case management organizations were more likely than those in outpatient programs to score lower on the Appeal scale (β = −.225, SE β = .114, p < .05) and providers working in programs with written practice policies were more likely to score higher on the Appeal scale (β = .219, SE β = .113, p < .05). For the Appeal scale, multilevel regression and standard regression models produced consistent results. Predictors accounted for 9.1% of the variance in Appeal scale scores.

Scores on the Openness scale were positively associated with intern status (β = .201, SE β = .105, p < .05) indicating higher Openness scores for interns. Openness was also positively associated with wraparound program type indicating higher Openness scores for providers working in wraparound programs (relative to outpatient programs; β = .298, SE β = .130, p < .054). Higher Openness scores were also found for providers working in low bureaucracy programs (β = .303, SE β = .138, p < .05), and programs with written practice policies (β = .370, SE β = .121, p < .01). Predictors accounted for 9.0% of the variance in Openness scale scores.

Providers from day treatment programs were more likely than those from outpatient programs to score higher on the Requirements scale indicating more positive attitudes toward adoption of EBPs if required to do so (β = .286, SE β = .150, p < .05). Providers working in less bureaucratic programs were more likely to score higher on the Requirements scale indicating a more positive attitude toward adopting EBPs if required to do so (β = .342, SE β = .172, p < .05). Predictors accounted for 5.9% of the variance in the Requirements scale scores.

Interns were more likely to score lower on the Divergence scale (β = −.216, SE β = .098, p < .05) indicating less perceived divergence between EBP and current practice. The model accounted for 6.4% of the variance in Divergence scale scores.

Openness and Appeal scales were moderately to highly correlated suggesting that openness to using new innovations may be facilitated by the intuitive appeal of a given EBP. The Requirements scale was moderately positively associated with both Appeal and Openness suggesting that providers with positive attitudes toward adoption of EBPs may also be more likely to comply with changes in practice that are part of work requirements. The moderate association between the Appeal and Requirements scales suggests that DI efforts may benefit from a careful balance of requiring changes in practice with making new innovations appealing to providers in order to facilitate the adoption of EBPs in real-world settings. The present findings converge with other studies showing that openness to innovation can be an important component of mental health program and organizational context that is important in the development of a learning organization (e.g., Anderson & West, 1998; Birleson, 1999; Garvin, 1993). The Divergence scale was moderately negatively associated with Requirements suggesting that those who perceive EBPs as being of little relevance are less likely to adopt a practice even when it is a job requirement.

The most consistent individual difference finding across EBPAS scales was that interns endorsed more positive attitudes toward adoption of EBPs relative to professional providers and this was consistent with hypotheses. Level of educational attainment was also associated with positive attitudes to adopting EBPs given their intuitive appeal. Specifically, interns were more likely to score higher on Appeal, Openness, total EBPAS, and lower on Divergence scales whereas higher educational attainment was associated with higher scores on the Appeal scale. Although education and internship overlap, level of educational attainment and internship represent related, but qualitatively different aspects of a mental health provider’s professional development trajectory. This synergistic scenario suggests that although professional education leads to a conditional openness to EBPs, professional internships may be an especially opportune stage of service providers’ professional development in which to “plant seeds” and reinforce the value of the use of EBPs. Related research has shown that during preprofessional status, workers may be especially predisposed to the acquisition of new practices because of more malleable knowledge structures (e.g., Day, Arthur, & Gettman, 2001; Rentsch & Klimoski, 2001). Such flexibility may facilitate the effectiveness of training in EBPs. It follows that a staged model of education and acquisition of professional skills and attitudes could serve as a template for recommending optimal times to promote a well-considered flexibility in practice.

No significant differences were found in attitudes toward adoption of EBPs across discipline. It may be that there are other factors that interact with provider discipline in complex ways. For example, providers with different personality characteristics may respond to organizational constraints in complex ways and such characteristics are important to consider in understanding and improving job performance, workforce development, and personnel selection (Barrick & Mount, 1991). Such complex constraints should be explored further with sample sizes large enough to comprehensively evaluate subgroup results.

In addition to provider characteristics, program characteristics were related to EBPAS scores. In contrast to outpatient providers, behavioral health care providers working in wraparound programs were more open and those working in case management programs were less open to adoption of EBPs suggesting that it is important to consider the programmatic context into which EBPs are to be disseminated. This finding requires further study to delineate how program type interacts with provider attitudes.

Level of bureaucracy was also associated with attitudes toward adoption of EBPs. Providers working in low bureaucracy programs were more predisposed to adoption of EBPs scoring higher on Openness and Requirements scales. This indicates a general openness to new practices and willingness to engage in new practices when required to do so. One implication for tailoring DI efforts to specific operational contexts is that some organizations may be poised to respond to environmental contingencies such as changes in contracting and practice demands, whereas others may be less flexible in regard to changes in policies or procedures that allow for practice change. The notion of the organization as an adaptive system extends the research on learning organizations and holds promise as an explanatory model of change in behavioral health services (Jankowicz, 2000). To survive, organizations must be able to adapt to market demands. However, in public mental health services, these demands are less likely to be market driven than in the private sector. It is plausible that a more market-driven approach could lead to more receptivity to the use of new and innovative technologies. For example, contracts could be structured in a way that requires use of EBPs in usual care placing funding and professional status as incentive structures. Further research should examine these issues and the potential for cross-fertilization between business models and public sector service models. This may lead to more responsive organizations in which providers and provider organizations can grow and thrive while providing high quality, effective services for consumers.

The impact of providers working in settings with written policies for treatment of youth mental health problems has not been well studied. In this study, the presence of written policies regarding treatment of mental disorders were part of internal program initiatives that appear to predispose providers to being more open to new practices. Although this confirmed the hypothesis posed above, further work is needed to examine what types of policies and under what conditions policies lead to more innovation and adaptability in organizations. However, it is likely that top-down models of imposing practice policies may engender higher Divergence, especially where high caseloads and administrative demands compete with the provision of services (Garland et al., 2003).

This study differs from previous inquiries of clinician perceptions of psychotherapy practices (e.g., Cohen et al., 1986; Morrow-Bradley & Elliott, 1986) in that it includes providers from a number of different disciplines with widely varying levels of education and practicing in real-world publicly funded mental health service settings. Thus, the ecological validity of this study is well supported. The present study provides preliminary support for the face and content validity of the EBPAS. Construct validity was explored in relation to both individual difference and organizational context variables. However, concurrent and predictive validity of the measure have yet to be assessed and further study is required.

Although structured approaches (e.g., manualization) may aid in the dissemination of EBPs, additional factors must be considered in order to most effectively change treatment practices. In keeping with other findings, this study suggests that both provider individual differences and contextual variation are important in understanding potential attitudes toward EBP (e.g., Glisson, 2002; Strupp & Anderson, 1997). However, there may be optimal times in a career trajectory in which to facilitate an ongoing openness to innovation. Further research should examine training models in various disciplines in order to identify components of training programs that would increase openness to adoption of EBPs while maintaining professional judgment and a balanced “scientist/practitioner” perspective on the judicious use of empirically supported interventions.

The EBPAS has practical utility in a number of ways. First, the scale taps provider attitudes likely to be related to elements of practice that may facilitate or hinder the adoption of EBPs in real-world settings. As suggested by findings for the Appeal scale, having a positively perceived local opinion leader to introduce and guide change in practice may facilitate receptivity and change in provider behavior (Denton, Smith, Faust, & Holmboe, 2001). Second, organizational factors may impact provider attitudes in ways that could facilitate or hinder DI efforts (Glisson, 2002). Finally, because the EBPAS is extremely brief (i.e., 15 items), taking about 1–2 min to complete, the measure can be used efficiently for research and in real-world practice settings to better understand the service context prior to DI of EBPs.

As an exemplar of a complex system, many factors will affect adoption of EBPs in community mental health settings. The experience, training, and primary discipline of the provider may affect what procedures are used to serve clients. For example, marriage and family therapists may be more open to adoption of interventions that are focused on family issues, psychologists may be more focused on individual treatment, and psychiatrists may be more focused on pharmacological treatments. The context of health care and behavioral health care is highly complex and inquiry into theory and change requires an equally complex perspective (Fraser & Greenhalgh, 2001). In addition to issues addressed in the present study, EBP dissemination efforts should consider the potential impact of financing structures, contractual constraints, political forces, and consumer preferences on how mental health programs provide services.