Using Mixed Methods to Adapt and Evaluate the Implementation of a Multicomponent Tobacco-Free Workplace Program within Behavioral Health Centers

Background: Behavioral health treatment centers (BHTCs) rarely implement existing evidence-based practices for treating tobacco dependence, despite high rates of tobacco use among their clients. Taking Texas Tobacco Free (TTTF) has successfully targeted this disparity by delivering an evidence-based, multilevel, tobacco-free workplace program providing policy implementation and enforcement, education, provider training in tobacco screenings and treatments, and nicotine replacement therapies (NRT) to BHTCs across Texas. We describe a mixed methods design used to conduct a formative evaluation process to adapt implementation strategies to local contexts, evaluate program outcomes and characterize processes influencing program implementation in two BHTCs serving 17 clinics. Methods: Varied data collection included pre and post-implementation leader, provider, and staff surveys; and pre, mid, and post-implementation provider, staff and consumer focus groups. During implementation, data were collected via various logs (tobacco screenings, NRT delivery) to monitor program content delivery. The RE-AIM ( R each, E ffectiveness, A doption, I mplementation, M aintenance) framework guided translation of behavioral interventions into sustainable practice. Results: While program implementation varied between each BHTC, all clinics adopted a 100% tobacco-free workplace policy, integrated tobacco screenings into routine practice, delivered evidence-based interventions, dispensed NRT to consumers and staff, and increased provider knowledge on how to address tobacco dependence. Pre, mid, and post-implementation qualitative findings served to: 1) develop program strategies and materials adapted to local contexts and populations and address barriers; 2) adjust delivery systems of key components to enhance implementation; 3) understand reasons for success or failure to implement specific practices; and 4) reveal program integration into clinic culture, enhancing sustainability. Conclusions: Implementation of TTTF at both BHTCs increased organizational capacity in the provision of evidence-based practices to treat tobacco dependence through successfully meeting most of our RE-AIM targets. Mixing methods involved program adopters and recipients as collaborators directly shaping core interventions to their individual context and needs, thus increasing

program fit, ownership, adoption and sustainability; closing the gap between research and practice.
These findings contribute to the development of flexible strategies and interventions capable of addressing variable implementation contexts and barriers, thus enhancing the effectiveness and sustainability of a tobacco-free workplace program.

Contributions To The Literature:
This is the first study using mixed methods to adapt and evaluate the implementation of a comprehensive tobacco-free workplace program within behavioral health settings.
We illustrate full application of the RE-AIM framework using mixed methods, yielding an understanding of contextual and multilevel (organizational, community, provider) factors influencing implementation, and explain results on each dimension.
Study findings present an effective model in building capacity for integrating evidence-based practices for treating tobacco dependence via a formative evaluation process that facilitated program uptake by developing flexible, tailored and sustainable strategies responsive to the real-world conditions of diverse settings.

Background
Behavioral health clients (BHCs) (those with mental and substance use disorders) represent only 21% of the population, yet consume about half of the cigarettes sold in the United States [1] with smoking prevalence rates nearly five times that of the general population [2][3][4][5][6][7]. Consequently, these individuals account for 50% of annual smoking-related deaths [8]. From 2005-2016, smoking rates among the general population have declined steadily from 20.9% to 15.5% [9]. However, smoking rates for BHCs remain relatively static, and significantly higher, at 35.8% [9]. Practice guidelines [2,10,11] recommend that smokers with and without behavioral challenges receive the same evidencebased-practices (EBPs) [12] proven effective in helping individuals in both groups to successfully quit smoking [13,14] at the same rates [15,16]. Nonetheless, BHCs are less likely to quit smoking compared to those in the general population [1,3] despite research showing equally high motivation to quit among both groups [17][18][19].
While many factors influence the persistently high smoking rates among BHCs, a significant contributor is the hesitancy of psychiatrists and behavioral health clinicians to offer specialized care for quitting [20,21]. In 2016, only 48.9% of mental health facilities screened patients for tobacco use, only 37.6% of programs provided tobacco cessation counseling, and only 26.2% offered nicotine replacement therapy (NRT) [22]. The main barriers to implementing tobacco cessation services within behavioral health treatment centers (BHTCs) include several clinician-level factors such as lack of effective training on treating tobacco dependence [23,24]; undervaluing of tobacco addiction as a problem [25]; lack of, or failure to, enforce tobacco-free workplace (TFW) policies [2]; and misconceptions regarding motivation for, and the behavioral effects of, quitting smoking for BHCs [12]. Clinicians also widely believe quitting smoking will jeopardize substance recovery efforts and trigger depression, aggressive behavior, suicide attempts or self-harming behaviors [20]. Although evidence indicates these clinician beliefs are myths [20], they persist and have the strongest negative impact [26,27], along with organizational barriers [28], on successful implementation of tobacco cessation interventions. Given the systemic nature of these barriers, organizational change is recommended [29,30].

Taking Texas Tobacco Free (TTTF) is a multicomponent TFW program assisting Local Mental Health
Authorities (LMHAs) throughout Texas to implement comprehensive and sustainable tobacco cessation [31], designed to target known implementation barriers across diverse levels of influenceorganizational, community, clinician, and client. LMHAs are state agencies that administer and provide behavioral health services across Texas according to geographic region through non-profit community mental health centers [32]. As a comprehensive TFW program, TTTF includes a TFW policy, clinician training in evidence-based treatments, integration of tobacco use assessments (TUAs) into routine practice, staff education on hazards of tobacco use, community outreach, and assistance with NRT provision [2]. TTTF was effective in building capacity for the integration of EBPs in assessing and treating tobacco dependence and helping clients and employees within 18 LMHAs successfully quit smoking [33].
To facilitate understanding and overcoming of barriers, implementation researchers are increasingly adopting mixed methods designs [34]. Combining qualitative and quantitative approaches provides a more comprehensive understanding of implementation issues and greater knowledge yield than either method alone [35].
Although widely recommended in implementation research [34,[36][37][38][39], few studies have applied mixed methods in evaluating smoking cessation interventions. Interventional studies have focused on evaluating the use of mono, rather than multicomponent smoking cessation interventions, such as the effectiveness of tailored smoking cessation training for the elderly [40] and for youth [41], and of financial incentives on quitting [42]. One exception is a mixed methods evaluation of policy implementation of a national public health guideline for smoking cessation in the UK, focused on preand post-implementation changes. [43] This study adopts a three-phase mixed methods design to examine the adaptation, dissemination and implementation of TTTF within two LMHAs overseeing 17 community clinics. The current study addresses limitations of prior research in implementation through using mixed methods to both adapt interventions and evaluate the processes and outcomes of implementing a comprehensive TFW program within BHTCs.

Study design
Following the successful implementation of TTTF within 18 LMHAs [31,33], we were funded to enable its statewide dissemination through active and passive means and to implement a more cost-effective TTTF within additional LMHAs, this time using a mixed methods approach not included in the prior implementation. Researchers recommend the use of qualitative and mixed methods approaches to address the complexities involved in implementing interventions across diverse contexts [44] and for developing strategies that facilitate the implementation of EBPs [34,36]. Mixed methods is particularly suited to evaluating complex interventions. The quantitative component assesses program impact, while the qualitative provides insight into contextual factors and change processes that are key to program adaptation and implementation [45,46], facilitating effective integration of EBPs into local contexts. Our purpose in using a mixed methods design was for, complementarity, (i.e. using quantitative data to examine outcomes and qualitative to examine processes), and expansion, (i.e. qualitative data is used to explain results of quantitative analyses) [47].
We used a multistage evaluation mixed methods design [48], consisting of three phases, each with a different aim and a different core mixed methods design (see Figure 1). Specifically: Phase 1) formative evaluation or needs assessment (pre-implementation) -to identify site-specific contextual factors that might affect implementation of TTTF to tailor the program to fit local contexts; Phase 2) program implementation -to assess the delivery and uptake of TTTF components within LMHAs; and Phase 3) summative evaluation (post-implementation) -to evaluate program outcomes and characterize processes influencing program reach, effectiveness, adoption, implementation, and maintenance [49]. RE-AIM was selected as a guiding framework as it was developed to translate behavioral interventions into sustainable practice [49]. Full application of the model requires in-depth understanding of contextual and multilevel factors influencing implementation, that use of both qualitative and quantitative methods provide, to understand and to explain reasons for outcomes on the RE-AIM dimensions [50]. Phases were interactive, with each phase building upon prior findings, and with iterative quantitative and qualitative data collection and analyses over the three phases [48]. The quantitative component had greater priority [48].

Participants and setting
This project has Institutional Review Board approval. Participation was open to any interested LMHAs in Texas who had not implemented the program previously. Our LMHA partner in the project, Integral Care, recruited centers through targeted mailings to ~20 LMHA chief executive officers (CEOs) assessing interest. Three LMHAs indicated interest, provided consent for participation, and enrolled: one withdrew mid-course due to competing organizational priorities. This paper focuses on the two remaining LMHAs. LMHA1 was in a large urban area and comprised two community BHTCs annually serving 10,247 unique clients (213,498 contacts). LMHA2 was in a mixed urban and rural area, spanning 25,000 square miles, managed 15 community BHTCs and annually served 6,538 unique clients (339,158 contacts).

Intervention: Taking Texas Tobacco Free
Previously described in detail [31,33,51,52], TTTF is a multicomponent, comprehensive, and sustainable TFW program that is implemented through an academic-community partnership. TTTF entails adoption of various EBPs in tobacco cessation [2,10], including: 1) policy implementation and enforcement; 2) integration of TUAs into routine practice; 3) provision of tobacco cessation services (e.g., behavioral counseling, NRT); 4) specialized provider training on treating tobacco dependence, i.e., Motivational Interviewing (MI); and 5) and community outreach ( Figure 2). Program components target challenges and misperceptions regarding tobacco dependence among BHCs, encourage program buy-in and integration into local settings to reduce tobacco use and exposure to environmental tobacco smoke (ETS), and ultimately, prevent tobacco-related cancers and other diseases. TTTF was designed to increase organizational capacity for the provision of EBPs for tobacco use, because the delivery of such interventions has proven effective in increasing quit attempts and cessation among clients [10]. Therefore, TTTF's success is measured through RE-AIM targets as opposed to smoking cessation, reduced morbidity, etc. (cf. [53,54]). We followed the Template for Intervention Description and Replication (TIDieR) Checklist [55] for describing interventions (Additional file 1).

Data collection
This study used a pre/post design with additional data collected during the active implementation phase to monitor program delivery and implementation. Data collection instruments used in this study are reported according to phase below, and described in Table 1.

PHASE 1: Formative evaluation
The initial step in adopting TTTF was preparing for program implementation through developing each LMHAs' 100% TFW policy that included electronic nicotine delivery systems (ENDS), training clinicians, and champions, in treating tobacco dependence, educating staff about tobacco hazards, and integrating EBPs in tobacco cessation into clinical practice [2,10], and tailoring materials to centers' needs. While TTTF team members consulted with LMHA leadership on policy drafting, we encouraged centers to confer with their members and community in developing their own TFW policy.
Each LMHA designated one program champion to oversee program implementation and maintenance according to a recommended timeline [33], and facilitate and steward successful organizational change [56,57]. TTTF team members provided guidance and practical advice throughout program implementation.
Quantitative data included pre-implementation clinic leader, clinician and employee surveys, LMHA demographics, pre-and post-knowledge tests, and the Organizational Readiness to Implement Change (ORIC) [58] survey that assessed organizational characteristics and needs regarding knowledge, skills, practice, and readiness to implement change [scale of 1 (agree) -5 (disagree), higher scores indicate greater commitment to change], (Table 1 describes instruments). Qualitative data included pre-implementation staff focus groups, field visits and champion consultations. Analysis of quantitative data informed sample selection for the pre-implementation focus groups. We purposively selected a heterogeneous sample for the focus groups, selecting participants who had expressed apprehension of TTTF to hear and address their concerns.

PHASE 2: Program implementation
The active implementation phase entailed adoption of various EBPs in tobacco cessation [2, 10] as described in Figure 2. TUAs are an empirically based method to increase quit attempts [2], which in this program consisted of documenting current, and history of, tobacco use; prior quit attempts and methods used; NRT use; and clients' readiness to quit. Counselors facilitated smoking cessation groups using a validated and effective smoking curriculum, developed specifically for BHCs [59,60].
TTTF staff delivered an 8-hour training in MI to participating LMHAs, as guidelines indicate the most successful evidence-based interventions combine the delivery of behavioral and pharmacological support [10,61]. In accordance with best practices, this combined therapy approach attends to clients' needs whether they are ready for smoking cessation treatments or motivational treatments to quit in the future [10,25]. Community outreach focused on expanding and sustaining community TFW programs to address high tobacco use rates among this population.
Quantitative data included LMHA quarterly reports submitted by champions, documenting TUAs administered and quantities and types of NRT distributed to staff and clients, and descriptions of monthly community outreach events. Various qualitative methods -mid-implementation staff, and separately, client, focus groups; site visits; and champion consultations -were used to engage program partners collaboratively, examine and understand implementation processes and challenges in greater depth, and make program adjustments as needed.
To facilitate documentation of client tobacco use, centers were asked to integrate TUAs into the electronic health record. Establishing use of these audit and monitoring methods are recognized as enhancing program success and sustainability [16]. Quarterly reports served two functions: (1) provided ongoing monitoring of program component delivery, allowing team members to intervene and suggest potential adjustments to these processes; and (2) provided data for summative evaluation.

PHASE 3: Summative evaluation
The summative evaluation phase consisted of LMHAs' ongoing implementation of evidence-based program components. Post-implementation, with the exception of the ORIC, we administered all surveys (center leader, clinician, and employee), continued monitoring implementation through quarterly TUA and NRT logs, and conducted staff focus groups, site visits and champion consultations.
All qualitative methods focused on examining and understanding how and why results were obtained for program outcomes.

Quantitative
Quantitative data were analyzed using SAS 9.4 [62] (Table 1). Descriptive statistics, including mean and standard deviation (SD) and percent, were provided for continuous and categorical variables of interest, respectively. The distribution of variables between pre-and post-implementation were examined using chi-square tests, as pre-and post-data were unmatched at the participant level.
Alpha was set at 0.05.

Qualitative
Focus groups and interviews were conducted using interview guides (available upon request from corresponding author), transcribed verbatim and uploaded to Atlas.ti8 (v8.4) along with all other qualitative data to facilitate data management. Qualitative data were coded inductively from themes drawn from the data as well as deductively, according to the RE-AIM dimensions. Three team members trained in qualitative research independently coded initial transcripts to develop a coding frame. Coding discrepancies were discussed and reconciled until agreement was reached, refining codes until a final coding frame was agreed upon and reapplied to all transcripts [63]. Coding and analysis proceeded iteratively across each stage of data collection. Constant comparison of data -a continuous process of comparing emerging data within and across previously coded transcripts -was used to refine themes, avoid redundancy, ensure fittingness of themes, and check accurate accounting of the data set [64]. Analyst triangulation, where two or more analysts independently code the same data and compare findings, was used to ensure congruency and credibility of findings [65].

Mixed methods integration
Quantitative and qualitative data were analyzed separately, and merged at the end of each phase.
Qualitative analysts remained blinded to quantitative results during qualitative data analysis, until data merging. Various types of integration were used to mix the quantitative and qualitative data, aligned with the different core mixed methods designs and aims of each study phase. In phase 1 (formative evaluation), the qualitative data was used to build and adapt intervention features to the local context. In phase 2 (program implementation and monitoring) qualitative and quantitative data were compared and connected to make ongoing program adjustments that facilitated successful implementation. In phase 3 (program evaluation) qualitative and quantitative data from phases 1 -3 were connected; that is, qualitative data on the implementation processes was used to explain the quantitative outcomes [66].
Regarding existing TFW policy, center leaders reported: 1 clinic (4.35%) had a TFW policy that included e-cigarettes; 4 (17.39%) regularly conducted TUAs and noted them within clinical records; 1 (4.35%) routinely provided tobacco cessation services for clients wanting to quit; and 14 (61%) had no existing TFW policy and did not provide assessments or any cessation services. However, regarding commitment to the TFW policy implementation at baseline, according to center leader surveys, 90% of employees (40% agree, 50% somewhat agree) were committed to implementing this policy change. Clinician survey results indicate only 25% of clinicians were providing TUAs preimplementation.
Qualitative findings were used to adapt the program to individual centers through development of: 1) program materials, i.e., posters representing center clients regarding age, ethnicity and language; and 2) additional resources tailored to the needs of special populations, including pregnant women, women with infants and young children. Materials included educational brochures providing guidance and specific recommendations on EBPs for treating tobacco dependence within these specialized subpopulations. These intervention adaptations enhanced the delivery, reach, and implementation of TTTF.
Qualitative analysis showed that while the majority of LMHA staff were onboard with becoming tobacco-free, some, particularly smokers, were apprehensive regarding program implementation. In fact, many felt that staff apprehension was primary and greater than that of clients: I anticipate a struggle with staff here who may be struggling with the new policy and then conveying that to consumers. 'You can do this [stop smoking] when I'm not sure I can do this'…I think the staff is going to be harder…I think our clients will follow what our staff does, for the most part, if we set a good example. (Program director, smoker, LMHA2) Other concerns and apprehensions regarding program implementation included perceived violation of staff and clients' rights to smoke, and staff's expectations that client resistance would manifest as violent behavior, as well as attrition, due to becoming tobacco-free: "Our staff's concern would be behaviors from our consumers that we're going to try to prevent smoking. They're going to act out and then they're going to hurt us." (Nursing director, LMHA1). However, generally, staff expressed confidence that once implementation had started, staff and clients would accept the program: Additionally, qualitative data revealed limitations to program delivery and implementation by LMHA.
At LMHA1, clients and clinicians reported a need to tailor educational materials, which TTTF staff complied with, to fit clients' advanced readiness to change, and challenges of not offering after-hours smoking cessation groups. Implementation at LMHA2 was hindered initially by not including employees in NRT distribution along with clients, which was soon corrected. Furthermore, LMHA2 did not offer any smoking cessation groups. Results from the qualitative and quantitative data collection during implementation were connected to build and adjust program strategies and materials, address implementation challenges and understand processes to further enhance program implementation and sustainability.

Phase 3
Detailed results evaluating program implementation according to how each RE-AIM dimension was measured within this study, along with qualitative and mixed methods findings on each dimension, is included in Table 2. Analysis of qualitative data produced five themes guided by the RE-AIM framework: factors related to Reach, Effectiveness, Adoption, Implementation and Maintenance.
( Table 3 provides quotes supporting each theme.)

Reach
Reach focused on changes in clinician and staff training regarding tobacco, and community outreach. Clinicians also related that attitudes in their greater communities were supportive of tobacco-free environments, which facilitated TFW program implementation at LMHAs.

Effectiveness
Effectiveness focused on the establishment, enforcement and acceptability of TFW policies, and decreases over baseline of numbers of LMHA employees using tobacco. Quantitative results indicated full implementation of TFW policies in both participating LMHAs, leading them to become 100% TFWs, protecting thousands of employees, clients and visitors from ETS. Although participating LMHAs established TFW policies, staff focus group participants reported variations in policy enforcement in the two LMHAs. The CEO of LMHA2 reported that in enforcing the TFW policy, they realized their initial fears regarding policy implementation were unfounded misconceptions, which she repeatedly stressed to staff as a strong supporter and implementation TFW program leader (Table 3). As a result, LMHA2 enforced the policy consistently with clients and reported no violations among employees.
Whereas during focus groups, LMHA1 staff repeatedly supported the misconception that enforcing TFW policies would precipitate client acting-out, and reported not enforcing the policy among clients due to fear of provoking violent behavior. Neither LMHA reported increases in client violent behavior, post-implementation. It is notable that these qualitative findings conflict with post-implementation survey data indicating adoption and enforcement of 100% TFW policies in both LMHAs. Additionally, LMHA2 employees were on board with enforcing their TFW policy and program, and reported policy integration into their workplace culture. LMHA1 staff reported that they should not be burdened with policy enforcement, signaling their reticence to accept and adopt the program fully. No employees reported quitting smoking at LMHA1; however, the percentage who smoked were few. At LMHA2, one employee reported quitting smoking because of TFW policy implementation during a focus group, and two others reported cutting down on smoking. Clinicians' and staff's views on tobacco training, as well as community attitudes toward tobacco use were supportive of TTTF and positively impacted reach 1) The more education you guys provided us, the better. Like h idea about that! I was really shocked at how much I didn't know this? So they're cutting their chances of getting into wellness w education. (CEO, LMHA2) 1) The training we received was fantastic, and getting it to staff on board. When you have your center saying 'We're behind you help you quit smoking…I think it really shows the staff as indivi something other than a warm body working.' (Program Director 2) My experience of (city) is that it's a very health-conscious co healthy living. A lot of people are into healthy eating, healthy liv Factors related to Effectiveness Categories: 1) Questioning or supporting myths/ fears about TFW programs 2) Staff attitudes towards TFW program 3) Staff experiences of quitting Attitudes towards Myths/Fears about TFW programs can aid or hinder program adoption Staff acceptance of TTTF facilitated program success TTTF helped employees to quit/ manage tobacco 1) But there were a lot of things that people predicted would ha smoking at respite or in the group homes. Most of the fears tha (CEO, LMHA2) 1) I'll be honest, I feel intimidated to go up to people to tell them someone who's already having a really bad day and then settin situation...I'm not going to risk myself to tell him to stop smokin LMHA1) 2) I can't see where there would be any barriers coming forward bought into it, because it IS policy. It's part of (LMHA2) now and environment for (LMHA2) now. (Program Director, LMHA2)) 3) Because we're going to a non-smoking campus I thought, we to smoke at work when I'm stressed out I might as well just qui my smoke breaks off campus and go somewhere where I can sm bad! (Counselor, LMHA2) Factors related to Adoption:

Categories: 1) Contextual factors affecting program uptake (facilitators and barriers) 2) TFW compliance 3) Clinician views of tobacco cessation interventions
Site-specific contextual factors can either hinder or aid program uptake Clinicians reported using novel practices compliant with TFW policy Clinicians reported benefits of using tobacco cessation interventions 1) The facilitator is going to be the staff and to a large degree t 'I'm going through this too.' I don't smoke on the grounds eithe (Program Director, smoker, LMHA2) 1) The support staff aren't trained in mental health and so they …And so, they don't say something when they're out there [sm we want to be smoke free, I don't feel like it's being addressed 2) Introducing mindfulness and those relaxed breathing skills… with them, and delaying the first cigarette too. A client told me, smoke, and then at the end of that hour, OK, just for another ho free].' (Counselor, LMHA1) 3) I think the best thing we've done so far is implementing the [ just 'Are you smoking?' It's 'Do you want to quit?' We're also gi they request it. (Program Manager, LMHA2) Factors related to Implementation Categories: 1) Tailoring material/strategies 2) Champion-initiated trainings 3) Program fidelity (consistent with Implementation Guide (IG); website) Program materials/ strategies tailored to staff suggestions for their center Champion's work load, and confidence-level limited implementing TTTF Staff shared faithfully implementing main TTTF components and use of TTTF's IG and website 1) My consumers, I go to their homes because they are 0 throug there is a pamphlet that we could give them the education abo danger. So we need to offer assistance to the parents who are p intervention specialist, LMHA2) 1) They [consumers] were really interested in how addiction eff activities in the book are more individualized…what was most im common ground of being able to support each other. (Counselo 2) I don't feel prepared to lead staff trainings…and don't really LMHA2) 3) We found it [website] really helpful…a lot of our stuff is base picked out the positive messaging, especially for our policy. ( 3) There was some withdrawal symptoms on the website that w them…there is a column for withdrawal symptoms and then an you're going to be depressed, and you might be irritable and he Factors related to Maintenance Categories: 1) Functional delivery systems 2) Attitudes towards sustaining program initiatives

3) Integration into organizational culture
Enabling uptake via adjustments to key initiatives delivery While aware of challenges of sustaining TFW program, staff valued the initiatives To varying degrees, changes were seen in organizational culture in participating LMHAs 1) The prescribers do them [TUAs]. Every time they [clients] see we've had is the doctors, and not having the time of really bein 2) The only thing I can foresee from a management point is just mindful not to forget and let it [TTTF] drop. Like already [5 mon we'll have to be continuously proactive on how we're going to, awareness fresh and keep these services available. (Anonymou 2) When you're interacting with a patient, instead of just saying want some assistance quitting?' And we weren't doing that. We knew whether they wanted help to quit or not…You got to take 3) It's a change in the culture and I think that

Adoption
Increases in adoption measures (TUAs conducted, EBPs provided, TFW policy compliance), were significant, with clinicians trained and conducting 13,659 (12,377 unduplicated) TUAs during implementation. Diverse NRT products (140.5 boxes) were delivered and distributed to clients and employees, which indicates people were making quit attempts. Qualitatively, staff attitudes towards enforcing the TFW policy served as both an implementation barrier and a facilitator. In LMHA1, staff apprehensions served as a barrier to full program implementation while at LMHA2, staff program support facilitated becoming tobacco-free. Clinicians at both LMHAs were enthusiastic about providing novel EBPs for tobacco cessation to clients, such as delaying gratification and reported their effectiveness in helping clients manage smoking. Likewise, clients reported that the smoking cessation groups had helped them to quit smoking: "And I've got these mints [NRT] which, you know, they really, really work…But without the group support, I wouldn't have been able to do it." (Sandy, Client, LMHA1) Within one smoking cessation group at LMHA1, 2 out of the 3 participants (66.66%) reported quitting smoking. Counselors of other groups reported that many clients had reduced their smoking significantly. Clinicians reported compliance with implementing practices consistent with upholding the TFW policy such as the TUAs, described as a vital step in helping clients in quitting smoking.

Implementation
Regarding implementation, quantitative assessment of program fidelity indicated each LMHA had successfully implemented all major TTTF components as intended. Qualitative implementation factors focused on staff and clients' suggested alterations to increase program fit to their center, and implementation consistency with the Implementation Guide (IG). Employees' suggestions led to the creation of various dissemination materials targeting their clients, including posters in Vietnamese, of young people and teens, and smokeless tobacco users, and information cards addressing the harms of ETS, especially on young children, and the hazards of smoking while pregnant. Employees reported consulting the website (www.takingtexastobaccofree.org) and the step-by-step IG and their usefulness in steering implementation efforts. Clients suggested adjustments to the smoking group curriculum, to reflect individuals in the preparation or action stage of change [67], rather than precontemplation: "It's [curriculum] geared towards smokers who haven't made the decision to stop smoking…but we have already made the decision to quit and that's why we signed up for the group" (Rick, Client, LMHA1)." Clients also asked for more group, rather than individual, smoking cessation materials and requested information on the neurobiology of nicotine addiction. We added these resources to the project website to address requests and improve implementation.

Maintenance
Regarding program maintenance, systems for monitoring and documenting TUA provision and NRT distribution were established, and modified where needed, to operate efficiently within each LMHA, and some tobacco education for new employees had been incorporated into annual trainings.
Qualitative data indicate employees' general attitudes towards ongoing program maintenance between the two LMHAs varied. At LMHA1, the tobacco-free program was seen as having been implemented and then forgotten about; a project requiring periodic reviving, rather than being integrated into center culture. At LMHA2, employees reported integration of the program and program maintenance into the center to become part of organizational culture. Both LMHAs valued NRT as essential to helping their clients manage or quit smoking.
Areas in which program implementation were weak included: 1) lack of significant decrease over baseline of the number of employees using tobacco, as the distribution of smokers did not change significantly from pre-to post-implementation (x 2 = 6.79, df = 3, p = 0.079; and 2) lack of sustainability initiatives, i.e. neither LMHA provided champion-initiated trainings, or continued smoking cessation groups. Challenges reported by champions to providing in-house trainings included competing organizational duties and lacking confidence to lead such trainings.

Discussion
Study findings demonstrate that while implementation of program components was differential within and between LMHAs, TTTF's implementation was largely successful. Both LMHAs integrated tobacco cessation interventions and TUAs into routine practice, delivered evidence-based tobacco cessation interventions, recorded significant increases in clinician knowledge on how to address tobacco dependence, and dispensed NRT to clients. Although both LMHAs reported adopting a 100% TFW policy, comparison of quantitative and qualitative data indicate inconsistent policy implementation and enforcement between the two agencies. While LMHA2 adopted and enforced a 100% TFW policy, qualitative findings from staff focus groups indicated LMHA1 staff inconsistently enforced the policy out of fear of provoking violent behavior from clients. Policy adoption without policy enforcement is limited.
As researchers have noted, clinician misconceptions regarding treating tobacco dependence within BHTCs serve as the strongest organizational barrier to successful implementation of tobacco cessation programs [26][27][28]. In applying what they learned in the TTTF trainings on treating tobacco dependence among BHCs, LMHA2 staff overcame their initial misconceptions and fears regarding addressing smoking among their clients to faithfully integrate TTTF into their organization. The overcoming of this fundamental barrier served as a catalyst allowing LMHA2 staff and leadership to incorporate effective clinician training on EBP's to treat tobacco dependence, value tobacco addiction as a serious problem, and establish and enforce tobacco-free policies -thus addressing the most commonly cited implementation barriers to tobacco cessation programs within BHTCs [2,12,[23][24][25]32]. Whereas at LMHA1, staff attitudes and misapprehensions regarding smoking and their BHCs persisted without change, resulting in partial program adoption due to lack of policy enforcement.
Although TTTF team members continually attempted to correct these misconceptions by providing research evidence and prior program experience to the contrary, these erroneous beliefs continued unabated among LMHA1 staff. Qualitative data indicate the difference between attitudes at each LMHA towards TFW policy enforcement was largely influenced by program support, or lack thereof, of center leaders. The CEO of LMHA2 actively championed and remained abreast of program implementation, while LMHA1 center leadership adopted a hands-off approach, delegating implementation exclusively to managerial staff. Our findings are consistent with various studies citing the critical importance of support and direction from organizational leaders to successful program implementation [68][69][70].
Qualitative data from clinicians and clients at LMHA1 indicate greater success in clients quitting smoking than at LMHA2. Differences in client quit rates at each LMHA may likely be attributed to the provision, or lack, of smoking cessation groups, which have proven effective in helping BHC's to quit as part of comprehensive tobacco-free programs [25,59,60]. Although LMHA1 only held one series of smoking cessation groups, clinicians and clients reported that in one group alone, two out of the three group members (66.66%) quit smoking with support from peers and NRT. LMHA2 did not provide any smoking cessation groups. At LMHA1, clinicians were enthusiastic, engaged and supportive in assisting clients in the smoking cessation groups, and clients attributed their success in quitting to the support they received from them. While staff at both LMHAs valued the provision of smoking cessation groups, their main obstacle was organizational, as neither operated in the evening or offered after-hours services, when clients were available for evening sessions.
Study findings on each of our individual program components is consistent with prior research supporting the effectiveness of TFW policies in reducing tobacco use [71,72]; TUAs in increasing quit attempts [2]; EBPs in treating tobacco dependence [10,73]; and increasing cessation services provided by training clinicians on treating tobacco dependence [24,74]. Moreover, as a comprehensive organization-wide TFW program, TTTF's multicomponent model proved effective in addressing the many challenges influencing successful program implementation [25,30] and affecting organizational change. Implementation scientists have stressed the importance of differentiating between core intervention components, and core implementation components, i.e., the core implementation drivers required to implement intervention components, such as coaching or training [75].
Employing a mixed methods design allowed key program stakeholders to participate collaboratively via qualitative methods to shape core intervention components to their needs, and researchers to understand context-specific implementation facilitators, barriers and processes at individual LMHAs.
Improving systems, organizational and community fit of intervention components enhanced core implementation components and addressed implementation challenges more effectively. Use of mixed methods also facilitated implementation by providing information on which intervention components were successfully adopted (or not), by whom, and why (or why not), expanding our understanding of what core implementation components need further development.
Identified areas of weakness in program implementation on the RE-AIM dimensions included: 1) absence of champion-initiated trainings (implementation); 2) inability to provide routine smoking cessation groups (adoption); and 3) continuing tobacco education -i.e., challenges integrating tobacco education into new employee orientation and training (maintenance). Although these components are on the clinician, systems, and organizational levels, the organization primarily determines them, as ultimately, they are due to prioritization of resources.
While champions' stewardship was vital in overseeing and organizing program implementation and maintenance efforts, neither took the next step to initiate any trainings in their LMHAs. Other studies support the importance of champions in leading successful implementation [76]. Champions reported competing duties and demands on their time and uncertainty in their ability to lead such trainings, as the main obstacles to providing in-house trainings. As both program champions reported similar challenges to providing and leading continuing tobacco education efforts, this indicates an area of needed improvement to the TTTF program, i.e., greater support and resources, such as train-thetrainer courses, to meet this implementation goal to ensure sustainability. Our future implementation programs will seek to address how to provide champions with the additional resources needed to transition from being managers to leaders of organizational change.
There are several limitations to this study. Some focus group participants may have had a stake in over-reporting program success. To diminish the potential for social desirability bias in the qualitative data, we intentionally included program sympathizers and detractors to get a more accurate picture of implementation, inquired about -and encouraged participants to share -negative and positive experiences, and used varied qualitative data collection methods, triangulating across data sources to ensure rigor [65,77]. Given the central role of context to successful implementation and the attention placed on this factor within this study, study findings are not necessarily applicable to other settings and populations. Our aim however, was not generalizability of findings; rather on the contrary, we sought to demonstrate and describe how we identified and responded to the needs of individual clinics to enhance program fit and implementation. Limitations for quantitative data included that not all surveyed stakeholders participated in data collection, despite solicitation, and inability to match pre-and post-data at participant-level. Also, as previously mentioned, client-level quantitative data on quit attempts among clients was not collected for comparison to qualitative reports.

Conclusions
Implementation of TTTF at both LMHAs increased organizational capacity in the provision of EBPs to treat tobacco use and dependence through successfully meeting the majority of our RE-AIM targets.
Adopting a mixed methods approach enhanced TTTF program implementation, which allowed us to conduct a formative evaluation process to adapt implementation strategies to local contexts; evaluate program outcomes; and characterize processes influencing program implementation in two LMHAs (17 clinics). Mixing methods also involved program adopters and recipients as collaborators who directly influenced implementation by shaping core interventions to their individual context and needs, facilitating uptake. The collaboration of such key stakeholders was vital to enhancing program buy-in, adapting delivery systems, program content and materials, and ensuring maintenance; and alerted us to needed improvements in core implementation components. Study findings suggest that successful implementation of multilevel, evidence-based tobacco interventions requires an in-depth understanding of the implementation culture at the level of the clinician, client, organization, and community level, to better address barriers and support facilitators.
Study findings contribute to the development of flexible strategies and tailored interventions responsive to real-world conditions in diverse settings, which are better equipped to address implementation barriers. The need to address tobacco dependence among BHCs is imperative. This evaluation of the TTTF program presents a successful model for the implementation of an effective and sustainable evidence-based, TFW program in community organizations interested in becoming tobacco-free.

Availability of data and materials
The datasets supporting the findings of this study are available, on reasonable request, from the corresponding author, Dr. Lorraine Reitzel at: Lrreitzel@uh.edu. Program materials are available online at our project website: www.takingtexastobaccofree.com.