This study analyzed the post-trial implementation of eSMART-TC conducted in the waitlist control group of a cluster randomized controlled trial (cRCT). In the cRCT, the intervention was delivered to the waitlist control group after completion of the primary intervention period, allowing for the observation and documentation of context-specific adaptations under routine implementation conditions.

The post-trial intervention involved 16 SMEs affiliated with the JHIA, excluding companies that withdrew before intervention delivery (n = 2). Characteristics of the participating SMEs are summarized in Table 1. Before study initiation, participating employers and HCMs received explanations of the study protocol and provided written informed consent. The study protocol was approved by the Ethics Committees of the National Cancer Center, Japan.

Figure 1 illustrates the overall study design and analytic workflow, including post-trial implementation, data collection, and analysis.

Provider discussion sessions were conducted as structured, framework-based consensus discussions rather than open-ended focus groups. The primary purpose of these sessions was to review documented modifications, classify implementation strategies using the Function and Forms Matrix, and evaluate whether each modification constituted a necessary adaptation for optimization, based on FRAME-IS elements.

Intervention providers were divided into three groups based on the timing of intervention initiation, with each group comprising four providers. Each group participated in three discussion sessions, held after sessions 3 or 4, after sessions 6 or 7, and after the final intervention session. Each discussion lasted approximately 1–2 h, and all providers attended all scheduled sessions for their assigned group.

The discussions were facilitated by members of the research team using a structured guide informed by FRAME-IS modules and the Function and Forms Matrix. Providers reviewed their completed intervention checklists, shared experiences regarding implementation challenges, and jointly assessed the appropriateness and implications of the observed modifications. Session outputs were documented through detailed facilitator notes taken during discussions. Sessions were not audio-recorded, and these documented outputs were used as analytic inputs for subsequent qualitative analysis.

In this study, co-creation was operationalized as a structured, iterative process through which intervention providers and the research team jointly reviewed, classified, and evaluated modifications to implementation strategies during and after intervention delivery, consistent with established co-creation methodologies in health and implementation research that emphasize collaborative knowledge generation through sustained researcher–practitioner partnerships (20).

The co-creation team comprised 12 intervention providers and 3 members of the research team (MO, JS, and TS), including the principal investigator and co-investigators with expertise in implementation science and workplace health promotion.

Co-creation was implemented through three sequential activities. First, intervention providers documented all deviations from planned implementation strategies using standardized checklists based on the FRAME-IS framework after each session. Second, these documented modifications were reviewed during structured provider discussion sessions, where implementation challenges were examined, and modifications were assessed in relation to optimization. Third, classifications of core functions and forms, identification of key forms, and determinations of necessary adaptations were finalized through consensus-based discussions informed by the Function and Forms Matrix and FRAME-IS elements.

Consensus was defined as agreement among all participating providers in each discussion session. Disagreements were resolved through facilitated discussion until all participants reached an agreement. All consensus decisions were documented in session notes and validated in subsequent sessions to ensure consistency across provider groups.

Intervention providers contributed information derived from implementation experience in workplace settings, and the research team facilitated the application of implementation frameworks during the review and classification process. Outputs from these co-creation activities informed the subsequent qualitative analysis.

The study assessed three primary outcomes. First, it identified the core functions and corresponding forms of eSMART-TC implementation strategies. Second, it determined the adaptations necessary to optimize these strategies. Finally, the program's effectiveness and acceptability were evaluated to contextualize adaptations. Effectiveness was measured by the smoking abstinence rate, while acceptability was assessed through participants' perceived effectiveness and satisfaction with the program.

Data were obtained from (1) checklists completed by intervention providers, (2) employee surveys conducted at baseline and post-intervention, and (3) company surveys administered after the intervention in the waitlist control group of the eSMART-TC cRCT. Details about adaptations were collected using intervention checklists completed by providers for each session (Additional file 1).

The checklist listed the strategies to be delivered in sessions 1–10, consistent with the intervention group protocol. It included a section to record whether each strategy was executed. When a strategy was not implemented, this was considered a modification. Providers documented such changes in a designated section based on FRAME-IS modules 1–6: (1) description of the treatment and implementation strategies, (2) content modified, (3) nature of the modification, (4) rationale, (5) timing and whether planned, and (6) individuals involved in the decision. After each session, providers submitted the completed checklists to the research team.

Smoking status was measured through pre- and post-intervention employee surveys. Self-reported 7-day point-prevalence abstinence was assessed by calculating the percentage of current smokers at baseline who answered “yes” to the question in the post-intervention survey: “Have you smoked even one mouthful of tobacco [including all cigarettes (rolled, heated, and other types)] in the last seven days?” Program acceptability was evaluated based on HCMs' responses to the company survey regarding effectiveness (scale 1: not effective at all to 5: very effective) and satisfaction (scale 1: not satisfied at all to 5: very satisfied).

To address the study objectives, analysis was organized in two parts: (1) defining core functions and forms of the implementation strategies, and (2) identifying modifications that support optimization and potential sustainment of the program.

Qualitative analysis was conducted to define core functions and forms of the eSMART-TC implementation strategies and to identify modifications considered necessary adaptations for optimization. Data sources included: (1) implementation strategy checklists completed by intervention providers for each company (n = 16 checklists, one per company, documenting all sessions from initial assessment through final review), and (2) documented outputs from nine structured provider discussion sessions (three sessions per provider group).

Analysis followed a directed, framework-based approach guided by FRAME-IS and the Function and Forms Matrix. Two members of the research team (MO and JS) independently reviewed submitted checklists to identify instances in which planned implementation strategies were not delivered as specified. Each discrepancy was documented and provisionally classified according to FRAME-IS modules (action, actor, action target, temporality, dose, implementation outcome, and justification).

To enhance trustworthiness and consistency of the interpretations, multiple validation strategies were employed. First, provisional classifications were independently reviewed by two analysts (MO and JS), with discrepancies resolved through discussion and, when necessary, consultation with a third team member (TS). Second, all classifications were validated through member checking during structured provider discussions, ensuring alignment between framework-based interpretation and actual implementation experience. Third, consensus decisions were documented and reviewed across multiple sessions to ensure consistency over time and across provider groups. This multi-layered validation process ensured that the results of the analysis reflected both rigorous framework application and practical implementation realities.

This study follows the StaRI and SRQR reporting guidelines; completed checklists are provided as Supplementary Files.

Among the 26 implementation strategies (forms), a subset was identified as key forms based on patterns of modification observed during implementation (Table 4). Eight forms (No. 1, 2, 4, 5, 7, and 23–25) were consistently delivered without modification across all companies. Furthermore, 10 forms were modified in some companies but were ultimately retained as key forms because the observed modifications were either undesirable (e.g., omissions due to oversight or modifications associated with negative outcomes) or limited to highly specific contexts not generalizable to typical SME settings (e.g., companies with very low smoking prevalence or medical institutions with specialized resources). Among these modified forms, No. 3, 8, 9, 10, and 13–18 were designated as key forms following consensus discussions with intervention providers.

Through discussion with intervention providers, one additional implementation strategy—No. 26: “The HCM shares the results of the review and evaluation with the entire company”—was designated a key form due to its role in promoting internal communication and organizational learning. As a result, 19 implementation strategies were officially designated as key forms (Table 4).

The key forms of eSMART-TC's implementation strategies were organized into four core functions addressing four motivating needs and 26 forms to operationalize each function effectively. The first core function, Ensuring Executive Engagement and Support, was developed to address the challenge that limited executive commitment impedes the creation of a health-conscious organizational culture and reduces the likelihood of achieving meaningful outcomes. Six forms were assigned to this function, including setting objectives for tobacco control and issuing a formal implementation declaration to employees. The second core function, Ongoing Facilitation Tailored to the Workplace Context, aimed to mitigate the tendency for health promotion to be deprioritized amid competing business demands and to ensure facilitation throughout all phases of implementation. Forms supporting this function included conducting regular intervention sessions, requiring the HCM to submit an action memo reporting on progress between sessions, and providing facilitation based on these updates. The third core function, Supporting Employees to Quit Smoking through the Cessation Campaign, was designed to increase motivation among employees to stop smoking. HCMs were responsible for disseminating evidence-based cessation methods and delivering support tailored to employees' readiness to quit. The final core function, Review and Evaluate the Campaign, addressed the difficulty of sustaining the Plan-Do-Study-Act (PDSA) cycle over time. To maintain ongoing improvements, forms required that tobacco control efforts be reviewed during established company health meetings, such as the hygiene committee, and that new goals be set based on these evaluations. Additionally, intervention providers, particularly PHNs serving as external change agents, emphasized the importance of conducting evaluations and sharing results internally, which was adopted as an additional form through consensus discussions. Table 4 summarizes the motivating needs, core functions, and forms using the Function and Forms Matrix.

Adaptations essential for optimization were specified within seven forms across three core functions. Based on a review of documented modifications using FRAME-IS elements and structured discussions with intervention providers, seven such adaptations were identified by consensus and categorized according to the core function and form they supported (Table 5).

To support interpretation of the adaptations, descriptive statistics summarized smoking abstinence rates and program acceptability. Across 16 companies, the average smoking cessation success rate was 9.5% (range: 0%–30.8%). In the company survey, the average effectiveness rating was 4.4 points (SD: 0.51, range: 4–5), and the average satisfaction rating was also 4.4 points (SD: 0.51, range: 4–5). These results served as supplementary outcome data that supported both the appropriateness of decisions regarding whether implementation strategy adaptations were necessary in practice and the feasibility and acceptability of the adapted strategies.

Clearly defining core functions and key forms supported fidelity while allowing context-specific flexibility. Using the Function and Forms Matrix, the eSMART-TC implementation strategy was categorized into four core functions designed to address four motivating needs or problems, along with 26 forms enforcing these functions. The adaptation frameworks FRAME–IS and MADI emphasize the importance of preserving the functionality of core functions during adaptation (15, 21). Therefore, establishing clear definitions of core functions and forms is considered a critical first step in adaptation (21). Throughout this process, fidelity was maintained in 19 of the 26 forms. Intervention providers reported that these forms contributed positively to outcomes when implemented. Accordingly, these 19 forms were identified as key forms of the implementation strategy, and it is recommended that they be implemented without modification as a general principle. However, if obstacles to implementation arise or a more effective approach is identified, planned adaptation is advised.

Previous research on workplace tobacco control has largely focused on intervention components delivered to employees rather than on implementation strategies for building organizational capacity (12). The four core functions identified in this study demonstrate both convergence with and extension of these approaches. Systematic reviews of workplace smoking cessation interventions have identified effective components, including counseling, pharmacological treatment, and smoke-free policies (22). Our “Employee cessation support” function aligns with these evidence-based elements. However, prior studies have given limited attention to the organizational implementation infrastructure required in resource-constrained settings such as SMEs (5, 6). Our three additional core functions—“Executive engagement,” “Facilitation for HCM activities,” and “Campaign design and implementation”—address this gap by building organizational capacity to deliver and sustain cessation support. Moreover, these functions correspond to strategies described in systematic reviews of workplace implementation strategies (23) and the Expert Recommendations for Implementing Change (ERIC) compilation (12), such as “develop stakeholder interrelationships,” “identify and prepare champions,” “facilitation,” and “use evaluative and iterative strategies.” These consistencies with well-established implementation strategies enhance the validity of our approach. However, our “Facilitation for HCM activities” reflects SME-specific elements not fully captured in existing frameworks: sustained, individualized support for designated healthcare managers through external public health nurses, with flexible formats and frequencies adapted to company readiness and resource constraints (5, 6). Furthermore, the systematic identification of 19 key forms advances adaptation science by operationally distinguishing essential vs. adaptable components—an issue emphasized in FRAME-IS and MADI (15, 20) but rarely specified in practice.

Context-specific adaptations, particularly to the duration and frequency of activities, effectively improved implementation feasibility. Modifications were made to seven forms to align with each company's context and were identified as necessary adaptations for optimizing the implementation strategy. Of these adaptations, five addressed the frequency and duration of interventions. These adjustments reflected intervention providers' proposals regarding the number of sessions and time allocation based on company readiness, industry characteristics, and unique challenges. Addressing barriers such as time constraints imposed by the program and implementation challenges within JHIA also motivated these adaptations. Modifications included reducing the number of sessions during the intervention period, adjusting the time allocated to planning and campaign execution, and extending the smoking cessation campaign beyond the intervention timeframe. These changes were determined through ongoing assessments of company needs and discussions between intervention providers and company representatives. Apart from modifications to session format (in-person or online), most modifications were reactive. Although proactive modifications are generally preferred, prior research (15) indicates that flexible, adaptive responses are often necessary to address unanticipated changes in real-world settings. Reactive modifications were more common in this study for several reasons. First, because the implementation strategy spanned preparation, implementation, and maintenance phases, it was often difficult to assess the need for adjustments until each phase was underway. Second, modifications incorporated findings from needs assessments conducted by HCMs among employees during implementation. This study distinguished between “modification” (any change to implementation strategies) and “adaptation” (intentional, contextually appropriate changes to improve fit), regardless of whether changes were planned or reactive (19, 24). This distinction follows frameworks such as FRAME-IS and MADI.

Therefore, it is necessary to distinguish between planned and reactive modifications when documenting them. In this study, reactive modifications were implemented flexibly through consensus-building between intervention providers and companies to ensure the most appropriate approach in each case. Because these modifications were systematically recorded, the knowledge gained about specific contexts and necessary adjustments can be shared among intervention providers. Over time, accumulating such records can help build evidence on effective modifications for particular settings.

However, compared to planned modifications, reactive modifications require caution, as their impact can be difficult to predict, they may deviate from core functions, and they risk inconsistency (15). Modifications should never be made arbitrarily; they must maintain fidelity to core functions, as deviations can undermine the intended effectiveness of the strategy (18). Implementing proactive adaptations rather than relying solely on reactive modifications can help increase implementation success while preserving core function integrity (25). Therefore, it is essential to train intervention providers to verify fidelity to core functions when making reactive modifications. Training should also include conducting systematic needs assessments, improving the fit of the strategy to the company's context, and planning modifications to address existing constraints effectively.

Collaborative refinement during implementation supports feasibility, motivation, and sustainability. In this study, intervention providers conducted three discussions during and after the intervention to finalize the core function matrix and determine necessary adaptations for optimization. During these discussions, intervention providers expressed positive expectations for the program and noted that adaptations helped sustain the motivation of participating companies. In future studies, combining FRAME-IS with a tracking framework such as the Longitudinal Implementation Strategy Tracking System (LISTS) may further enhance the ability to capture how implementation strategies are applied and evolve over time (26).

Although adaptations aim to improve processes, if forced or introduced hastily, they may compromise service delivery and effectiveness by reducing feasibility and fidelity (27). Modifications to implementation strategies align with a co-creation model, in which researchers and stakeholders—including implementers, policymakers, and community members—collaborate to identify needs, generate knowledge, and adapt strategies to local contexts. This approach emphasizes equal partnerships, shared decision-making, and mutual learning, and has been applied across health, welfare, and community-based research settings (28–30). Within this framework, contextual adaptation is not only expected but essential for sustainable and acceptable implementation (20). Moreover, adapting strategies through co-creation with frontline healthcare providers, patients, and policymakers has been reported to improve acceptability and enhance sustainment in practice (31). Therefore, rather than delivering standardized programs or implementation strategies, adopting a flexible approach consistent with the co-creation model will be crucial for achieving effective and lasting implementation.

This study offers a new perspective and framework for advancing fidelity-aligned yet adaptive implementation. It has several strengths, including the real-time documentation of modifications, theoretical systematization of core functions and forms, addressing implementation barriers, and application of the co-creation model. By integrating these elements, this study goes beyond merely recording modifications and provides a novel perspective on optimizing and sustaining implementation strategies, contributing significantly to the academic field. Furthermore, the proposal of key forms represents a major strength. While previous research has extensively discussed the adaptation of implementation strategies, systematic efforts to identify which forms are essential for successful implementation have been limited (15, 21). In this study, by applying the core function and forms framework, 19 of 26 forms were identified as key forms recommended for implementation without modification in principle. This approach offers a new guideline for balancing fidelity and adaptation and has substantial practical value.

This study has several limitations. First is the limited number of participating companies and the difficulty in evaluating the impact of adaptations. Because adaptations were made continuously and tailored to each company's context, variations arose in the implementation process and approach, making it impossible to directly compare the effects of interventions with and without adaptations under equivalent conditions.

Ideally, a study design comparing groups with and without adaptations across multiple companies sharing similar contexts—such as industry type, size, and employer leadership—would have enabled a more rigorous evaluation of adaptation effects. Future research should document adaptation processes in greater detail and employ impact-oriented frameworks such as MADI (21) to assess adaptation effects more clearly.

The second limitation is the lack of quantitative measurement of intervention providers' evaluation and acceptability of the program. This study did not conduct quantitative assessments of how intervention providers perceived and accepted the program. Consequently, it was not possible to quantify the extent to which the adapted program was accepted or to identify factors influencing implementation and outcomes based on systematic data. Although qualitative data, including interviews and feedback, were collected, combining them with quantitative measures would allow for a more robust assessment of providers' perspectives. Future research should incorporate quantitative evaluation methods to clarify how factors such as program acceptability, feasibility, and sustainment at the provider level affect implementation success. Moreover, the designation of key forms is inherently context dependent. The determination that specific forms are essential for maintaining fidelity was based on implementation experiences within Japanese SME settings and should not be generalized to other organizational contexts, cultural settings, or healthcare systems. As implementation data accumulate across more diverse settings, the classification of forms as “key” vs. “adaptable” may need to be revisited. Future research should examine whether these key forms remain critical across varied contexts or whether different forms emerge as essential in other implementation environments.