The input phase includes WP1 (preparation for data collection) and WP2 (cross-country data collection) and spans months 1-17 (see Figures 2 and 4). This phase will map and identify factors influencing resilience and mental well-being among health care workers, informal caregivers, and leaders in older adult care across 6 European countries (Finland, Italy, the Netherlands, Norway, Romania, and Spain) and Australia. It aims to provide a holistic understanding of the key factors that support leaders in older adult care. Furthermore, it will explore the needs and perspectives of health care workers and informal caregivers to enable successful adaptation to changing work conditions, ethical dilemmas, and adverse events and to identify potential solutions. The main activities in this phase include the following: (1) literature reviews; (2) contextual mapping of health care systems; (3) mapping of resilience and mental well-being factors among health care workers, informal caregivers, and leaders; and (4) exploration of health care workers’ and informal caregivers’ perspectives.

We will supplement the existing research literature [15-17] by conducting 3 reviews on the following topics: (1) factors supporting individual and organizational resilience and the relationship between individual and organizational resilience [18]; (2) factors and interventions supporting the mental well-being of older adult care workers [19]; and (3) the effectiveness and cost-effectiveness of organizational management interventions for promoting mental well-being in older adult care [20].

First, the findings will enable the development of a cross-disciplinary table of definitions, measures, psychometric properties, and identified research gaps, which will be combined into novel, holistic theoretical concepts and used to develop a Resilience Capacity Scale (RCS). Second, the relationships between resilience, mental well-being, and economically relevant outcomes, such as sick leave, disability, turnover, burnout, and productivity, will be mapped. Results from the literature review on intervention and implementation design in older adult care will inform the S4R intervention design, including final decisions on the selection of instruments for quantitative and cost-effectiveness analyses. Third, the findings will inform the content of questionnaires and interview guides for collecting data from leaders, health care workers, and informal caregivers.

The organization of older adult care differs across European countries and Australia. The project will map a broad range of external and internal contextual factors that may influence the research and implementation of interventions to improve resilience, mental well-being, quality, and patient safety in older adult care services. Information on external contextual factors will be collected through publicly available sources, including national, regional, and local government websites, as well as from relevant authorities. Internal contextual factors will be collected from the recruited sites. The context mapping tool is based on an existing tool designed for mapping older adult care contexts across different countries [21,22] and is anchored in the Consolidated Framework for Implementation Research developed by Damschroder et al [23].

An online survey will be conducted in the 6 European partner countries and Australia. WP2 aims to map resilience and mental well-being factors and explore their associations across 7 health care systems. A total sample of 2500-3000 participants (minimum of 300 leaders, 1000 informal caregivers, and 1200 health care workers) will ensure:

WP2 is descriptive and exploratory; multivariable regression models will be used to explore associations among resilience, mental well-being, job characteristics, and demographic factors and to inform instrument development and intervention design. These analyses are not intended for confirmatory hypothesis testing. Consequently, no formal power calculation was conducted for WP2. Instead, the sample size was determined based on feasibility, stakeholder representation, and the requirements for psychometric evaluation. Moreover, the limited availability of leaders in older adult care must be taken into account.

The participating countries represent highly diverse health care systems (see Figure 3). Informants will represent different genders and age groups, and health care workers and health care leaders will have diverse professional backgrounds and experiences. The survey will be available in the official language of each country. Informants at the recruited sites will be recruited through leaders and the use of QR codes, whereas informal caregivers will be recruited through the participating sites as well as through additional networks and social media. The baseline assessment will measure mental well-being, resilience capacities, individual resilience, patient safety perceptions, job demands and resources, work engagement, job satisfaction, leadership support, burnout, intention to quit, sick leave, and health-related quality of life at the recruited sites in each country. The instruments used for health care workers and health care leaders are the Short Warwick-Edinburgh Mental Well-Being Scale (SWEMWBS) [24], the RCS (in development) [25], the Copenhagen Burnout Inventory [26], the Motivation at Work Questionnaire intention-to-turnover measure [27], and the Self-Assessed Health measure [28]. For informal caregivers, the Family Involvement in Care Questionnaire [29] and the Caregiver Self-Assessment Questionnaire [30] will be used (see Multimedia Appendix 1).

The project spans 7 countries with diverse languages (Finnish, Norwegian, Italian, Dutch, Romanian, Spanish, and English). Where validated translations of the instruments already exist, these will be used in the corresponding country. For tools that are not yet available in 1 or more study languages (including the newly developed RCS), we will follow internationally recognized procedures for translation and cultural adaptation [31]. Specifically, the process will include translation into the target language by multiple translators, resolution of discrepancies, expert committee review (researchers from the participating countries), cognitive testing, and piloting with a small sample, with the exception of the back-translation stage of the framework described by Beaton et al [31]. The translation process will be documented in a report.

This approach ensures that the measurement tools are conceptually equivalent across settings while recognizing that the project does not aim for strict cross-country comparability but instead focuses on within-country validity and the context-sensitive evaluation of the S4R Toolbox.

Given the heterogeneity of health care systems across the participating countries, the purpose of these measures is not to enable cross-country score comparisons but rather to provide robust, contextually appropriate assessments of resilience and mental well-being within each national system.

Analysis in WP2 will describe the distribution of mental well-being, resilience capacities, burnout, intention to quit, support, and other relevant variables across countries. The analysis plan includes the following (with ad hoc customizations depending on the stakeholder group: health care workers, leaders, or informal caregivers):

These analyses will provide contextual understanding and empirical input for the development of the S4R Toolbox.

The RCS was developed to capture organizational resilience capacities relevant to older adult care (eg, adaptive capacity, learning, coordination, and supportive structures), grounded in resilience-in-health care theory and organizational approaches to workforce well-being. WP1 covers theoretical specification, item generation, expert review (content validity), translation and cultural adaptation, and cognitive testing and piloting before field deployment.

Psychometric evaluation will be conducted using the WP2 baseline dataset. To avoid circularity, we will use internal cross-validation: exploratory factor analysis will be performed in a development subsample, and confirmatory factor analysis will be conducted in a separate holdout subsample. Reliability (eg, internal consistency) and construct validity will also be assessed. The final scoring algorithm and measurement model will be specified (locked) before the WP4 effectiveness analyses. In WP4, the RCS will be used as a secondary outcome to assess changes in organizational resilience capacities. If minor refinements are made after baseline validation, the WP4 analyses will use the locked version, and sensitivity analyses will be reported where relevant.

Patterns and mechanisms of missing data will be examined. Missing data will be handled using multiple imputation under a missing-at-random assumption. The imputation model will include all analysis variables (primary and secondary outcomes, baseline covariates, treatment group, and cluster identifiers), as well as auxiliary variables associated with missingness, to improve the plausibility of the missing-at-random assumption. Approximately 20-50 imputations will be generated, with the final number guided by the proportion of missing information.

Patterns of missingness will be examined to distinguish monotone (eg, dropout after baseline) and nonmonotone patterns, and appropriate imputation methods will be applied accordingly. For the cRCT, potential differential attrition between the intervention and control groups and across clusters will be assessed and reflected in the imputation model.

Sensitivity analyses will compare imputed results with complete-case analyses and alternative assumptions (eg, pattern-mixture approaches) to evaluate robustness.

Qualitative data collection will take place at the recruited sites in the same 6 European countries and Australia. The data will be collected through semistructured interviews with leaders, health care workers, and informal caregivers to explore their diverse challenges in everyday tasks and care situations, types of ethical stress, adverse events, situational support, needed adaptations, solutions to improve working and caring conditions, how they integrate and experience patient and user involvement, and the role of stakeholders in resilient service performance and in improving involvement. This will include 3-5 focus groups with leaders, 3-5 focus groups with health care workers, and 10 individual interviews with informal caregivers in each participating country (see Multimedia Appendix 2).

The interviews will be conducted in the informants’ native languages to allow for rich data and nuanced insights. The qualitative data will be analyzed using inductive thematic analysis [32], in which the data are analyzed inductively to allow themes and insights to emerge. The findings will be reported by the researchers in each country using a predefined, standardized template in English. This approach will allow for cross-country analysis of patterns across cases, as well as integration and comparison across health care systems and informant groups. Informants will be recruited in collaboration with the participating sites in all countries. Interviews will take place at the workplace for health care workers and leaders and at the preferred location of the informal caregivers. Focus groups will last approximately 2 hours each, whereas individual interviews will last approximately 1 hour each.

The second phase (WP3) will lead to the development of a digital S4R Toolbox for leaders to promote resilient performance and mental well-being among health care workers and informal caregivers. WP3 begins in month 5 and continues through month 30. The features and content of the S4R Toolbox will be developed based on findings from the first input phase, systematic literature reviews, and cross-country qualitative and quantitative data collection. Cross-country data from diverse settings will be used to provide an understanding of factors and challenges that are relevant across settings (residential care homes, home care services, and the hospital-to-home interface) and target groups (leaders, health care workers, and informal caregivers), ensuring that the S4R Toolbox provides value across the included older adult care settings. The development phase applies an iterative co-design methodology using an Agile software development process [33], in which a development team continuously develops the content and features of the S4R Toolbox by organizing 8 collaborative workshops (6 national and 2 transnational) with a co-design panel consisting of consortium partners and collaborators, as well as representatives of the target groups (3-5 health care leaders, 3-5 health care workers, and 3-5 informal caregivers), to ensure relevance and usability.

The S4R Toolbox includes 3 tools: (1) Mapping and Identification Tool (to increase awareness), (2) Reflection and Education Tool (to increase understanding), and (3) Reorganization Tool (to support reframing; Figure 1). While the S4R Toolbox is intended for leaders, the tools rely on the involvement of different stakeholders, including health care workers and informal caregivers, to ensure collaboration across organizational levels and stakeholder groups. All test sites will implement and evaluate the S4R Toolbox through a 1-year intervention program. The S4R Toolbox aims to bring about organizational- and system-level changes through leaders as change agents to improve working conditions for frontline health care workers and informal caregivers. Accordingly, the intervention requires sufficient time for these changes to take place.

This tool will provide several reports and graphs based on data collected from different stakeholders (ie, informal caregivers, frontline leaders, and health care workers) to enable leaders to gain a comprehensive overview of the current status of their organization. The main input will be gathered through questionnaire responses from the involved stakeholders. The questionnaire will be intentionally developed for this tool and will pertain to 3 main areas identified through previous literature reviews: Supporting the Worker, Improving Working Conditions, and Improving Work Culture. The questions will be answered on a 5-point Likert scale and will address areas such as how well the unit fosters learning and professional growth, creates a balance between control and autonomy, develops a shared understanding of professional practice, and promotes communication and positive relationships at work.

Suitable graphs (eg, histograms and bar graphs) will be used to visualize the results, while natural language processing techniques will be used to facilitate leaders’ exploration of the findings. Presenting the data in multiple formats will further enhance the user-friendliness of the tool. The purpose is to provide leaders with a quick, easy-to-use overview of the current status of the unit as a basis for using tool 2.

Building on the same 3 themes addressed in tool 1, this tool is structured into 3 distinct learning modules: Supporting the Worker, Improving Work Conditions, and Improving Work Culture. Based on the scores obtained in tool 1, leaders are encouraged to begin with the module in which their unit demonstrates the strongest performance. Each module incorporates short informational videos, reflective questions, and brief learning scenarios depicting everyday work situations relevant to the thematic area. Through these activities, leaders, together with their staff, are expected to identify existing strengths and the underlying reasons for them.

The overarching aim is to enhance understanding of how work is carried out within the unit, identify resilient and effective practices, and promote improvement by learning from what works well. Once leaders have gained insight into successful practices, they are encouraged to translate this knowledge into targeted actions to address areas that require further development.

This tool addresses a well-known challenge in improvement work: the tendency to continually add new tasks and procedures without removing outdated ones. Over time, this can lead to overlapping practices and reduced system coherence. The aim of tool 3 is therefore to help units identify and safely discontinue practices that no longer add value.

The tool will include a function for gathering input from leaders, staff, and family members on activities they consider unnecessary or burdensome. Based on these insights, leaders will be provided with clear guidance and a structured action plan for removing practices in a responsible and transparent manner. The tool will also offer practical guidance on risk assessment, communication, and monitoring to ensure that discontinuation is carried out safely.

Given the pace of work and the stressful environment of older adult care, the intervention design will be aligned with everyday work to reduce the likelihood that leaders will perceive the intervention as an additional stressor [34]. The S4R intervention will consist of 3 phases, each corresponding to 1 of the 3 tools in the S4R Toolbox. During each phase, frontline leaders will engage with different elements of the tool and participate in different activities (see Multimedia Appendix 3). The implementation of the intervention is planned to take 12 months. Six local learning collaboratives (including consortium researchers and co-researchers from the clinical setting) will be established, 1 in each European partner country, in addition to 1 principal learning collaborative comprising representatives from all 6 local learning collaboratives. Australia will have a separate learning collaborative because of time zone differences.

For each country, a designated intervention team consisting of consortium partners will be formed. The teams will be responsible for administering the local learning collaborative and coordinating the overall implementation activities throughout the intervention in their respective countries.

The S4R intervention design and Toolbox will be open access through the S4R Resource Bank, which includes self-directed instructions on how to implement the intervention. Furthermore, the infrastructure supporting deployment of the Toolbox will ensure server maintenance for 5 years after project completion. This period will also allow time to implement a successful exploitation strategy.

The third phase (WP4) will implement and evaluate the use, relevance, effectiveness, and cost-effectiveness of the S4R Toolbox across the different health care systems involved (months 22-43; see Figures 2 and 4). Implementation of the S4R Toolbox will be carried out in line with the intervention design and the MRC guidelines [13,35-37] (see Multimedia Appendix 4). This design prioritizes the development of interventions that are both theoretically robust and practically adaptable to real-world conditions. The S4R intervention offers a coherent structure—with defined phases, tools, and recommended practices—while still enabling teams to tailor its components to their specific context, consistent with the MRC guidelines. The S4R Toolbox is intentionally flexible: although leaders are encouraged to follow the recommended structure, they may adjust the format, frequency, and integration of its elements to suit the needs and dynamics of their particular setting.

Implementation and evaluation will contribute to exploring the value of the S4R Toolbox in terms of usability, technical features, leadership support, performance, effectiveness, and cost-effectiveness, as well as provide input for the development of policy recommendations and theoretical frameworks. Evaluation of the S4R intervention will follow the study designs used in each country. Four analyses are planned for WP4: (1) process evaluation, (2) pre-post evaluation (test cases 2-4), (3) a cRCT (Norway and Finland, test case 1), and (4) cost-effectiveness evaluation.

A qualitative process evaluation will be performed [13]. Interviews with leaders, health care workers, and informal caregivers conducted in the initial phase of the intervention will be complemented by interviews conducted during and after the intervention period. Additionally, observations will be made during workshops to explore work processes and interactions among participants during implementation of the intervention. To minimize interference with implementation, we will develop standardized informational materials that will be used across all participating units. The researchers will limit their involvement to working directly with leaders through separate workshops and will not be present in the units when the Toolbox is applied in practice. This approach is intended to reduce researcher influence on everyday workflows and ensure that the Toolbox is used as intended by the local leadership. In addition, we will systematically document the contextual circumstances of each unit through context mapping and use standardized observation forms to record events or conditions that may have influenced implementation. These strategies will help differentiate between natural contextual variation and external factors that could interfere with the implementation process.

The data collected will contribute to exploring how the S4R Toolbox supports leaders, how health care workers and informal caregivers experience involvement, and whether capacity-demand challenges improve. In addition, the content of the Toolbox will be explored (eg, comprehension of the learning materials, factors important for sustainability, and levels of engagement and reflection), as well as the nonfunctional and quality aspects of the technical platform (eg, usability, tool design, and accessibility). The knowledge gained will inform further refinements to the Toolbox. Each participating site will include 3-5 focus group interviews with leaders, 3-5 focus group interviews with health care workers (nurses, physicians, and other health care workers), and 10 individual interviews with informal caregivers.

We will assess the effectiveness of the S4R Toolbox through a cRCT (test case 1), a pre-post evaluation (test cases 2-4), and a cost-effectiveness evaluation (test case 1). In both study designs, the primary outcome will be health care workers’ mental well-being, measured using the SWEMWBS. Secondary outcome measures include individual resilience, organizational resilience capacities, burnout, turnover intention, self-assessed health, and, for informal caregivers, family involvement in care and caregiver self-assessment. In addition, sick leave will be assessed using public sick leave records where available. The effectiveness, cost-effectiveness, and cRCT analyses of the S4R intervention will provide information for decision makers in organizations at both the national and European Union (EU) levels.

Changes from baseline to postintervention will be examined using mixed-effects regression models. Individuals will be nested within organizational units, with random intercepts included for organizational units. Time will be entered as a fixed effect. Standardized effect sizes (Cohen d) will be calculated to aid interpretation.

For countries using a pre-post evaluation, a sample of 250-500 participants per country is sufficient to

The S4R Toolbox will be evaluated in a 2-arm cRCT in 2 countries (Finland and Norway). An equal allocation ratio will be used for the 2 groups: an intervention arm, in which the S4R Toolbox is implemented by health care leaders, and a control arm, in which the S4R Toolbox is not implemented by health care leaders. The cRCT will include 30 clusters with approximately 30 health care workers per cluster. Accounting for clustering (ICC=0.05), an expected 30% response rate (approximately 9 respondents per cluster), and 6 repeated measurements (baseline plus 5 follow-up assessments) with a plausible within-person correlation of 0.5-0.7, the cRCT has 95% power to detect standardized effect sizes of approximately d=0.28-0.32 for mental well-being. The 6 measurement time points were chosen to align with the planned phases of Toolbox use and feasible data collection windows in the participating services. This pragmatic schedule enables monitoring of early adoption and engagement, as well as whether changes are sustained over the intervention period.

The sample size calculations assumed an ICC of 0.05 and an average of approximately 9 respondents per cluster (30% response from approximately 30 eligible workers). The assumed ICC of 0.05 was selected based on previous cluster randomized trials in primary care, home care, and organizational intervention studies, in which ICCs for psychosocial and workforce-related outcomes commonly ranged from 0.01 to 0.10. Given the lack of precise prior ICC estimates for mental well-being outcomes in older adult home care services, an ICC of 0.05 was considered a conservative and appropriate planning value. This corresponds to a cluster design effect of 1.40. To assess the robustness of the findings to uncertainty in the ICC, we conducted sensitivity analyses using ICC values of 0.03, 0.08, and 0.10 (design effects of 1.24, 1.64, and 1.80, respectively). Holding all other assumptions constant, the detectable standardized effect size ranged from d=0.26-0.30 (ICC=0.03) to d=0.30-0.35 (ICC=0.08) and d=0.32-0.36 (ICC=0.10), compared with d=0.28-0.32 under the base-case ICC of 0.05. These findings indicate that the trial is sufficiently powered to detect small-to-moderate improvements in the primary outcome.

A 30% response rate was used as a conservative planning assumption for this pragmatic workforce trial in older adult care to reduce the risk of overestimating the effective sample size under real-world conditions. Participation and retention will be supported through leadership engagement, brief, mobile-friendly questionnaires, repeated reminders, and monitoring of response rates at each measurement time point.

The primary outcome measure, the SWEMWBS, will be treated as a continuous outcome. The primary analysis will follow the intention-to-treat principle at the home care service (cluster) level. Noncompliance with the intervention may attenuate the observed intervention effect. Linear mixed-effects models will be used, including random intercepts for clusters. Group (intervention vs control) will be included as a fixed effect, and SWEMWBS (mental well-being) will be the primary outcome. Other outcomes (burnout, turnover intention, caregiver burden, etc) will be analyzed as secondary end points.

Analyses will follow the intention-to-treat principle. A complier average causal effect analysis will also be conducted to account for variation in exposure to the intervention.

Cost-effectiveness analysis will be conducted only for the cRCT study design (Norway and Finland).

The primary outcome will be SWEMWBS, with effectiveness assessed using the same methods described in the previous section.

The costs of the intervention will be based on the time spent by leaders, health care workers, and researchers implementing the S4R Toolbox. The learning collaboratives support the leaders and are therefore expected to influence the effectiveness of the Toolbox. The cost data will be based on the following:

The time horizon for the analysis is 1 year; therefore, no discounting will be applied. The analysis adopts the perspectives of 2 funders: the participating organizations, which cover the costs of leaders’ and health care workers’ time, and the S4R project funding organization (EU), which covers the costs of researchers’ support for implementing the Toolbox.

We use intention-to-treat based analyses and evaluated the incremental net monetary benefit of the Toolbox compared to no Toolbox ("practice as usual"). We evaluated incremental net monetary benefit (NMB) in comparing the alternatives. NMB is calculated by multiplying the observed effects by the decision-makers’ willingness to pay, λ, for an extra unit of effectiveness, and then subtracting the observed cost [37,38]. As the willingness to pay for one extra unit of effectiveness in SWEMWBS is unstated, we use a range of values.

As a sensitivity analysis, we will use an algorithm to convert changes in SWEMWBS scores into WELLBY (Wellbeing-Adjusted Life Year) units [39-41]. WELLBYs have been monetized and can therefore provide a monetary estimate of effectiveness. This analysis will also provide an opportunity to interpret the results of the cost-effectiveness analysis across the range of willingness-to-pay values.

To further support the robustness of our analysis, we will also conduct a cost-benefit analysis. For this analysis, we will collect data from the participating organizations (work units) on sickness absence (spells and days), staff resignations, turnover, disability pensions, and personnel survey results for the years 2023-2026. Where possible, these measures will be monetized using national recommendations, statistics, reports, and research.

The S4R project does not aim to estimate a single pooled effect across countries. The participating health care systems differ substantially in structure, governance, workforce composition, and the availability of formal older adult care services. Because of this heterogeneity, the project follows the MRC guidance for evaluating complex interventions, which explicitly recommends that evaluation designs be adapted to local feasibility and contextual conditions.

In Norway and Finland, organizational structures allow randomization of home care units, making a cRCT feasible and appropriate for evaluating both effectiveness and cost-effectiveness. In the other participating countries, randomization was not feasible because of organizational differences and the absence of comparable units. For these countries, a pre-post evaluation is methodologically appropriate and enables examination of within-country change. This hybrid approach preserves scientific rigor by using the strongest possible design where feasible while ensuring validity and contextual relevance across diverse older adult care systems.