Modernizing health systems with digital technologies offers a practical path to tackle persistent primary health care (PHC) challenges across sub-Saharan Africa (SSA) [1]. Adoption of digital innovations—from mobile health (mHealth) apps to electronic health records—is accelerating and has the potential to improve service delivery, data quality, and access to care, particularly for vulnerable populations [2,3]. However, fragmentation across digital tools can hinder strategic planning, monitoring and evaluation, and data-driven decision-making [4,5]. Real-world deployment also faces substantial barriers, including infrastructure gaps, limited digital literacy among health care workers, and resource constraints [6,7], which create a persistent gap between the promise of digital health and its system-level impact [6,7]. In Burkina Faso, where Primary Health and Social Promotion Centers (Centres de Santé et de Promotion Sociale [CSPSs]) form the backbone of the health system [8], the Minimal Digital Ecosystem (MDE) was introduced to strengthen service delivery, health financing and financial management, medicines oversight, governance, and data use while supporting the national free-care policy (gratuité) [9]. The MDE is not a single application but a set of integrated tools spanning point-of-care and managerial functions (eg, electronic consultation registers, child and maternal care modules, individual electronic care records/Fiche Individuelle de Soins [individual care sheets], free care management, stock and financial management, quality assessment, and community health platforms) [9]. Given this complexity, a rigorous evaluation of implementation is essential for understanding the determinants of success and guiding sustainability and scale-up.

This study addresses two critical gaps in the research on digital health implementation. First, validated, standardized instruments to measure the maturity of complex digital health ecosystems in resource-constrained settings remain scarce [10,11]. Although frameworks such as the Consolidated Framework for Implementation Research (CFIR) and Normalization Process Theory (NPT) enumerate potential determinants, few tools enable their quantification and examination alongside key implementation outcomes (adoption, fidelity, penetration, sustainability) [10,12]. Second, this measurement gap creates a translational challenge: Without evidence-based benchmarks, policymakers cannot define “implementation success” to target support or guide scaling decisions. In Burkina Faso, the widespread, nonrandom rollout of the MDE precluded a traditional quasi-experimental design but created a pragmatic opportunity to answer a stakeholder question from the Ministry of Health: At what threshold can a PHC facility be considered to have fully implemented the MDE? This underscores the need for an evidence-based, measurable definition of “implementation success” to inform resource allocation and policy [10]. Using a pragmatic, mixed methods design [13], this protocol aims to generate a contextualized understanding of MDE implementation, identify levers for optimization, and produce validated instruments to assess implementation maturity—evidence intended to inform digital health policy in Burkina Faso and similar resource-constrained settings.

The MDE evaluation was initially structured into 5 complementary surveys across its implementation stages [9]. The initial phase profiled users, analyzed the implementation, and evaluated the early effects; the second explored the deployment context, including resources, preparedness, user perceptions, and the free access application; and the third assessed deployment levels, utilization, satisfaction, service governance, and quality. Although this multiphase approach provided foundational data, the objectives for the ongoing fourth survey and the prospective fifth were substantively revised in response to real-world implementation challenges that necessitated a shift in the evaluation’s focus.

We adopted a pragmatic paradigm [14], an approach in implementation science that prioritizes useful, real-world solutions over theoretical purity. We aim to generate actionable knowledge in the complex setting of actual practice, not to test a hypothesis under perfect, controlled conditions. Our approach is pragmatic in 3 specific ways.

First, from a methodological standpoint, we embrace flexibility in response to field realities. When intervention contamination threatened the validity of our original quasi-experimental design, we shifted to an implementation-focused, cross-sectional observational evaluation better suited to the real-world context.

Second, from an epistemological standpoint, we integrate multiple frameworks (CFIR 2.0, NPT) and methods (quantitative surveys, qualitative inquiry). We recognize that no single approach can adequately answer the complex questions our stakeholders have.

Third, contextually, we are evaluating the MDE “as deployed” rather than under idealized conditions. We treat the natural variation in implementation not as statistical “noise” to be controlled but as a rich source of insight into what is actually happening on the ground.

This pragmatic orientation ensures that our findings will be directly relevant to policymakers and program managers navigating the complexities of scaling up digital health in resource-constrained settings.

Initially, our evaluation plan for the MDE was a quasi-experimental design. We intended to compare two intervention districts (Ténado and Ziniaré) with two control districts (Manga and Sapouy) to assess the ecosystem’s impact. However, the operational reality of field deployment necessitated a significant deviation from this a priori approach.

However, a key challenge emerged early on: The MDE tools began to progressively and nonrandomly diffuse into our designated control districts. This “contamination” effect fundamentally compromised our original quasi-experimental design—a common problem in real-world implementation studies.

As a result, a robust, direct comparison between the groups became untenable, making it impossible to isolate the MDE’s specific effects. This cross-contamination severely threatened the study’s internal validity; we could no longer confidently attribute observed outcomes to the intervention and faced a high risk of biased or artificially weakened (attenuated) effect sizes.

Consequently, we reoriented the focus of the evaluation. Rather than pursuing a simple effectiveness question, our approach now aligns more closely with an implementation science perspective, which acknowledges and addresses the complexities of real-world contexts [14].

This real-world deployment scenario has raised a critical question from Ministry of Health stakeholders: At what threshold of adoption and routine use can a health facility be considered to have fully implemented the MDE initiative? This new question aligns perfectly with the field’s focus on understanding and measuring implementation outcomes—the direct effects of actions taken to implement an innovation, which serve as key indicators of success [10].

To address both the contamination challenge and this new, emergent research question, we adapted our methodology. Instead of a rigid comparative analysis, we are now using a pragmatic, integrated evaluation design. This shift ensures our evaluation remains relevant and provides actionable insights for future scale-up [15].

Our revised focus will assess the varying levels of MDE tool adoption (or “dose”), its integration into clinical and management workflows, and the maturity of its use across all districts. We will adjust our analytical approach to explore the relationship between these implementation outcomes and health service outcomes [10]. This will allow us to better understand the associations and mechanisms, even in the context of widespread tool availability.

Ultimately, this approach will provide stakeholders with more nuanced and actionable insights into how to optimize MDE implementation to achieve greater impact on the health system.

Specifically, this study aims to (1) assess the maturity of MDE implementation, (2) identify multilevel determinants, and (3) evaluate the association between MDE implementation and health facility outcomes.

To assess the maturity of MDE implementation, we will evaluate the level of adoption; fidelity to its intended use; penetration into existing practices, perceived sustainability, and institutionalization of its various tools. We will use the NPT framework to guide our understanding of institutionalization.

To identify multilevel determinants, we will use the CFIR 2.0 to identify the facilitators and barriers that influence the degree and quality of MDE implementation. This will encompass a broad range of factors, including intervention characteristics, outer and inner settings, individual user characteristics, and the implementation process itself.

To evaluate the association between MDE implementation and health facility outcomes, we will assess the relationship between the degree of MDE implementation and the readiness and quality of health facility services. This evaluation will particularly focus on key areas such as service delivery, financial and medication management, the efficiency of free health care provision, and the use of data for decision-making.

Building on the study’s objectives to evaluate the implementation and effects of the MDE, we propose the following 4 hypotheses to guide our investigation.

Hypothesis 1 (H1) involves implementation maturity variance. We hypothesize that the maturity of MDE implementation will vary significantly across different MDE tools, CSPS, and the districts. This reflects the complexity of real-world implementation contexts.

H2 involves multilevel determinants. We hypothesize that the degree and quality of MDE implementation are associated with multilevel determinants. Specifically, positive factors at the inner setting level (eg, engaged leadership, adequate resources, a favorable implementation climate) and within the implementation process (eg, quality of training, ongoing support) will be positively correlated with a higher degree and better quality of MDE implementation.

H3 involves user perceptions and adoption. We hypothesize that the degree of MDE implementation is associated with user perceptions of the intervention’s characteristics. A positive perception of MDE characteristics (eg, relative advantage, low complexity) by individual users will be positively correlated with higher adoption and utilization rates of MDE tools, contributing to a greater overall degree of implementation.

H4 involves implementation and facility outcomes. We hypothesize that a higher degree of MDE implementation will be positively associated with improved health facility readiness and performance. This association will be particularly evident in enhanced service delivery, more efficient financial and medication management, better management of free health care services, and increased utilization of data for informed decision-making within CSPS.

This study is being conducted in 4 health districts in Burkina Faso: Ziniaré (Plateau Central region), Ténado and Sapouy (Centre-Ouest region), and Manga (Centre-Sud region). Although these districts were initially selected for a quasi-experimental design [16], the widespread, real-world deployment of the MDE has reframed the study. This collection of districts now serves as a setting with valuable natural variation in implementation levels [17], allowing us to assess varying levels of MDE implementation and identify the multilevel determinants influencing its adoption and institutionalization across diverse settings.

The MDE represents a novel approach to digital health implementation in SSA. Unlike standalone digital interventions, such as isolated electronic medical record systems or single-purpose mHealth apps, the MDE deliberately integrates 9 complementary tools into a cohesive ecosystem spanning clinical care, supply chain management, financial tracking, quality assurance, and community health [9]. The “minimal” designation reflects a strategic choice: Rather than implementing a comprehensive, resource-intensive hospital information system, Burkina Faso’s Ministry of Health prioritized essential, interoperable tools that can be deployed at the PHC level within existing infrastructure [9]. This ecosystem model has not been systematically evaluated elsewhere in SSA, making this study among the first to rigorously assess the implementation of an integrated digital health suite—rather than discrete applications—in a low-resource setting. Additionally, although Burkina Faso has previously piloted individual digital health tools (eg, DHIS2 for health information, isolated mHealth interventions), the MDE’s distinguishing feature is the intentional harmonization and codeployment of multiple tools, designed to generate synergies in data flow, reduce parallel workflows, and support holistic PHC strengthening [9].

The inclusion of all 4 districts enables the study to leverage their distinct characteristics—such as urban/rural profiles, population density, and pre-existing digital infrastructure—to robustly address the implementation-focused research objectives. This ensures a comprehensive evaluation that considers a broad range of contexts.

The intervention under evaluation is the MDE, a suite of integrated digital tools designed to strengthen PHC in the 4 study districts of Burkina Faso. The MDE aims to enhance operational efficiency, service quality, governance, and data-driven decision-making by harmonizing various digital applications into a cohesive system. The core tools of the MDE are detailed in Table 1 and Figure 1.

This study is not an interventional clinical trial; however, the MDE is a complex intervention for which the implementation is being evaluated. Accordingly, we adapted the SPIRIT 2013 guidance to structure the protocol, particularly the sections describing the intervention, objectives, outcomes, target population, and analysis plan. We also completed the TIDieR (Template for Intervention Description and Replication) checklist to enable precise reporting of interventions. The finalized checklists are provided as Checklist 1.

The MDE’s theory of change posits that core health system functions will be digitized and streamlined by deploying an integrated suite of digital tools and supporting their use through targeted training. This is expected to improve clinical protocol adherence, financial and supply chain efficiency, and systematic data collection at the facility and community levels. By aggregating this newly captured, real-time data into a central dashboard, the MDE then aims to foster evidence-based decision-making among managers. The ultimate goal of this combined improvement in service quality, operational efficiency, governance, and data utilization is to produce a modernized and strengthened PHC system capable of delivering higher-quality and more equitable care.

This study will use a cross-sectional mixed methods design [18] guided by a pragmatic approach [14]. This design is particularly well-suited to our revised objectives, as it allows us to pivot from a rigid comparative analysis to an observational one, an ordinary and necessary adjustment when researchers cannot fully control interventions in real-world settings [19]. The design will be implemented in two distinct phases where the quantitative findings inform the subsequent qualitative inquiry.

The first phase of the study will consist of a cross-sectional quantitative survey designed to address both the primary and secondary objectives. This phase will assess the extent of MDE implementation across all participating health facilities and districts. To achieve this, we will quantify key implementation outcomes based on the CFIR, with a specific focus on adoption, fidelity, penetration, and sustainability. The findings will support the development of a typology of implementation profiles, which will inform and guide the strategic deployment of a digital health ecosystem in Burkina Faso (Table 2).

We will estimate the associations between the implementation level and (1) multilevel determinants specified by CFIR 2.0 and NPT and (2) key health facility performance indicators using appropriate multivariable models.

This phase will answer the “what” and “how much” questions, providing a broad overview of implementation maturity and identifying the significant relationships outlined in our hypotheses (H1-H4).

The second phase will consist of a qualitative inquiry designed to explain and elaborate upon the findings from the quantitative phase. Using a multiple-case study approach, we will purposively select a subset of health facilities that demonstrate particularly high, low, or mixed levels of implementation based on the quantitative results. This phase will explore the “how” and “why” behind the statistical patterns through semistructured interviews, focus groups, and direct observation. For example, it will help us understand why specific MDE tools are well-integrated while others are not, how specific facilitators (eg, leadership engagement) operate in practice, and what contextual barriers hinder the translation of the MDE into improved performance.

Following quantitative analysis, we will use maximum variation sampling to select facilities representing the full implementation spectrum. Target qualitative sample sizes, informed by prior implementation research and saturation principles [25-27], are described in the following sections.

The study’s primary and secondary outcomes are defined and measured as described in the following sections.

Data for this study will be collected using a suite of mixed methods instruments designed to capture both the degree of MDE implementation and its underlying determinants.

Prior to the main data analysis, a rigorous process will be undertaken to develop and validate the quantitative measurement scales from the questionnaire data [34]. This process applies to the composite implementation scores for each MDE tool and for the MDE overall, as well as to scales measuring key implementation determinants (eg, constructs from the CFIR framework).

We are evaluating the measurement properties in line with COSMIN (Consensus-Based Standards for the Selection of Health Measurement Instruments) guidance, covering content validity, structural validity, internal consistency, reliability, measurement error, and hypothesis-testing construct validity. Content validity was addressed through expert review and piloting (relevance, comprehensiveness, comprehensibility) [35]. For structural validity, we will first assess sampling adequacy (Kaiser-Meyer-Olkin≥0.70; Bartlett test P<.001) and conduct exploratory factor analysis (EFA) using polychoric correlations with minimum residual extraction and oblique rotation; item retention requires primary loading ≥0.40, cross-loading difference ≥0.30, and communality ≥0.30. We will then run confirmatory factor analysis (CFA) using weighted least squares mean and variance adjusted for ordinal Likert items (or maximum likelihood/maximum likelihood robust estimation if distributions are approximately continuous), reporting fit indices with a priori thresholds of comparative fit index/Tucker-Lewis index ≥0.90 (preferably ≥0.95), root mean square error of approximation≤0.08 (preferably ≤0.06), and standardized root mean squared residual≤0.08; models failing these will be revised only based on theory and prespecified modification rules [36]. Internal consistency will be summarized using Cronbach α and McDonald ω, targeting α and ω values of 0.70 or higher for newly adapted scales and 0.80 or higher where feasible [28]. For multidimensional scales, α/ω will be reported per subscale; when a bifactor structure is indicated, we will report ωH (omega hierarchical).

Only scales demonstrating satisfactory psychometric properties (ie, strong validity and reliability) will be retained for subsequent hypothesis testing and regression analyses.

Data collection will take place over 1 month. Quantitative data (via questionnaires) will be collected first, followed by qualitative data (including interviews, focus groups, and observations). System usage data (logs) will be collected retrospectively and prospectively during the study period.

Quantitative data will be collected electronically using the Open Data Kit platform, with forms programmed to include quality controls such as skip logic, input validation, and required fields. Encrypted data will be submitted daily to a secure central server. Qualitative audio recordings will be transcribed verbatim and fully anonymized by replacing personal identifiers with unique codes. All study data will be stored on a secure, password-protected server with access restricted to the research team and will be securely archived after the study, in accordance with relevant ethical guidelines.

We anticipated and proactively managed typical field constraints. To address geographic barriers across 72 often-remote facilities, data collection was scheduled before the rainy season, local enumerators were recruited, and contingency transport plans were prepared. Respondent availability was improved through pre-arranged appointments during low-volume periods and multiday visits at busier sites. Technology limitations were mitigated with offline-enabled forms, portable chargers, and standardized backup procedures. To minimize attrition, we used a census approach and engaged stakeholders throughout, yielding a 94.3% response rate. Qualitative saturation will be monitored in real time by two researchers conducting concurrent analysis, with flexible recruitment to target emerging gaps. Risks of social desirability bias are reduced through neutral-questioning training, use of external data collectors, and triangulation with system usage logs and observational notes. Timeline feasibility is supported by deploying 32 enumerators concurrently over 4 weeks with strong district-level coordination. Finally, coding reliability will be ensured through dual independent coding of 25% of transcripts, followed by consensus meetings to reconcile discrepancies and balance analytic rigor with interpretive depth.

In line with this study’s sequential explanatory design (quantitative → qualitative) [18], the quantitative and qualitative strands will be integrated at multiple points. First, the quantitative results will directly inform the qualitative phase by guiding the purposive sampling of cases (eg, high- and low-performing sites) and the development of tailored interview and focus group guides. At the final interpretation stage, both datasets will be synthesized to generate meta-inferences [44], and team synthesis meetings will explicitly identify convergent and divergent findings. Joint displays linking facility-level quantitative implementation profiles with dominant qualitative themes will be constructed. Note that qualitative findings will inform the interpretation of quantitative associations but will not be used to respecify quantitative models post hoc [45]. Qualitative data will be used to explain and contextualize quantitative statistical patterns, while triangulation across the two strands will serve to confirm findings and strengthen the overall validity of the study’s conclusions [46].

This study received full ethical approval from the National Ethics Committee for Health Research of Burkina Faso (approval number 2023-06-136) and obtained additional authorizations from the Ministry of Health and relevant district authorities. Data will be immediately anonymized using unique codes to ensure confidentiality and stored on secure, password-protected servers with restricted access, and all consent forms will be stored separately. No financial compensation was provided; however, travel costs for focus group participants were reimbursed. Additionally, no personal identifiable information will be used in any publications.

Prior to participation, all individuals provided informed consent, either written or, for nonliterate participants, witnessed verbal consent, after being fully informed of the study’s objectives, their right to withdraw, and all confidentiality measures.

Given the minimal-risk, nonclinical nature of this study, a formal Data and Safety Monitoring Board has not been constituted. Any unexpected adverse events or breaches of confidentiality will be reported to the Ethics Committee for Health Research in accordance with national regulations.

This study will produce two categories of outcomes: (1) empirical findings on the maturity and determinants of MDE implementation and (2) methodological products—validated, ready-to-use measurement scales and scoring procedures.

This study will generate a nuanced portrait of implementation maturity across PHC facilities and districts, capturing adoption, fidelity, penetration, and perceived sustainability while quantifying between-facility heterogeneity through multilevel models (including ICCs and random-intercept variance). It will further produce an evidence map of multilevel determinants, anchored in CFIR 2.0 domains and complemented by NPT-informed patterns of normalization, integrating quantitative estimates with qualitative explanations. We will also document the extent to which MDE tools are institutionalized and routinely used within clinical and managerial workflows and examine their associations with PHC readiness and performance indicators (service delivery, stockouts, financial reporting, and data use).

Although the study uses a cross-sectional design at each measurement wave, the standardized and psychometrically validated instruments are purposefully designed for repeated administration. This means that implementation maturity can be measured at successive time points across the same facilities, enabling a repeated cross-sectional architecture that supports change-on-change analytical strategies—comparing shifts in contextual determinants with shifts in implementation scores across waves. This capacity to move beyond static, single-point associations toward identifying factors associated with improvement over time represents a methodological advance rarely achieved in digital health evaluations in SSA and positions the protocol as a foundation for prospective monitoring of the MDE’s trajectory.

The validated implementation measures enable a bidimensional strategic profiling of facilities along two complementary axes—implementation quality (how well the tools are used, as captured by psychometrically validated scores) and utilization (the extent to which available tools are actually integrated into routine practice). This dual profiling can generate a maturity typology (Success, Adoption Challenge, Deployment Challenge, Critical Zone) that translates complex statistical outputs into an actionable, intuitive classification directly usable by policymakers and district health management teams to target differentiated support strategies and monitor progress over successive evaluation cycles.

Together, these findings will culminate in actionable, context-specific recommendations to optimize and scale the MDE tailored to district and facility conditions.

This study will develop a suite of validated scales to quantify implementation outcomes for each MDE tool and for the ecosystem as a whole, following COSMIN guidelines (content and structural validity, reliability, measurement error, and hypothesis-testing construct validity). For each scale, we will report full psychometric evidence, including EFA/CFA results with model fit indices and item loadings, internal consistency (α and ω), and ωH where a bifactor structure is indicated—alongside measurement invariance checks across districts and staff cadres. Building on these metrics, we will construct a composite MDE implementation index with alternative weighting schemes (equal vs policy-informed weights), accompanied by prespecified sensitivity analyses. To facilitate uptake and reuse, user materials will be packaged for dissemination: Finalized instruments with item wording, a scoring codebook, and worked examples of score calculations will be provided for routine monitoring and future studies.

Collectively, these outputs will enable policymakers and program managers to define, measure, and track “implementation success” using transparent, psychometrically sound tools. The empirical findings will guide the prioritization of support (training, supervision, infrastructure) to accelerate the effective scale-up of the MDE.

This study will advance implementation science for complex digital health interventions in resource-constrained settings by delivering a CFIR 2.0– and NPT-guided, mechanism-focused analysis of the multilevel determinants that shape adoption, integration into routine PHC workflows, and long-term institutionalization; generating generalizable insights across districts; and contributing methodological assets in the form of psychometrically validated implementation measurement scales (tool-specific and composite) with transparent scoring rules and invariance checks. Together, these outputs will provide an evidence-based roadmap for policymakers and practitioners—identifying leverage points such as leadership engagement, readiness, and support systems—to prioritize investments in training, supervision, infrastructure, and data use, thereby bridging the gap between the technological potential of the MDE and measurable, system-level impact.

Although this study is conducted in Burkina Faso, its findings are expected to have transferability to similar contexts. The study’s generalizability stems from three features: First, theoretical grounding in CFIR 2.0 and NPT—frameworks validated across diverse settings—ensures findings are interpreted through broadly applicable constructs rather than context-specific idiosyncrasies. Second, the validated measurement instruments produced by this study are intentionally designed for adaptability; items reference generalizable implementation constructs (eg, “perceived ease of use,” “organizational readiness”) rather than Burkina Faso–specific details. Third, the multilevel analytical approach explicitly models contextual variation (between districts, facilities, individuals), allowing identification of which findings are setting-dependent versus universal.

That said, we acknowledge limitations to generalizability: The MDE’s specific configuration (choice of 9 tools, CommCare/DHIS2 platforms) may differ from digital ecosystems elsewhere, Burkina Faso’s health system structure (CSPS as the PHC backbone, the gratuité policy) may not mirror other countries, and infrastructure constraints in the study districts may be more or less severe than other SSA settings.

Researchers and policymakers in other contexts should adapt the study’s approach—integrated ecosystem evaluation using CFIR/NPT with validated scales—rather than directly applying the specific findings without contextual consideration. Our instruments and methods provide a replicable template while acknowledging that determinants and thresholds of “implementation success” may vary by setting.

The expected results carry implications from near-term operational improvements to long-term health system strengthening. Identifying the factors most strongly associated with implementation outcomes would justify targeted actions—such as upgrading infrastructure, refining workflows, and institutionalizing ongoing professional development—to deliver a more robust and responsive MDE rollout. Over time, sustained use of the MDE is poised to enhance data quality, streamline clinical and administrative processes, and improve accountability and resource use. Collectively, these gains can translate into measurable improvements in access to and quality of essential PHC services, inform scalable strategies for digital health investment, and contribute to system resilience and equity in resource-constrained settings.

This protocol has several strengths that enhance its scientific rigor and policy relevance. First, the sequential explanatory mixed methods design (quantitative → qualitative) integrates the breadth of quantitative findings with the descriptive depth of qualitative inquiry, enabling not only the measurement of implementation maturity but also the explanation of the mechanisms underlying observed patterns. Second, the explicit dual grounding in CFIR 2.0 and NPT provides complementary analytical lenses—CFIR 2.0 to identify what multilevel factors shape implementation and NPT to illuminate how digital tools become normalized (or fail to become normalized) within clinical and managerial workflows. Third, the multilevel data collection strategy captures perspectives from providers, facility managers, community health workers, and health service users across 72 facilities and 4 districts, enabling robust triangulation and sufficient statistical power for multilevel modeling. Fourth, the rigorous psychometric development and validation of all implementation measurement scales—including EFA and CFA, reliability assessment, and measurement invariance testing across districts and staff cadres—ensure that observed differences reflect genuine variation in implementation rather than measurement artifacts. Finally, the pragmatic orientation of the study—evaluating the MDE “as deployed” rather than under idealized conditions—ensures that findings are directly applicable to the real-world contexts in which scale-up decisions will be made.

This protocol has several potential limitations. First, reliance on self-reported data introduces risk of social desirability bias; we mitigate this through intensive enumerator training, neutral interviewing, and strict anonymity procedures. Second, the cross-sectional, observational survey design is suitable for characterizing implementation at a point in time but precludes causal inference; all multilevel model estimates reflect statistical associations rather than causal effects. The subsequent qualitative phase is intended to contextualize and explain observed associations. Third, generalizability may be constrained to settings with comparable health system structures and digital readiness. Finally, logistical challenges—such as reaching remote sites and accommodating busy clinical schedules—may affect recruitment and timing, necessitating flexible fieldwork plans and persistent follow-up.

We will pursue a multi-audience strategy. For policymakers and program managers, comprehensive evaluation reports and concise policy briefs will be delivered to the Ministry of Health, district authorities, and funding partners. For frontline providers and communities, we will convene interactive feedback workshops to review the results and discuss the practical implications. To contribute to the global evidence base, we will publish in reputable peer-reviewed journals (including this protocol) and present at major international health conferences. To engage the public and media, we will produce plain-language summaries and fact sheets for broad distribution. Where permissible, we will share an anonymized dataset under a Data Use Agreement to support secondary analyses and reproducibility.

This protocol outlines a pivotal study to evaluate the implementation of a national digital health ecosystem in Burkina Faso. By generating robust evidence on implementation maturity, fidelity, and their key determinants, the research will directly inform optimization of the MDE and guide future digital health strategies. Ultimately, these insights are expected to support stronger, more equitable PHC across SSA.