From stakeholder mapping to statistical modeling: An illustrative demonstration of end-to-end Net-Map methodology for health governance analysis

From stakeholder mapping to statistical modeling: An illustrative demonstration of end-to-end Net-Map methodology for health governance analysis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Method Article From stakeholder mapping to statistical modeling: An illustrative demonstration of end-to-end Net-Map methodology for health governance analysis Bianca-Elena Mihăilă, Marian-Gabriel Hâncean, Marius Geantă, Cosmina Cioroboiu, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7837431/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Coordinating primary cancer prevention in Europe requires alignment across the EU, national, and local actors. Yet governance relations are challenging to observe. We used a participatory network approach to make relationships visible and test which ties are associated with funding. This study serves as a methodological demonstration to illustrate the complete Net-Map research cycle from data collection to statistical analysis. Methods Three Net-Map sessions (January–February 2024) with a national NGO expert panel elicited stakeholders and ties for authority, influence, and funding, followed by validation. Canvases were used to record the information, and they were digitized afterwards into edge and attribute files. We described network structure and estimated exponential random graph models (ERGMs) for the directed funding network, including overall tie propensity, governance level (EU, national, local), and dyadic influence and authority relations. Results The final network comprises 128 organizations and 836 funding ties (density 5.14%). Funding is predominantly cross-level (71.4%). EU actors show the strongest outward activity (EU-national density 17.8%; EU-local 16.1%), while national-local is lower (6.2%). In ERGMs, influence is the strongest correlate of funding (OR 168.54, p < 0.001), authority is positive but smaller (OR 5.76, p < 0.001), and same-level funding is less likely (OR 0.35, p < 0.001). Model comparisons favored the full specification; structural goodness-of-fit was poor, typical for large, sparse networks, so we interpret effects as dyadic associations. Conclusions This illustrative analysis claims that participatory Net-Map workshops, combined with formal network analysis are helpful in diagnosing coordination patterns in decentralized health systems. The findings draw from a single embedded panel and cross-sectional design that can be extended in future interventions. Still, they are primarily serving as an example of methodological application rather than providing generalizable insights about Romanian cancer prevention governance. Health Policy Preventive Medicine Net-Map methodology stakeholder mapping primary cancer prevention public health policy social network analysis Figures Figure 1 Figure 2 Background Net-Map was developed as a participatory way to visualize stakeholder networks and power relations [ 1 , 2 ]. Particularly, it combines stakeholder engagement with social network analysis to generate structured, analyzable data. Applications span food systems, child nutrition, reproductive health, obesity prevention, and natural resource governance [ 3 – 7 ]. These studies show their value for mapping influence, authority, and resource flows in complex settings. Recent work argues for using Net-Map to integrate relational evidence into policy analysis [ 8 , 9 ]. Despite its promise, several gaps remain in the literature. First, most applications have been in the Global South (e.g., Ghana, Nigeria, Ecuador, Nepal), while European health governance has received little attention [ 3 , 7 , 10 – 12 ]. Second, few studies have applied Net-Map across the full research cycle: while the participatory phase is well developed, systematic integration with advanced statistical analysis remains rare, limiting its contribution to policy and systems research. Cancer provides a compelling test case for addressing these gaps. Globally, there were 19.3 million new cancer cases in 2020, with projections of 28.4 million annually by 2040 [ 13 ]. Europe, home to less than 10% of the world's population, carries 25% of this burden [ 14 ]. In the European Union (EU), incidence and mortality continue to rise [ 15 ], with particularly high rates in Eastern Europe. In Romania, cancer is the second leading cause of death, responsible for 19% of all deaths in 2022 [ 16 , 17 ]. Between 30% and 50% of cancer cases are considered preventable [ 18 ]. The European Commission [ 19 ] and the World Health Organization [ 20 ] highlight primary prevention strategies, including reducing tobacco and alcohol use, promoting healthier diets and physical activity, vaccination, and mitigating environmental risks. Implementing these strategies requires coherent governance frameworks that extend beyond biomedical interventions to address structural determinants of health. Yet, responsibilities for prevention are distributed across local, national, and supranational actors, including ministries, regulatory agencies, professional associations, civil society, patient groups, and EU institutions. This complexity often produces fragmented strategies, weak coordination, and inequitable allocation of resources. Consequently, this makes cancer prevention governance a vital framework for understanding how decentralized health systems operate. Analyzing such governance processes requires approaches that can identify key stakeholders, map relationships of influence, collaboration, authority, and financial flows, and make visible both formal and informal power dynamics. Net-Map [ 1 , 21 ] is particularly well suited to this task. While it has been used to study governance in other domains, it remains underutilized in European health systems and has not previously been applied to cancer prevention. This underuse is notable, as European health systems often privilege hierarchical decision-making and technical expertise, limiting opportunities for participatory analysis of governance [ 22 ]. Our study addresses these gaps by applying Net-Map to primary cancer prevention governance in Romania, with a focus on stakeholders involved in implementing the European Code Against Cancer [ 20 , 23 ]. Building on recent research into social determinants and prevention [ 24 – 26 ], we combine participatory mapping with advanced social network analysis (exponential random graph models) to examine how influence and authority ties are associated with directed funding across governance levels (EU, national, local). We contribute to health policy and systems research in three ways: (1) we offer one of the few European applications of participatory network mapping in public health, (2) we extend Net-Map across the full research cycle (from stakeholder elicitation and consensus validation to reproducible, model-based analysis), and (3) we show how multi-level structures and informal influence relate to the distribution of prevention resources, generating transferable, system-diagnostic insights for strengthening collaborative governance in decentralized health systems. Together, these contributions position Net-Map not only as a methodological innovation but also as a diagnostic tool for informing prevention policy and guiding more equitable allocation decisions in complex systems. Methods Exercise design and facilitation Net-Map is a participatory interview-based mapping method that integrates social network analysis with influence mapping to visualize stakeholder relationships and perceived power dynamics [ 1 ]. In this study, we applied Net-Map to analyze the multi-layered governance of primary cancer prevention in Romania, with a particular focus on stakeholders involved in implementing the European Code Against Cancer [ 20 ]. The choice of Net-Map was motivated by two considerations. First, Romania's health system context: while cancer is the country's second leading cause of death, the absence of a national cancer registry has led to fragmented data on incidence, outcomes, and survival. In this environment, prevention is crucial, yet it requires coordinated action across a fragmented governance landscape. Lessons from the Karelia Project on cardiovascular prevention argued that prevention strategies are more effective when supported by collaborative networks [ 27 , 28 ]. Second, the methodological rationale: given the diversity and number of actors involved, Net-Map provides a systematic means of eliciting, visualizing, and analyzing governance networks. The methodological choices prioritized demonstrating the complete Net-Map workflow over maximizing representativeness. This analysis serves as an illustrative case study to show researchers how to implement the full research cycle. The Net-Map exercise was conducted through three in-person sessions (January–February 2024) in Bucharest, Romania. Each session lasted approximately five hours. Splitting the exercise into multiple meetings minimized participant fatigue and allowed for reflection between stages. Namely, eliciting stakeholders and detecting their relationships are cognitively demanding tasks for participants [ 29 , 30 ]. Session 1 (January 20 2024) consisted of identifying relevant actors and mapping within-level relationships (i.e., relationships among organizations of the same level). Session 2 (January 26, 2024) focused on the mapping of cross-level relationships (EU, national, and local). Session 3 (February 2, 2024) addressed the validation of the network maps. All three sessions were facilitated by a lead sociologist with extensive experience in qualitative sociology, supported by two assistants. Consistent with Net-Map principles, facilitators ensured clarity in problem framing, guided discussions, and moderated dynamics to avoid dominance by individual participants. All sessions were video-recorded to capture participant reasoning and support coding. Participants and setting We purposively recruited five experts from the same non-profit organization (NGO), a national organization with longstanding, multi-level engagement in primary cancer prevention (local implementation partnerships, national policy dialogue, and interfaces with European initiatives). This embedded, practice-based panel was selected for three reasons. First, the NGO maintains continuous operational links across levels and sectors, providing participants with comprehensive visibility into stakeholders and inter-organizational ties relevant to the European Code Against Cancer. Second, shared technical language and institutional memory facilitated the efficient and high-fidelity elicitation of complex networks within the time and cognitive constraints of Net-Map. Third, panel members covered complementary functional roles (policy design/regulation, financing/commissioning, implementation/coordination, data analysis and reporting), ensuring breadth of perspectives despite a single organizational home. Eligibility required ≥ 10 years of experience in public health/health policy, direct involvement in primary prevention and/or European Code Against Cancer (ECAC) related activities, and routine collaboration with actors at local, national, and European levels. We did not aim for statistical representativeness; the goal was to provide credible coverage of system-critical perspectives, thereby generating and validating the stakeholder map. We recognized the risk of shared-organization bias (common frames or blind spots). To address this, we employed a scripted facilitation protocol with turn-taking and probe questions, conducted three sessions to facilitate reflection and correction, required explicit consensus on actors and ties, and performed structured digitization and cross-checks of the map. We also triangulated the elicited actor list against public policy documents, e.g., national cancer plans, ECAC guidance, and documented disagreements and their resolution. We note that Net-Map can be implemented with individual key informants or small groups. We chose a small-group format to enable triangulation and consensus-building while managing cognitive load. Larger panels should be intentionally avoided: they have an increased risk of proving logistically and cognitively unwieldy (crowded canvases, slower facilitation, harder consensus, higher coding error), which jeopardizes data quality. Using a single-NGO panel may over-represent organizational perspectives. We address this by transparently reporting recruitment, providing open materials/code, and emphasizing that findings are diagnostic and meant to inform further multi-stakeholder validation rather than to claim population representativeness. Concerning the expert panel, there were no conflicts of interest. Additionally, the experts did not receive any honoraria for their study participation. In total, twelve individuals were present at the sessions: five expert participants, one lead facilitator, two assistant facilitators, and four student observers responsible for logistics. To minimize potential bias, all experts were consulted equally during the mapping. Problem framing followed a standardized script derived from Schiffer [ 1 ] to ensure consistency. We conducted the sessions in Romanian (including the mapping). The dataset is translated into English. Three guiding questions structured the mapping: (1) Who are the actors responsible for implementing the European Code Against Cancer in Romania? (2) What relationships connect them (i.e., financial flows, influence, authority)? (3) What are their primary motivations or roles in relation to cancer prevention? Stakeholders were classified into four categories: government ( public ), private , academia , and civil society [ 29 – 31 ]. Actors were also categorized by governance level: local (Argeș county), national (Romania), and European (European Union). This study is designed as a methodological illustration. The choice of a single-NGO panel, while providing operational advantages for demonstrating the Net-Map process, necessarily limits the generalizability of findings. The results should be understood primarily as showing how the methodology works rather than providing comprehensive insights into the Romanian healthcare system Ethics All research procedures complied with the Declaration of Helsinki and the European General Data Protection Regulation (GDPR). The study protocol was approved by the Ethics Committee of the Center for Innovation in Medicine (EC-INOMED Decision No. D001/09-06-2023 and No. D001/19-01-2024). Written informed consent was obtained from all participants. We anonymized personal identifying information after each interview and securely stored data on encrypted drives accessible only to authorized personnel. Materials and mapping procedure Mapping was carried out using large sheets of paper, post-it notes (for stakeholder names), and colored markers (for relational ties). Stakeholders were written on post-its, attached to the canvas, and linked according to three types of relationships. Namely, authority , i.e., formal power or decision-making capacity (policy enforcement, regulation, command); influence , i.e., ability to shape others' decisions or actions through expertise, advocacy, or strategic positioning; financial resources , i.e., transfer of funding or economic support. In other Net-Map studies, the number of nominated stakeholders varied from 38 [ 32 ] to 83 [ 6 ]. In our case, we did not limit the experts in nominating stakeholders. Without a cap in the stakeholder elicitation process, the final map included 128 actors across governance levels. Parts of the original canvas are presented in Fig. 1 . Data structuring and preparation The hand-drawn maps (the stakeholders and their relationships) were digitized into two datasets. An edge list file included source and target actors with relational type (financial flow, influence, authority), binary coded (1 = tie present, 0 = absent). An attribute file consisted of actor ID, name, organizational category (coded as: public, private, non-profit, and academia), and governance level (local, national, and European). Digitization was conducted in two different working sessions. Student assistants entered data into Excel, followed by cross-checking against the original maps. Experts validated the digital dataset in the third meeting. The dataset was then imported into RStudio (2023.06.1), where we made use of various packages for the analysis (e.g., igraph , statnet , ergm ). Reliability and validation We used a two-step reliability procedure. First, a consensus mapping. During elicitation, participants jointly discussed stakeholder inclusion and ties. Ties were recorded only when all participants reached explicit consensus. Disagreements were resolved through targeted prompts and, where applicable, by consulting official documents (e.g., legal mandates, program guidelines). Items without consensus were flagged and revisited in the validation session. Only ties with full agreement were retained in the final dataset. Second, double data entry and adjudication . We implemented double data entry and adjudication: hand-drawn maps were independently digitized by pairs of observing student assistants, and discrepancies were reconciled by returning to the original canvases, session notes, and video; when ambiguity persisted, we consulted the expert panel. We did not compute a formal inter-coder agreement statistic during digitization because the task was a transcription of consensus-based, binary ties using pre-specified definitions. We acknowledge this as a potential limitation and recommend that future studies complement consensus and double-entry checks with blind recoding and κ/α estimates on a random subset. Based on our experience, limiting the number of stakeholders and tie types per session helps contain cognitive load and transcription errors. While Net-Map is stakeholder-driven, facilitator influence is inherent [ 33 ]. To reduce bias, we: (i) used a standardized script and predefined guiding questions, (ii) moderated dominant participants while encouraging quieter voices, (iii) validated all actors and ties across three sessions, and (iv) required participant consensus before finalizing the maps. Observers (students) remained passive and uninvolved in discussions. Statistical analysis Prior to modeling, we calculated standard network descriptive statistics, including density (ratio of observed and theoretically possible ties) and cross-level tie patterns for each relationship type (authority, influence, financial flows). We examined the overlap between different relationship types to understand how formal and informal mechanisms relate to resource allocation patterns. We specifically analyzed cross-level coordination patterns by calculating the proportion of ties occurring within governance levels versus across levels (EU-National, EU-Local, National-Local) for each relationship type. Given the directed nature of funding relationships, we examined hierarchical resource distribution patterns by analyzing funding flows in both directions between governance levels (upward vs. downward funding). We used Exponential Random Graph Models (ERGMs) [ 34 ] to examine which relational mechanisms are associated with the presence of a directed funding tie between organizations ( i funds j ). ERGMs treat the observed network as one realization from a reference distribution in which the log-odds of a tie depend on a set of network statistics capturing baseline tie propensity (the tendency of tie formation within the observed network), homophily (the tendency of same-level organizations to share a tie), and dyadic covariates (influence and authority ties). This approach accounts for the interdependence of ties, an inherent feature of social and governance networks, and yields coefficients interpretable as log-odds (reported alongside odds ratios). The dependent variable was the binary directed funding network derived from the Net-Map sessions. All nodes appearing on the map were retained; multiple mentions were reconciled during digitization and validation. Self-ties were not permitted. We estimated four nested ERGM specifications to assess the relative importance of different predictors: (1) baseline model containing only network density ( edges ), (2) adding governance level homophily effects ( same governance level ), (3) incorporating formal authority relationships, and (4) full model including informal influence relationships alongside all previous predictors. Table 1 briefly introduces the model terms. Table 1 ERGM parameter specifications for funding network analysis Term Label (parameter) Interpretation Baseline density Edges (baseline) This captures the overall tendency to form funding ties Same-governance level Nodematch(level) This tests whether actors at the same level (EU, national, local) are more or less likely to exchange funding. Given our focus on cross-level coordination, a negative coefficient indicates a preference for cross-level funding. Dyadic covariates (edgecov) This incorporates other mapped relations as predictors of funding. These covariates test whether influence and authority co-occur with funding beyond baseline density and level structure. Influence ties edgecov(influence) 1 if i reports influencing j Authority ties edgecov(authority) 1 if i has formal decision power over j Note: Definitions and interpretations of exponential random graph model terms used to predict funding relationships in the Romanian cancer prevention governance network. The baseline density parameter captures overall tie formation tendency, while governance level homophily tests preferences for within-level versus cross-level funding coordination. Dyadic covariates examine whether existing authority and influence relationships predict funding ties beyond baseline network structure. Models were estimated using maximum pseudolikelihood (MPLE) as implemented in the ergm package. For interpretability, we present log-odds coefficients with standard errors (SE) and odds ratios (OR). We compared the full specification against nested models that omit each dyadic predictor in turn to assess incremental contribution using AIC/BIC comparisons. We evaluated (i) numerical convergence and parameter stability (including seed sensitivity) and (ii) structural Goodness of Fit (GOF) via simulated vs. observed distributions of degree, shared partners, and geodesic distances. In large, sparse governance networks, global structural GOF often remains imperfect even when dyadic effects are stable; accordingly, we interpret ERGMs as identifying dyadic association patterns rather than attempting to reproduce all higher-order structure. The full R code, as well as the data files, is available in the Additional file 1 [ 35 ]. Results The following results illustrate what researchers might expect to find when applying Net-Map methodology, rather than definitive conclusions about Romanian cancer prevention governance. The patterns described demonstrate the analytical capabilities of the combined Net-Map–ERGM approach. Table 2 reports the descriptive statistics of the networks generated by the panel of experts in the Net-Map exercise. Specifically, the Romanian cancer prevention governance network comprises 128 organizations distributed across three administrative levels: 63 national-level organizations (49.2%), 40 local-level organizations (31.3%), and 25 European-level organizations (19.5%). Table 2 Network structure and funding flow patterns in Romanian cancer prevention governance. Characteristic Value Characteristic Value Network Composition Cross-Level Funding Patterns Total organizations 128 Within-level ties 239 (28.6%) European level 25 (19.5%) Cross-level ties 597 (71.4%) National level 63 (49.2%) Cross-level percentage 71.4% Local level 40 (31.2%) Funding Flows by Direction Network Structure European → National 280 (17.8% density) Total funding ties 836 European → Local 161 (16.1% density) Possible dyads 16,256 National → Local 155 (6.2% density) Network density 0.0514 (5.14%) National → European 1 (0.1% density) Reciprocity 0.01 (1.0%) Local → National 0 (0.0% density) Local → European 0 (0.0% density) Relationship Networks Authority ties 150 Relationship Overlap with Funding Authority density 0.0092 (0.92%) Authority-funding overlap 43/150 (28.7%) Influence ties 1,012 Influence-funding overlap 647/1,012 (63.9%) Influence density 0.0623 (6.23%) Note: Illustrative network structure and funding flow patterns in Romanian cancer prevention governance. These descriptive statistics demonstrate the type of insights that can be generated through Net-Map analysis. Here, we report descriptive statistics for the multi-level cancer prevention governance network (N = 128 organizations across European, national, and local levels). Cross-level funding comprises 71.4% of all funding relationships, with European organizations showing the highest outward funding activity, connecting to 17.8% of national organizations and 16.1% of local organizations. Upward funding is virtually absent, confirming hierarchical resource distribution. Influence relationships show greater overlap with funding ties (63.9%) than authority relationships (28.7%), suggesting informal networks better predict resource flows than formal structures. Overall network connectivity = 5.14%; reciprocity = 1.0%. The funding network exhibits low overall density (5.14%), with 836 observed funding relationships among 16,256 possible directed dyads. Reciprocal funding relationships are rare (1% reciprocity), indicating predominantly unidirectional resource flows. Two distinct relationship networks were mapped alongside the funding network. The authority network contains 150 formal hierarchical relationships (density = 0.92%), while the influence network is substantially denser with 1,012 informal influence ties (density = 6.23%). The influence network density exceeds the funding network density, suggesting that influence relationships extend beyond direct resource allocation patterns. The funding network shows strong hierarchical coordination patterns, with 597 ties (71.4%) occurring across governance levels rather than within them. European-level organizations demonstrate the highest outward funding activity, directing resources to both national (280 ties from 1,575 possible connections, 17.8%) and local levels (161 ties from 1,000 possible connections, 16.1%). National-to-local funding is substantially lower (155 ties from 2,520 possible connections, 6.2%), while upward funding from local to national or European levels is virtually absent (1 tie total). This pattern claims a clear hierarchical resource distribution model where higher administrative levels fund lower ones. The overlap between relationship types and funding patterns reveals important differences in how formal and informal mechanisms translate into resource allocation. Influence relationships show substantial overlap with funding ties (647/1,012, 63.9%), indicating that informal influence networks closely mirror actual resource flows. In contrast, authority relationships show limited overlap with funding (43/150, 28.7%), suggesting that formal hierarchical structures are less predictive of actual resource allocation than informal influence patterns. These descriptive statistics illustrate that while formal authority structures exist, the governance network operates primarily through cross-level resource flows guided more by influence relationships than formal organizational hierarchies. The plot panel in Fig. 2 illustrates the distribution of ties (money, influence, and authority) among the stakeholders identified during the Net-Map exercise. To illustrate how exponential random graph models can be used to understand mechanisms driving funding relationships, we estimated a series of ERGM models. These results exemplify the analytical insights possible through this methodological approach. Thus, to understand the mechanisms driving funding relationships in Romanian cancer prevention governance, we estimated a series of exponential random graph models (ERGMs) that progressively incorporated different theoretical predictors of tie formation. We started with a baseline model (Model 1) containing only the network density parameter (edges). Further, we sequentially added governance-level homophily (Model 2), formal authority relationships (Model 3), and informal influence relationships (Model 4) to assess their relative importance in explaining funding ties among organizations. Specifically, Table 3 presents the results of the four nested ERGM specifications. Model 1 establishes the baseline probability of funding ties in the absence of any structural predictors. Model 2 introduces governance-level effects to test whether organizations preferentially fund within or across administrative levels (local, national, and European). Model 3 adds formal authority relationships to examine whether hierarchical organizational structures predict resource flows. Model 4 presents the full specification, incorporating informal influence networks alongside formal authority and governance level effects. Table 3 Authority vs. Influence: Predictors of funding ties in health governance networks. Parameter Model 1 Baseline Model 2 Cross-level Model 3 Authority Model 4 Full Model Edges (baseline) -2.915*** -2.777*** -2.831*** -4.140*** (SE) (0.036) (0.042) (0.043) (0.077) Same governance level — -0.417*** -0.387*** -1.040*** (SE) (0.078) (0.079) (0.112) [OR] 0.66 0.68 0.35 Authority relationship — — 1.989*** 1.751*** (SE) (0.185) (0.305) [OR] 7.31 5.76 Influence relationship — — — 5.127*** (SE) (0.106) [OR] 168.54 Log-likelihood -3295.0 -3280.1 -3238.2 -1615.6 AIC 6592.1 6564.3 6482.4 3239.3 BIC 6599.8 6579.7 6505.5 3270.1 Deviance 6590.1 6560.3 6476.4 3231.3 n (dyads) 16,256 16,256 16,256 16,256 n (edges) 836 836 836 836 n (nodes) 128 128 128 128 Note: Exponential random graph models predicting funding relationships among 128 organizations in Romanian cancer prevention governance. Coefficients are shown with standard errors (in parentheses) and odds ratios [in brackets]. Influence relationships are the strongest predictor of funding ties (OR = 168.54), vastly exceeding authority relationships (OR = 5.76). Organizations prefer cross-level funding coordination (OR = 0.35). All models were estimated using MPLE. N = 16,256 dyads; 836 funding ties. ***p < 0.001 The full model reveals several key patterns in network formation processes. Most strikingly, informal influence relationships emerge as the dominant predictor of funding ties, with organizations connected by influence relationships being 168.54 times more likely to have funding connections (p < 0.001). This effect vastly exceeds that of formal authority relationships, which increase funding odds by 5.76 times (p < 0.001). Put differently, influence relationships are 29.3 times more predictive than authority relationships (168.54 ÷ 5.76 = 29.3) in the flow of money. The governance level parameter indicates a strong preference for cross-level coordination, with same-level funding relationships being significantly less likely (OR = 0.35, p < 0.001). These findings suggest that while formal organizational hierarchies matter for resource allocation, informal influence networks are far more predictive of actual funding flows. The preference for cross-level funding coordination indicates that cancer prevention governance operates through hierarchical mechanisms rather than horizontal collaboration within administrative levels. Model diagnostics and quality assessment are available in Table 4 . All ERGM models converged successfully using maximum pseudolikelihood estimation (MPLE). Parameter stability tests across different random seeds (12345, 54321, 99999) showed coefficient variation below 0.1 for all parameters, indicating highly stable estimates. Standard errors were successfully computed for all coefficients, enabling reliable statistical inference. Model comparison revealed substantial improvement from the baseline to full specification. The full model (AIC = 3,239.3) dramatically outperformed the baseline model (AIC = 6,592.1), representing an improvement of 3,352.8 AIC points. This consistent improvement indicates that relationship-based predictors capture fundamental network formation processes not explained by baseline density alone. Specification sensitivity tests confirmed the necessity of both authority and influence predictors. Removing the influence network from the full model increased AIC by 3,243.2 points, indicating that this predictor is critical for model performance. Removing authority relationships increased AIC by 27.7 points, showing this predictor also contributes meaningfully to model fit, though to a lesser degree than influence relationships. Table 4 ERGM Model Diagnostics and Quality Assessment. Diagnostic Value Interpretation Diagnostic Value Interpretation Model Convergence Model Specification Tests Estimation method MPLE Reliable estimation Full vs. No Authority (ΔAIC) 27.7 Authority contributes MCMC convergence Successful No convergence issues Full vs. No Influence (ΔAIC) 3,243.2 Influence essential Parameter stability (CV) < 0.1 (Stable) Highly stable results Influence effect necessity Critical Cannot omit influence Authority effect necessity Significant Authority adds value Model Fit Comparison Network Coverage Baseline model AIC 6,592.1 Poor baseline fit Observed ties modeled 836/836 (100%) Complete coverage Full model AIC 3,239.3 Substantial improvement Dyads analyzed 16,256 Full network analyzed AIC improvement (ΔAIC) 3,352.8 Very large improvement Missing data None Complete data BIC improvement (ΔBIC) 3,329.7 Strong model preference Goodness-of-Fit Assessment Estimation Quality Total GOF tests 635 Comprehensive testing Standard errors available Yes Robust estimation Significant deviations 468 Many significant tests Confidence intervals computed Yes Full inference possible Percentage significant 73.7% Poor overall fit P-values < 0.001 (all params) Yes Highly significant effects Overall fit assessment Poor structural fit Dyadic patterns reliable Parameter Precision Authority coefficient SE 0.305 Moderate precision Influence coefficient SE 0.106 High precision Cross-level coefficient SE 0.112 High precision All z-values > 2 Yes Strong statistical power Note: Diagnostic assessment of ERGM estimation quality and model specification. All models converged successfully with stable parameters (CV < 0.1). The full model outperforms baseline specifications (ΔAIC = 3,352.8), with influence relationships essential (removing increases AIC by 3,243.2) and authority relationships significant (ΔAIC = 27.7). Goodness-of-fit testing shows poor structural fit (73.7% of 635 tests are significant), common in sparse networks, but dyadic relationship patterns remain reliable. All parameters demonstrate high significance (p < 0.001) and adequate precision Goodness-of-fit testing using degree distributions, shared partner statistics, and geodesic distance revealed poor overall structural fit, with 468 of 635 tests (73.7%) showing significant deviations from the observed network. This pattern is common in large, sparse networks where ERGMs struggle to capture all structural dependencies. However, this limitation does not invalidate the dyadic relationship patterns that are the primary focus of our analysis. The poor structural fit indicates that while our model successfully identifies which types of relationships predict funding ties, it may not fully capture higher-order network properties like clustering patterns or degree distributions. This is consistent with ERGM limitations in modeling complex network structures with simple dyadic predictors. All relationship parameters demonstrated high statistical significance (p < 0.001) with adequate precision. The influence relationship parameter showed particularly high precision (SE = 0.106), while the authority relationship parameter had moderate precision (SE = 0.305). The cross-level preference parameter also showed high precision (SE = 0.112), indicating reliable estimation of governance level effects. The ERGM analysis reliably captures dyadic relationship patterns and their relative importance for funding tie formation. However, the poor goodness-of-fit suggests caution in interpreting results about global network properties. The models are best understood as identifying which dyadic characteristics predict funding relationships rather than explaining the full complexity of network structure. The cross-sectional design limits causal inference about network formation processes. The models identify associations between relationship types and funding patterns, but cannot determine whether influence relationships lead to funding or whether funding creates influence relationships. Discussion We demonstrate the methodological application of Net-Map to map governance of primary cancer prevention in Romania and illustrate how funding relationships can be modeled using ERGMs. This methodological demonstration reveals three types of patterns researchers might expect when applying this approach. First, funding may be predominantly cross-level: in this illustrative case, 71.4% of ties span EU, national, and local levels. Second, influence ties may emerge as the strongest correlate of funding. In this demonstration, influence relationships are 29.3 times more predictive than authority relationships of the flow of money. Third, authority ties may matter but potentially less so (OR = 5.76). Same-level funding may be disfavored (OR = 0.35) while the funding network remains sparse (density 5.14%). These illustrative findings suggest that resources flow downward and follow informal influence more than formal hierarchy. As an illustrative case, this analysis shows how influence appears to organize coordination and access, acting as a practical mechanism for moving money across levels. The methodological demonstration indicates that authority contributes, but potentially far less. The cross-level pattern indicates reliance on vertical linkages rather than horizontal collaboration within levels. A further pattern exemplifies insights relevant for system design: EU-Local flows (16.1% of possible dyads) are present but smaller than EU-national flows (17.8%). In this illustrative setting where influence strongly predicts funding, the relative paucity of direct EU–local ties and the lack of upward connections suggest weak feedback from implementers to EU programs. This example provides grounds to strengthen direct EU–local connectivity, not to bypass national roles, but to shorten paths, improve information flow, and align funding with delivery realities. This illustrative analysis suggests three types of implications that might emerge from our Net-Map study. First, the methodology reveals that formal mandates alone may not predict allocation. These findings suggest that policy teams should identify influential brokers early in program design. Doing so can shorten implementation chains and reduce coordination losses. Second, the analytical approach indicates the importance of strengthening vertical interfaces, with explicit EU–local links. Most funding is cross-level. Current flows favor EU-national more than EU-local. These examples demonstrate how establishing clearer EU–local contact points, co-design sessions, and direct technical channels can improve alignment while keeping national stewardship. Third, the methodology shows that authority networks do not map cleanly onto funding. The analytical approach suggests that budget processes should incorporate structured stakeholder checks before finalizing allocations, especially when new prevention actions are introduced. The primary contribution of this work is methodological. We demonstrate an end-to-end workflow linking participatory mapping to model-based inference. Net-Map generated a validated list of actors and ties. ERGMs then quantified which relations are associated with funding while accounting for interdependence. This illustration shows that the approach is transparent and reproducible and appears appropriate for decentralized systems where formal documents under-represent how decisions are made. Additionally, this methodological demonstration describes patterns of allocation rather than distributional fairness. We therefore refrain from equity claims. Future work should add explicit equity metrics (e.g., funding in-degree by level or sector; concentration indices) and relate them to prevention needs to test whether cross-level flows are also fair flows. As an illustrative methodology demonstration, several important limitations must be emphasized. Most critically, the network reflects the perceptions of a single, highly embedded NGO panel. This methodological choice prioritized demonstrating the complete research workflow over achieving representative findings. While this brings strong system knowledge for illustrative purposes, it creates significant risks of shared blind spots that limit generalizability. We mitigated this through scripted facilitation, consensus rules, triangulation with documents, and multi-session validation, yet residual bias is possible. The primary purpose of this analysis is to show how the methodology works rather than to provide definitive insights into Romanian health governance. The design is cross-sectional; therefore, causality cannot be inferred. Influence could enable funding, but funding could also create influence. ERGM structural goodness-of-fit is poor, which is common in large, sparse graphs. As stated, we interpret models as identifying dyadic associations, not reproducing higher-order structure. These choices keep the model interpretable but may omit additional dependencies. Several steps would extend this methodological approach. First, broaden participation to include government, clinical, and local implementers, followed by cross-panel comparison to achieve more representative findings. Second, link administrative data (budgets, grants, contracts) to the mapped ties to validate reported funding flows. Third, track change over time using repeated mapping or longitudinal models to examine how reforms, budgets, or EU initiatives alter governance patterns. These extensions would transform the illustrative demonstration into a comprehensive empirical analysis suitable for policy conclusions. Conclusion This methodological demonstration illustrates that primary cancer prevention governance can be analyzed through cross-level relationships in which informal influence appears to align with funding flows. The illustrative findings suggest that formal authority helps but may not determine allocation. In this example, the lower EU-local density, coupled with the central role of influence, demonstrates the type of insights that might provide grounds to strengthen direct EU-local connectivity. This Net-Map–ERGM workflow demonstrates a practical diagnostic approach to reveal where coordination occurs and where to target improvements within decentralized health systems. Future applications of this methodology with broader, more representative samples would be needed to generate definitive policy insights about specific health governance systems. Declarations Data availability The dataset analyzed in the current study and the R code are made openly available, as an additional file, in the Zenodo data repository as Mihăilă, B.-E., et al. Replication data and code for stakeholder mapping and Net-Map health governance analysis. Zenodo https://doi.org/10.5281/zenodo.17329451 (2025). Funding B.E.-M., M.-G.H., M.G., C.C., B.-A.V., A.T.-L., and I.O. were supported by the 4P-CAN project, HORIZON-MISS-2022-CANCER-01, project ID 101104432, programme HORIZON. Views and opinions expressed are those of the authors only and do not necessarily reflect those of the European Union or the granting authority. Neither the European Union nor the granting authority can be held responsible for them. J.L. was supported by Deutsche Forschungsgemeinschaft (DFG 555455503). Acknowledgement We thank Dr. Bianca Cucoş, Florin Găină and Simona-Elena Puncioiu for their valuable contributions to this study. Competing interests The authors declare no competing interests. References Schiffer E. Manual: Net-Map toolbox: influence mapping of social networks. Presented at the Sunbelt Conference of the International Network for Social Network Analysis. 2007. https://netmap.wordpress.com/wp-content/uploads/2008/06/net-map-manual-long1.pdf. Accessed 16 May 2024. Schiffer E, Waale D. Tracing power and influence in networks: Net-map as a tool for research and strategic network planning. 2008. Washington, DC. International Food Policy Research Institute. Aryeetey R, Harding K, Hromi-Fiedler A, Pérez-Escamilla R. Analysis of stakeholder networks for breastfeeding policies and programs in Ghana. Int Breastfeed J. 2020;15(1):74. doi:10.1186/s13006-020-00311-x. Mahmood H, Suleman Y, Hazir T, Akram DS, Uddin S, Dibley MJ, et al. Overview of the infant and young child feeding policy environment in Pakistan: Federal, Sindh and Punjab context. BMC Public Health. 2017;17:474. doi:10.1186/s12889-017-4341-5. Machado JG, Buccini G, Recine E. An analysis of key actor networks for scale-up strategies for childhood obesity prevention and the care of children with obesity in Brazil. Curr Dev Nutr. 2023;7(7):101961. doi:10.1016/j.cdnut.2023.101961. Litwan K, Lara-Mejía V, Chahine T, Hernández-Cordero S, Vilar-Compte M, Pérez-Escamilla R. An analysis of actors participating in the design and implementation of workplace breastfeeding interventions in Mexico using the NetMap analysis approach. Front Public Health. 2023;11:1192600. doi:10.3389/fpubh.2023.1192600. Hauck J. Managing social-ecological systems for resilience: fisheries in the small reservoirs of northern Ghana [dissertation]. Rheinische Friedrich-Wilhelms-Universität Bonn; 2010. https://nbn-resolving.org/urn:nbn:de:hbz:5N-23570. Accessed 26 Jun 2024. Boyle E, Ó Gallachóir B, Mullally G. Participatory network mapping of an emergent social network for a regional transition to a low-carbon and just society on the Dingle Peninsula. Local Environ. 2021;27(12):1431-45. doi:10.1080/13549839.2021.1936472. Cortés-Calderón SV, López-Rodríguez MD, Jiménez-Aceituno A, Castro AJ, Mancilla-García M. Contributions of Net-Map to sustainability action research. Curr Opin Environ Sustain. 2025;75:101542. doi:10.1016/j.cosust.2025.101542. Schiffer E, Mustapha AY, Mustaph AL. Planning, budgeting and disbursing funds for newborn survival in Katsina State, Nigeria – a Net-Map analysis. 2012. https://netmap.files.wordpress.com/2007/11/schiffer_12_net-map_case_study_newborn_survival_nigeria.pdf. Accessed 24 Jul 2024. Rosales Ap, Linnemann AR, Luning PA. A Net-Map analysis to understand the roles and influence of stakeholders in street food safety—a study in Ecuador. Food Control. 2023;154:109966. doi:10.1016/j.foodcont.2023.109966. Karn S, Devkota MD, Uddin S, Thow AM. Policy content and stakeholder network analysis for infant and young child feeding in Nepal. BMC Public Health. 2017;17:421. doi:10.1186/s12889-017-4340-6. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209-49. doi:10.3322/caac.21660. European Parliament. Fighting cancer in the EU: statistics and action (infographics). 2022. https://www.europarl.europa.eu/topics/en/article/20200131STO71517/fighting-cancer-in-the-eu-statistics-and-action-infographics. Accessed 29 Jul 2024. EU Science Hub. Cancer cases and deaths on the rise in the EU. 2023. https://joint-research-centre.ec.europa.eu/jrc-news-and-updates/cancer-cases-and-deaths-rise-eu-2023-10-02_en. Accessed 29 Jul 2024. OECD. Romania: country cancer profile. 2023. https://www.oecd.org/content/dam/oecd/en/publications/reports/2023/02/eu-country-cancer-profile-romania-2023_b7601b86/267467c6-en.pdf. Accessed 29 Jul 2024. Global Cancer Observatory. Romania: fact sheet. 2022. https://gco.iarc.who.int/media/globocan/factsheets/populations/642-romania-fact-sheet.pdf. Accessed 29 Jul 2024. World Health Organization. Preventing cancer. 2025. https://www.who.int/activities/preventing-cancer. Accessed 27 Sep 2024. European Commission. Cancer prevention. 2024. https://knowledge4policy.ec.europa.eu/cancer/cancer-prevention-2024_en. Accessed 27 Sep 2024. Espina C, Ritchie D, Feliu A, Canelo-Aybar C, D’Souza E, Mitrou PN, et al. Developing evidence-based cancer prevention recommendations: methodology of the World Code Against Cancer Framework to create region-specific codes. Int J Cancer. 2025;1-10. doi:10.1002/ijc.70068. Hauck J, Schiffer E. Between intuition and indicators – using Net-Map for visual and qualitative social network analysis. In: Gamper M, Reschke L, Schönhuth M, editors. Knoten und Kanten 2.0: soziale Netzwerkanalyse in der Medienforschung und der Kulturanthropologie. Bielefeld: Transcript; 2012. p. 231-257. Yasnoff WA, O’Carroll PW, Friede A. Public health informatics and the health information infrastructure. In: Shortliffe EH, Cimino JJ, editors. Biomedical informatics: computer applications in health care and biomedicine. New York: Springer; 2006. p. 537-63. doi:10.1007/0-387-36278-9_15. Schüz J, Espina C, Villain P, Herrero R, Leon ME, Minozzi S, et al. European Code against Cancer 4th edition: 12 ways to reduce your cancer risk. Cancer Epidemiol. 2015;39 Suppl 1:S1-10. doi:10.1016/j.canep.2015.05.009. Oană I, Hâncean MG, Perc M, Lerner J, Mihăilă BE, Geantă M, et al. Online media use and COVID-19 vaccination in real-world personal networks: quantitative study. J Med Internet Res. 2024 Oct 25;26:e58257. doi:10.2196/58257. Mihăilă BE, Hâncean MG, Perc M, Lerner J, Oană I, Geantă M, et al. Cross-sectional personal network analysis of adult smoking in rural areas. R Soc Open Sci. 2024;11:241459. doi:10.1098/rsos.241459. Hâncean MG, Lerner J, Perc M, Molina JL, Geantă M, Oană I, et al. Processed food intake assortativity in the personal networks of older adults. Sci Rep. 2025;15:10459. doi:10.1038/s41598-025-94969-0. Puska P, Vartiainen E, Nissinen A, Laatikainen T, Jousilahti P. Background, principles, implementation, and general experiences of the North Karelia Project. Glob Heart. 2016;11:173-8. doi:10.1016/j.gheart.2016.04.010. Puska P, Jaini P. The North Karelia Project: prevention of cardiovascular disease in Finland through population-based lifestyle interventions. Am J Lifestyle Med. 2020;14:495-9. doi:10.1177/1559827620910981. McCarty C, Killworth PD, Rennell J. Impact of methods for reducing respondent burden on personal network structural measures. Soc Networks. 2007;29:300-15. doi:10.1016/j.socnet.2006.12.005. Stadel M, Stulp G. Balancing bias and burden in personal network studies. Soc Networks. 2022;70:16-24. doi:10.1016/j.socnet.2021.10.007. Aminullah A, Wusko AU. Pentahelix model to create shared value in empowering the people of Winong Village, Gempol District, Pasuruan Regency. Global Review of Tourism and Social Sciences. 2025;1:59-67. doi:10.53893/grtss.v1i2.348. Cassaniti J, Kumoji K, Rariewa F, Ohkubo S, Oyenubi O. Influence networks relating to health knowledge among Nairobi’s micro-retailers and their clients. Electron J Knowl Manag. 2021;18:302-24. doi:10.34190/ejkm.18.3.2068. Choi BC, Pak AW. A catalog of biases in questionnaires. Prev Chronic Dis. 2005;2(1):A13. Lusher D, Koskinen J, Robins G, editors. Exponential random graph models for social networks: theory, methods, and applications. Cambridge: Cambridge University Press; 2013. doi:10.1017/cbo9780511894701. Mihăilă B-E, Hâncean M-G, Geantă M, Cioroboiu C, Lerner J, Molina JL, et al. Replication data and code for stakeholder mapping and Net-Map health governance analysis. Zenodo; 2025. doi: 10.5281/zenodo.17329451 Additional Declarations The authors declare no competing interests. Supplementary Files SupplementaryMaterial.html Additional information 1 Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7837431","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Method Article","associatedPublications":[],"authors":[{"id":528179991,"identity":"9c452335-e226-4ac7-9379-0535516110b1","order_by":0,"name":"Bianca-Elena Mihăilă","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABGklEQVRIie3QPUvDQBjA8ScELku0a+R8+QonhYZiMR/E5Q7BLhYyOXUoBOKizZrNr1ARnB85iJtZA+mQLpkzhQ5STLAVhFMzOtx/ucuRH5cnADrdP4yZ2w1xALCA0e4cOxIOVx3I164hwEHuHn8mrmWWxXoKF/s0WCGfpl4vQkJ9fymi2d6qUJBhQNzTuwQm4WHCkCe5iDNOaMxKEaPVZwrCJAwokIY4vCEk55ABoTaTYgGkHVBBrJrCpiXjCvnmzTtJ8S9iD6gRtuSaoQjRWDR3/UqGgX1zcD93WuKjmF+Kx0wEZzYr+7EkpmoWt/f67Kzr0eQhHj9VVX3uHafyJbffl0fRbWio/ti2bx9gzD5XU/GiTqfT6Tr1AYM/YgOneU5BAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0003-4930-450X","institution":"Department of Sociology, University of Bucharest","correspondingAuthor":true,"prefix":"","firstName":"Bianca-Elena","middleName":"","lastName":"Mihăilă","suffix":""},{"id":528181747,"identity":"d9b84784-ee4c-428c-a5f1-651e29de07df","order_by":1,"name":"Marian-Gabriel Hâncean","email":"","orcid":"https://orcid.org/0000-0001-9358-5287","institution":"Department of Sociology, University of Bucharest","correspondingAuthor":false,"prefix":"","firstName":"Marian-Gabriel","middleName":"","lastName":"Hâncean","suffix":""},{"id":528181748,"identity":"e0ac3c17-06d9-4ad4-8c58-2aead1da951f","order_by":2,"name":"Marius Geantă","email":"","orcid":"https://orcid.org/0000-0003-1510-8885","institution":"Center for Innovation in Medicine","correspondingAuthor":false,"prefix":"","firstName":"Marius","middleName":"","lastName":"Geantă","suffix":""},{"id":528181749,"identity":"3239b614-d3c1-4dfd-a8ac-975d4c1457b7","order_by":3,"name":"Cosmina Cioroboiu","email":"","orcid":"https://orcid.org/0000-0002-7959-4926","institution":"Center for Innovation in Medicine","correspondingAuthor":false,"prefix":"","firstName":"Cosmina","middleName":"","lastName":"Cioroboiu","suffix":""},{"id":528181750,"identity":"384c9ba3-c7ad-41b6-ad53-ff16d42052bb","order_by":4,"name":"Jürgen Lerner","email":"","orcid":"https://orcid.org/0000-0001-6565-1873","institution":"Department of Computer and Information Science, University of Konstanz","correspondingAuthor":false,"prefix":"","firstName":"Jürgen","middleName":"","lastName":"Lerner","suffix":""},{"id":528181751,"identity":"e2d00f2c-062e-403b-81ab-152c119708d7","order_by":5,"name":"José Luis Molina","email":"","orcid":"https://orcid.org/0000-0002-2932-2690","institution":"GRAFO, Department of Social and Cultural Anthropology, Universitat Autònoma de Barcelona","correspondingAuthor":false,"prefix":"","firstName":"José","middleName":"Luis","lastName":"Molina","suffix":""},{"id":528181752,"identity":"59089cf3-a76b-4e6d-926a-bc3b8a1e4160","order_by":6,"name":"Antonin Tron-Lozai","email":"","orcid":"","institution":"Department of Sociology, University of Bucharest","correspondingAuthor":false,"prefix":"","firstName":"Antonin","middleName":"","lastName":"Tron-Lozai","suffix":""},{"id":528181753,"identity":"ad8a3e44-c5ae-48f3-8577-9e0b66d115b0","order_by":7,"name":"Bogdan-Adrian Vidrașcu","email":"","orcid":"","institution":"Department of Sociology, University of Bucharest","correspondingAuthor":false,"prefix":"","firstName":"Bogdan-Adrian","middleName":"","lastName":"Vidrașcu","suffix":""},{"id":528181754,"identity":"92fd78ae-0ca5-4056-909f-e2af0d1d980c","order_by":8,"name":"Iulian Oană","email":"","orcid":"https://orcid.org/0000-0002-6052-6595","institution":"Department of Sociology, University of Bucharest","correspondingAuthor":false,"prefix":"","firstName":"Iulian","middleName":"","lastName":"Oană","suffix":""}],"badges":[],"createdAt":"2025-10-11 20:56:52","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-7837431/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7837431/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":93459210,"identity":"0d5ccf08-6067-44ff-b5fa-c5d8cb12d999","added_by":"auto","created_at":"2025-10-14 05:54:18","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":722523,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7837431/v1/2f0c41737c5a664010c23949.docx"},{"id":93459208,"identity":"5f4db065-9c73-42ef-961a-e396d8d018b0","added_by":"auto","created_at":"2025-10-14 05:54:17","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342,"visible":true,"origin":"","legend":"","description":"","filename":"rs7837431.json","url":"https://assets-eu.researchsquare.com/files/rs-7837431/v1/84358c57cdc014a241f94624.json"},{"id":93459215,"identity":"262c4fbc-237d-4063-bd3f-ac40167bdb0a","added_by":"auto","created_at":"2025-10-14 05:54:18","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":129241,"visible":true,"origin":"","legend":"","description":"","filename":"rs78374310enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7837431/v1/b46548327c9210691ef43424.xml"},{"id":93459859,"identity":"79717b29-011a-41a3-8c36-69ca7b2a275a","added_by":"auto","created_at":"2025-10-14 06:02:18","extension":"jpeg","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":787488,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7837431/v1/162fe3a53bc24bc8abebbc52.jpeg"},{"id":93459862,"identity":"eda42e42-091c-4280-88a0-b730a634c3de","added_by":"auto","created_at":"2025-10-14 06:02:18","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":731152,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7837431/v1/a1e9d213b437479102c1a56c.jpeg"},{"id":93459860,"identity":"02d990c0-47e0-43d8-b9d6-8d9b4211314d","added_by":"auto","created_at":"2025-10-14 06:02:18","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":279000,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7837431/v1/b91689dfe1f83259dc6de840.png"},{"id":93459861,"identity":"67104e55-fc9b-4d90-8669-ba2aa4555b3b","added_by":"auto","created_at":"2025-10-14 06:02:18","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":193315,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7837431/v1/58de0dfae82d939744bf1066.png"},{"id":93459219,"identity":"347d5d91-b042-444f-9b69-3b6fd7ee4a72","added_by":"auto","created_at":"2025-10-14 05:54:18","extension":"xml","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":126051,"visible":true,"origin":"","legend":"","description":"","filename":"rs78374310structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7837431/v1/50fd3bae3f1011e0394a0cda.xml"},{"id":93459217,"identity":"3e01d8e8-2474-40a0-bcd9-9b5788106ab6","added_by":"auto","created_at":"2025-10-14 05:54:18","extension":"html","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":136102,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7837431/v1/b1833c7e2901f9946fb9b2a0.html"},{"id":93459214,"identity":"a4b5b106-d9f7-4f6b-8ad2-4d0d6e2db5bb","added_by":"auto","created_at":"2025-10-14 05:54:18","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":8777501,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAn illustration of the multi-level stakeholder network elicited by the Net-Map exercise experts.\u003c/strong\u003ePost-it colors mark the type of stakeholders (private, public, non-profit, and academic). Colored lines define the type of relationships connecting the stakeholders (funding, influence, authority). The canvas is split into three to delineate the stakeholders' levels (local, national, and EU).\u003c/p\u003e","description":"","filename":"Fig1.png","url":"https://assets-eu.researchsquare.com/files/rs-7837431/v1/91ed8269e3a9b028daf61c00.png"},{"id":93459212,"identity":"adf4eab4-f621-4120-84f8-9cac5d3510c1","added_by":"auto","created_at":"2025-10-14 05:54:18","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":5088483,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMulti-level Stakeholder networks in primary prevention health policy: authority, financial, and influence relationships. \u003c/strong\u003eThe panel depicts node-and-line network visualizations of the money, authority, and influence multi-level ties among the 128 stakeholders identified by the panel of experts. Networks are organized by governance level (EU, National, Local) with nodes colored by organizational level: EU level (pink), National level (blue), and Local level (orange). Edge characteristics indicate relationship direction and scope: solid arrows represent cross-level ties, while dotted lines show within-level connections. Edge sources are differentiated by color (EU=pink, National=blue, Local=orange) and edge span is indicated by line type (within level=dotted, cross level=solid arrows). (A) Authority network showing formal hierarchical relationships and regulatory oversight patterns with 150 directed ties. (B) Money network illustrating financial flows, including funding and resource transfers, with 836 directed ties. (C) Influence network depicting informal power relationships and policy influence pathways with 1,012 directed ties. The hierarchical layout reveals the multi-level governance structure in health policy networks, with varying connectivity patterns across the three relationship types.\u003c/p\u003e","description":"","filename":"Fig2.png","url":"https://assets-eu.researchsquare.com/files/rs-7837431/v1/ef98b2f306f71dc86f89fdb4.png"},{"id":93460541,"identity":"f667e807-2cdb-44bf-8fe7-259e7f64b930","added_by":"auto","created_at":"2025-10-14 06:10:26","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":20210564,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7837431/v1/4e367824-4d19-44fe-8303-56eaec5bf799.pdf"},{"id":93459209,"identity":"6e54b1c5-637c-4150-998c-f7ebf1af1607","added_by":"auto","created_at":"2025-10-14 05:54:17","extension":"html","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":128238,"visible":true,"origin":"","legend":"\u003cp\u003eAdditional information 1\u003c/p\u003e","description":"","filename":"SupplementaryMaterial.html","url":"https://assets-eu.researchsquare.com/files/rs-7837431/v1/dcd1c5989d5ed2f8d9d1365e.html"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eFrom stakeholder mapping to statistical modeling: An illustrative demonstration of end-to-end Net-Map methodology for health governance analysis\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Background","content":"\u003cp\u003eNet-Map was developed as a participatory way to visualize stakeholder networks and power relations [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Particularly, it combines stakeholder engagement with social network analysis to generate structured, analyzable data. Applications span food systems, child nutrition, reproductive health, obesity prevention, and natural resource governance [\u003cspan additionalcitationids=\"CR4 CR5 CR6\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. These studies show their value for mapping influence, authority, and resource flows in complex settings. Recent work argues for using Net-Map to integrate relational evidence into policy analysis [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eDespite its promise, several gaps remain in the literature. First, most applications have been in the \u003cem\u003eGlobal South\u003c/em\u003e (e.g., Ghana, Nigeria, Ecuador, Nepal), while European health governance has received little attention [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan additionalcitationids=\"CR11\" citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Second, few studies have applied Net-Map across the full research cycle: while the participatory phase is well developed, systematic integration with advanced statistical analysis remains rare, limiting its contribution to policy and systems research.\u003c/p\u003e\u003cp\u003eCancer provides a compelling test case for addressing these gaps. Globally, there were 19.3\u0026nbsp;million new cancer cases in 2020, with projections of 28.4\u0026nbsp;million annually by 2040 [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Europe, home to less than 10% of the world's population, carries 25% of this burden [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. In the European Union (EU), incidence and mortality continue to rise [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], with particularly high rates in Eastern Europe. In Romania, cancer is the second leading cause of death, responsible for 19% of all deaths in 2022 [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eBetween 30% and 50% of cancer cases are considered preventable [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. The European Commission [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e] and the World Health Organization [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] highlight primary prevention strategies, including reducing tobacco and alcohol use, promoting healthier diets and physical activity, vaccination, and mitigating environmental risks. Implementing these strategies requires coherent governance frameworks that extend beyond biomedical interventions to address structural determinants of health. Yet, responsibilities for prevention are distributed across local, national, and supranational actors, including ministries, regulatory agencies, professional associations, civil society, patient groups, and EU institutions. This complexity often produces fragmented strategies, weak coordination, and inequitable allocation of resources. Consequently, this makes cancer prevention governance a vital framework for understanding how decentralized health systems operate.\u003c/p\u003e\u003cp\u003eAnalyzing such governance processes requires approaches that can identify key stakeholders, map relationships of influence, collaboration, authority, and financial flows, and make visible both formal and informal power dynamics. Net-Map [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e] is particularly well suited to this task. While it has been used to study governance in other domains, it remains underutilized in European health systems and has not previously been applied to cancer prevention. This underuse is notable, as European health systems often privilege hierarchical decision-making and technical expertise, limiting opportunities for participatory analysis of governance [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eOur study addresses these gaps by applying Net-Map to primary cancer prevention governance in Romania, with a focus on stakeholders involved in implementing the European Code Against Cancer [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Building on recent research into social determinants and prevention [\u003cspan additionalcitationids=\"CR25\" citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e], we combine participatory mapping with advanced social network analysis (exponential random graph models) to examine how influence and authority ties are associated with directed funding across governance levels (EU, national, local). We contribute to health policy and systems research in three ways: (1) we offer one of the few European applications of participatory network mapping in public health, (2) we extend Net-Map across the full research cycle (from stakeholder elicitation and consensus validation to reproducible, model-based analysis), and (3) we show how multi-level structures and informal influence relate to the distribution of prevention resources, generating transferable, system-diagnostic insights for strengthening collaborative governance in decentralized health systems. Together, these contributions position Net-Map not only as a methodological innovation but also as a diagnostic tool for informing prevention policy and guiding more equitable allocation decisions in complex systems.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eExercise design and facilitation\u003c/h2\u003e\u003cp\u003eNet-Map is a participatory interview-based mapping method that integrates \u003cem\u003esocial network analysis\u003c/em\u003e with \u003cem\u003einfluence mapping\u003c/em\u003e to visualize stakeholder relationships and perceived power dynamics [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. In this study, we applied Net-Map to analyze the multi-layered governance of primary cancer prevention in Romania, with a particular focus on stakeholders involved in implementing the \u003cem\u003eEuropean Code Against Cancer\u003c/em\u003e [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThe choice of Net-Map was motivated by two considerations. First, Romania's health system context: while cancer is the country's second leading cause of death, the absence of a national cancer registry has led to fragmented data on incidence, outcomes, and survival. In this environment, prevention is crucial, yet it requires coordinated action across a fragmented governance landscape. Lessons from the \u003cem\u003eKarelia Project\u003c/em\u003e on cardiovascular prevention argued that prevention strategies are more effective when supported by collaborative networks [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Second, the methodological rationale: given the diversity and number of actors involved, Net-Map provides a systematic means of eliciting, visualizing, and analyzing governance networks. The methodological choices prioritized demonstrating the complete Net-Map workflow over maximizing representativeness. This analysis serves as an illustrative case study to show researchers how to implement the full research cycle.\u003c/p\u003e\u003cp\u003eThe Net-Map exercise was conducted through three in-person sessions (January\u0026ndash;February 2024) in Bucharest, Romania. Each session lasted approximately five hours. Splitting the exercise into multiple meetings minimized participant fatigue and allowed for reflection between stages. Namely, eliciting stakeholders and detecting their relationships are cognitively demanding tasks for participants [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eSession 1 (January 20 2024) consisted of identifying relevant actors and mapping within-level relationships (i.e., relationships among organizations of the same level). Session 2 (January 26, 2024) focused on the mapping of cross-level relationships (EU, national, and local). Session 3 (February 2, 2024) addressed the validation of the network maps. All three sessions were facilitated by a lead sociologist with extensive experience in qualitative sociology, supported by two assistants. Consistent with Net-Map principles, facilitators ensured clarity in problem framing, guided discussions, and moderated dynamics to avoid dominance by individual participants. All sessions were video-recorded to capture participant reasoning and support coding.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eParticipants and setting\u003c/h3\u003e\n\u003cp\u003e We purposively recruited five experts from the same non-profit organization (NGO), a national organization with longstanding, multi-level engagement in primary cancer prevention (local implementation partnerships, national policy dialogue, and interfaces with European initiatives). This embedded, practice-based panel was selected for three reasons. First, the NGO maintains continuous operational links across levels and sectors, providing participants with comprehensive visibility into stakeholders and inter-organizational ties relevant to the European Code Against Cancer. Second, shared technical language and institutional memory facilitated the efficient and high-fidelity elicitation of complex networks within the time and cognitive constraints of Net-Map. Third, panel members covered complementary functional roles (policy design/regulation, financing/commissioning, implementation/coordination, data analysis and reporting), ensuring breadth of perspectives despite a single organizational home. Eligibility required\u0026thinsp;\u0026ge;\u0026thinsp;10 years of experience in public health/health policy, direct involvement in primary prevention and/or European Code Against Cancer (ECAC) related activities, and routine collaboration with actors at local, national, and European levels. We did not aim for statistical representativeness; the goal was to provide credible coverage of system-critical perspectives, thereby generating and validating the stakeholder map.\u003c/p\u003e\u003cp\u003eWe recognized the risk of shared-organization bias (common frames or blind spots). To address this, we employed a scripted facilitation protocol with turn-taking and probe questions, conducted three sessions to facilitate reflection and correction, required explicit consensus on actors and ties, and performed structured digitization and cross-checks of the map. We also triangulated the elicited actor list against public policy documents, e.g., national cancer plans, ECAC guidance, and documented disagreements and their resolution.\u003c/p\u003e\u003cp\u003eWe note that Net-Map can be implemented with individual key informants or small groups. We chose a small-group format to enable triangulation and consensus-building while managing cognitive load. Larger panels should be intentionally avoided: they have an increased risk of proving logistically and cognitively unwieldy (crowded canvases, slower facilitation, harder consensus, higher coding error), which jeopardizes data quality.\u003c/p\u003e\u003cp\u003eUsing a single-NGO panel may over-represent organizational perspectives. We address this by transparently reporting recruitment, providing open materials/code, and emphasizing that findings are diagnostic and meant to inform further multi-stakeholder validation rather than to claim population representativeness. Concerning the expert panel, there were no conflicts of interest. Additionally, the experts did not receive any honoraria for their study participation.\u003c/p\u003e\u003cp\u003eIn total, twelve individuals were present at the sessions: five expert participants, one lead facilitator, two assistant facilitators, and four student observers responsible for logistics. To minimize potential bias, all experts were consulted equally during the mapping. Problem framing followed a standardized script derived from Schiffer [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e] to ensure consistency. We conducted the sessions in Romanian (including the mapping). The dataset is translated into English. Three guiding questions structured the mapping: (1) \u003cem\u003eWho are the actors responsible for implementing the European Code Against Cancer in Romania?\u003c/em\u003e (2) \u003cem\u003eWhat relationships connect them\u003c/em\u003e (i.e., financial flows, influence, authority)? (3) \u003cem\u003eWhat are their primary motivations or roles in relation to cancer prevention?\u003c/em\u003e Stakeholders were classified into four categories: \u003cem\u003egovernment\u003c/em\u003e (\u003cem\u003epublic\u003c/em\u003e), \u003cem\u003eprivate\u003c/em\u003e, \u003cem\u003eacademia\u003c/em\u003e, and \u003cem\u003ecivil society\u003c/em\u003e [\u003cspan additionalcitationids=\"CR30\" citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Actors were also categorized by governance level: \u003cem\u003elocal\u003c/em\u003e (Argeș county), \u003cem\u003enational\u003c/em\u003e (Romania), and \u003cem\u003eEuropean\u003c/em\u003e (European Union).\u003c/p\u003e\u003cp\u003eThis study is designed as a methodological illustration. The choice of a single-NGO panel, while providing operational advantages for demonstrating the Net-Map process, necessarily limits the generalizability of findings. The results should be understood primarily as showing how the methodology works rather than providing comprehensive insights into the Romanian healthcare system\u003c/p\u003e\n\u003ch3\u003eEthics\u003c/h3\u003e\n\u003cp\u003e All research procedures complied with the Declaration of Helsinki and the European General Data Protection Regulation (GDPR). The study protocol was approved by the Ethics Committee of the Center for Innovation in Medicine (EC-INOMED Decision No. D001/09-06-2023 and No. D001/19-01-2024). Written informed consent was obtained from all participants. We anonymized personal identifying information after each interview and securely stored data on encrypted drives accessible only to authorized personnel.\u003c/p\u003e\n\u003ch3\u003eMaterials and mapping procedure\u003c/h3\u003e\n\u003cp\u003eMapping was carried out using large sheets of paper, post-it notes (for stakeholder names), and colored markers (for relational ties). Stakeholders were written on post-its, attached to the canvas, and linked according to three types of relationships. Namely, \u003cem\u003eauthority\u003c/em\u003e, i.e., formal power or decision-making capacity (policy enforcement, regulation, command); \u003cem\u003einfluence\u003c/em\u003e, i.e., ability to shape others' decisions or actions through expertise, advocacy, or strategic positioning; \u003cem\u003efinancial resources\u003c/em\u003e, i.e., transfer of funding or economic support.\u003c/p\u003e\u003cp\u003eIn other Net-Map studies, the number of nominated stakeholders varied from 38 [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e] to 83 [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. In our case, we did not limit the experts in nominating stakeholders. Without a cap in the stakeholder elicitation process, the final map included 128 actors across governance levels. Parts of the original canvas are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\n\u003ch3\u003eData structuring and preparation\u003c/h3\u003e\n\u003cp\u003eThe hand-drawn maps (the stakeholders and their relationships) were digitized into two datasets. An \u003cem\u003eedge list\u003c/em\u003e file included source and target actors with relational type (financial flow, influence, authority), binary coded (1\u0026thinsp;=\u0026thinsp;tie present, 0\u0026thinsp;=\u0026thinsp;absent). An \u003cem\u003eattribute\u003c/em\u003e file consisted of actor ID, name, organizational category (coded as: public, private, non-profit, and academia), and governance level (local, national, and European).\u003c/p\u003e\u003cp\u003eDigitization was conducted in two different working sessions. Student assistants entered data into Excel, followed by cross-checking against the original maps. Experts validated the digital dataset in the third meeting. The dataset was then imported into RStudio (2023.06.1), where we made use of various packages for the analysis (e.g., \u003cem\u003eigraph\u003c/em\u003e, \u003cem\u003estatnet\u003c/em\u003e, \u003cem\u003eergm\u003c/em\u003e).\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eReliability and validation\u003c/h2\u003e\u003cp\u003eWe used a two-step reliability procedure. First, a \u003cem\u003econsensus mapping.\u003c/em\u003e During elicitation, participants jointly discussed stakeholder inclusion and ties. Ties were recorded only when all participants reached explicit consensus. Disagreements were resolved through targeted prompts and, where applicable, by consulting official documents (e.g., legal mandates, program guidelines). Items without consensus were flagged and revisited in the validation session. Only ties with full agreement were retained in the final dataset.\u003c/p\u003e\u003cp\u003eSecond, \u003cem\u003edouble data entry and adjudication\u003c/em\u003e. We implemented double data entry and adjudication: hand-drawn maps were independently digitized by pairs of observing student assistants, and discrepancies were reconciled by returning to the original canvases, session notes, and video; when ambiguity persisted, we consulted the expert panel. We did not compute a formal inter-coder agreement statistic during digitization because the task was a transcription of consensus-based, binary ties using pre-specified definitions. We acknowledge this as a potential limitation and recommend that future studies complement consensus and double-entry checks with blind recoding and κ/α estimates on a random subset. Based on our experience, limiting the number of stakeholders and tie types per session helps contain cognitive load and transcription errors.\u003c/p\u003e\u003cp\u003eWhile Net-Map is stakeholder-driven, facilitator influence is inherent [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. To reduce bias, we: (i) used a standardized script and predefined guiding questions, (ii) moderated dominant participants while encouraging quieter voices, (iii) validated all actors and ties across three sessions, and (iv) required participant consensus before finalizing the maps. Observers (students) remained passive and uninvolved in discussions.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003eStatistical analysis\u003c/h2\u003e\u003cp\u003ePrior to modeling, we calculated standard network descriptive statistics, including \u003cem\u003edensity\u003c/em\u003e (ratio of observed and theoretically possible ties) and cross-level tie patterns for each relationship type (authority, influence, financial flows). We examined the overlap between different relationship types to understand how formal and informal mechanisms relate to resource allocation patterns. We specifically analyzed cross-level coordination patterns by calculating the proportion of ties occurring within governance levels versus across levels (EU-National, EU-Local, National-Local) for each relationship type. Given the directed nature of funding relationships, we examined hierarchical resource distribution patterns by analyzing funding flows in both directions between governance levels (upward vs. downward funding).\u003c/p\u003e\u003cp\u003eWe used Exponential Random Graph Models (ERGMs) [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e] to examine which relational mechanisms are associated with the presence of a directed funding tie between organizations (\u003cem\u003ei\u003c/em\u003e funds \u003cem\u003ej\u003c/em\u003e). ERGMs treat the observed network as one realization from a reference distribution in which the log-odds of a tie depend on a set of network statistics capturing baseline tie propensity (the tendency of tie formation within the observed network), homophily (the tendency of same-level organizations to share a tie), and dyadic covariates (influence and authority ties). This approach accounts for the interdependence of ties, an inherent feature of social and governance networks, and yields coefficients interpretable as log-odds (reported alongside odds ratios). The dependent variable was the binary directed funding network derived from the Net-Map sessions. All nodes appearing on the map were retained; multiple mentions were reconciled during digitization and validation. Self-ties were not permitted.\u003c/p\u003e\u003cp\u003eWe estimated four nested ERGM specifications to assess the relative importance of different predictors: (1) baseline model containing only network density (\u003cem\u003eedges\u003c/em\u003e), (2) adding governance level homophily effects (\u003cem\u003esame governance level\u003c/em\u003e), (3) incorporating formal authority relationships, and (4) full model including informal influence relationships alongside all previous predictors. Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e briefly introduces the model terms.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eERGM parameter specifications for funding network analysis\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTerm\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eLabel (parameter)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eInterpretation\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBaseline density\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEdges (baseline)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eThis captures the overall tendency to form funding ties\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSame-governance level\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNodematch(level)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eThis tests whether actors at the same level (EU, national, local) are more or less likely to exchange funding. Given our focus on cross-level coordination, a negative coefficient indicates a preference for cross-level funding.\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDyadic covariates (edgecov)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eThis incorporates other mapped relations as predictors of funding. These covariates test whether influence and authority co-occur with funding beyond baseline density and level structure.\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eInfluence ties\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eedgecov(influence)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1 if \u003cem\u003ei\u003c/em\u003e reports influencing \u003cem\u003ej\u003c/em\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAuthority ties\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eedgecov(authority)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1 if \u003cem\u003ei\u003c/em\u003e has formal decision power over \u003cem\u003ej\u003c/em\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"3\"\u003eNote: Definitions and interpretations of exponential random graph model terms used to predict funding relationships in the Romanian cancer prevention governance network. The baseline density parameter captures overall tie formation tendency, while governance level homophily tests preferences for within-level versus cross-level funding coordination. Dyadic covariates examine whether existing authority and influence relationships predict funding ties beyond baseline network structure.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eModels were estimated using maximum pseudolikelihood (MPLE) as implemented in the \u003cem\u003eergm\u003c/em\u003e package. For interpretability, we present log-odds coefficients with standard errors (SE) and odds ratios (OR). We compared the full specification against nested models that omit each dyadic predictor in turn to assess incremental contribution using AIC/BIC comparisons. We evaluated (i) numerical convergence and parameter stability (including seed sensitivity) and (ii) structural Goodness of Fit (GOF) via simulated vs. observed distributions of degree, shared partners, and geodesic distances. In large, sparse governance networks, global structural GOF often remains imperfect even when dyadic effects are stable; accordingly, we interpret ERGMs as identifying dyadic association patterns rather than attempting to reproduce all higher-order structure. The full R code, as well as the data files, is available in the Additional file 1 [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eThe following results illustrate what researchers might expect to find when applying Net-Map methodology, rather than definitive conclusions about Romanian cancer prevention governance. The patterns described demonstrate the analytical capabilities of the combined Net-Map\u0026ndash;ERGM approach. Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e reports the descriptive statistics of the networks generated by the panel of experts in the Net-Map exercise. Specifically, the Romanian cancer prevention governance network comprises 128 organizations distributed across three administrative levels: 63 national-level organizations (49.2%), 40 local-level organizations (31.3%), and 25 European-level organizations (19.5%).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eNetwork structure and funding flow patterns in Romanian cancer prevention governance.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCharacteristic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eValue\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eCharacteristic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eValue\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNetwork Composition\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eCross-Level Funding Patterns\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTotal organizations\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e128\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eWithin-level ties\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e239 (28.6%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEuropean level\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e25 (19.5%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eCross-level ties\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e597 (71.4%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNational level\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e63 (49.2%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eCross-level percentage\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e71.4%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLocal level\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e40 (31.2%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003eFunding Flows by Direction\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eNetwork Structure\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eEuropean \u0026rarr; National\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e280 (17.8% density)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTotal funding ties\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e836\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eEuropean \u0026rarr; Local\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e161 (16.1% density)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePossible dyads\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e16,256\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNational \u0026rarr; Local\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e155 (6.2% density)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNetwork density\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.0514 (5.14%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNational \u0026rarr; European\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1 (0.1% density)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eReciprocity\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.01 (1.0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eLocal \u0026rarr; National\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0 (0.0% density)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eLocal \u0026rarr; European\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0 (0.0% density)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eRelationship Networks\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAuthority ties\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e150\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003eRelationship Overlap with Funding\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAuthority density\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.0092 (0.92%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eAuthority-funding overlap\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e43/150 (28.7%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eInfluence ties\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1,012\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eInfluence-funding overlap\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e647/1,012 (63.9%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eInfluence density\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.0623 (6.23%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"5\"\u003eNote: Illustrative network structure and funding flow patterns in Romanian cancer prevention governance. These descriptive statistics demonstrate the type of insights that can be generated through Net-Map analysis. Here, we report descriptive statistics for the multi-level cancer prevention governance network (N\u0026thinsp;=\u0026thinsp;128 organizations across European, national, and local levels). Cross-level funding comprises 71.4% of all funding relationships, with European organizations showing the highest outward funding activity, connecting to 17.8% of national organizations and 16.1% of local organizations. Upward funding is virtually absent, confirming hierarchical resource distribution. Influence relationships show greater overlap with funding ties (63.9%) than authority relationships (28.7%), suggesting informal networks better predict resource flows than formal structures. Overall network connectivity\u0026thinsp;=\u0026thinsp;5.14%; reciprocity\u0026thinsp;=\u0026thinsp;1.0%.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eThe funding network exhibits low overall density (5.14%), with 836 observed funding relationships among 16,256 possible directed dyads. Reciprocal funding relationships are rare (1% reciprocity), indicating predominantly unidirectional resource flows. Two distinct relationship networks were mapped alongside the funding network. The authority network contains 150 formal hierarchical relationships (density\u0026thinsp;=\u0026thinsp;0.92%), while the influence network is substantially denser with 1,012 informal influence ties (density\u0026thinsp;=\u0026thinsp;6.23%). The influence network density exceeds the funding network density, suggesting that influence relationships extend beyond direct resource allocation patterns.\u003c/p\u003e\u003cp\u003eThe funding network shows strong hierarchical coordination patterns, with 597 ties (71.4%) occurring across governance levels rather than within them. European-level organizations demonstrate the highest outward funding activity, directing resources to both national (280 ties from 1,575 possible connections, 17.8%) and local levels (161 ties from 1,000 possible connections, 16.1%). National-to-local funding is substantially lower (155 ties from 2,520 possible connections, 6.2%), while upward funding from local to national or European levels is virtually absent (1 tie total). This pattern claims a clear hierarchical resource distribution model where higher administrative levels fund lower ones.\u003c/p\u003e\u003cp\u003eThe overlap between relationship types and funding patterns reveals important differences in how formal and informal mechanisms translate into resource allocation. Influence relationships show substantial overlap with funding ties (647/1,012, 63.9%), indicating that informal influence networks closely mirror actual resource flows. In contrast, authority relationships show limited overlap with funding (43/150, 28.7%), suggesting that formal hierarchical structures are less predictive of actual resource allocation than informal influence patterns.\u003c/p\u003e\u003cp\u003eThese descriptive statistics illustrate that while formal authority structures exist, the governance network operates primarily through cross-level resource flows guided more by influence relationships than formal organizational hierarchies. The plot panel in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e illustrates the distribution of ties (money, influence, and authority) among the stakeholders identified during the Net-Map exercise.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTo illustrate how exponential random graph models can be used to understand mechanisms driving funding relationships, we estimated a series of ERGM models. These results exemplify the analytical insights possible through this methodological approach. Thus, to understand the mechanisms driving funding relationships in Romanian cancer prevention governance, we estimated a series of exponential random graph models (ERGMs) that progressively incorporated different theoretical predictors of tie formation. We started with a baseline model (Model 1) containing only the network density parameter (edges). Further, we sequentially added governance-level homophily (Model 2), formal authority relationships (Model 3), and informal influence relationships (Model 4) to assess their relative importance in explaining funding ties among organizations. Specifically, Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e presents the results of the four nested ERGM specifications. Model 1 establishes the baseline probability of funding ties in the absence of any structural predictors. Model 2 introduces governance-level effects to test whether organizations preferentially fund within or across administrative levels (local, national, and European). Model 3 adds formal authority relationships to examine whether hierarchical organizational structures predict resource flows. Model 4 presents the full specification, incorporating informal influence networks alongside formal authority and governance level effects.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eAuthority vs. Influence: Predictors of funding ties in health governance networks.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eParameter\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eModel 1\u003c/p\u003e\u003cp\u003eBaseline\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eModel 2\u003c/p\u003e\u003cp\u003eCross-level\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eModel 3\u003c/p\u003e\u003cp\u003eAuthority\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eModel 4\u003c/p\u003e\u003cp\u003eFull Model\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEdges (baseline)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-2.915***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-2.777***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-2.831***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-4.140***\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e(SE)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e(0.036)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(0.042)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e(0.043)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e(0.077)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSame governance level\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-0.417***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.387***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-1.040***\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e(SE)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(0.078)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e(0.079)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e(0.112)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e[OR]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.66\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.68\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.35\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAuthority relationship\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.989***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1.751***\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e(SE)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e(0.185)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e(0.305)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e[OR]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e7.31\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e5.76\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eInfluence relationship\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e5.127***\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e(SE)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e(0.106)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e[OR]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e168.54\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLog-likelihood\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-3295.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-3280.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-3238.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-1615.6\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAIC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e6592.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6564.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e6482.4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e3239.3\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBIC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e6599.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6579.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e6505.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e3270.1\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDeviance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e6590.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6560.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e6476.4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e3231.3\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003en (dyads)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e16,256\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e16,256\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e16,256\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e16,256\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003en (edges)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e836\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e836\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e836\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e836\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003en (nodes)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e128\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e128\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e128\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e128\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"5\"\u003eNote: Exponential random graph models predicting funding relationships among 128 organizations in Romanian cancer prevention governance. Coefficients are shown with standard errors (in parentheses) and odds ratios [in brackets]. Influence relationships are the strongest predictor of funding ties (OR\u0026thinsp;=\u0026thinsp;168.54), vastly exceeding authority relationships (OR\u0026thinsp;=\u0026thinsp;5.76). Organizations prefer cross-level funding coordination (OR\u0026thinsp;=\u0026thinsp;0.35). All models were estimated using MPLE. N\u0026thinsp;=\u0026thinsp;16,256 dyads; 836 funding ties. ***p\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eThe full model reveals several key patterns in network formation processes. Most strikingly, informal influence relationships emerge as the dominant predictor of funding ties, with organizations connected by influence relationships being 168.54 times more likely to have funding connections (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). This effect vastly exceeds that of formal authority relationships, which increase funding odds by 5.76 times (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Put differently, influence relationships are 29.3 times more predictive than authority relationships (168.54\u0026thinsp;\u0026divide;\u0026thinsp;5.76\u0026thinsp;=\u0026thinsp;29.3) in the flow of money. The governance level parameter indicates a strong preference for cross-level coordination, with same-level funding relationships being significantly less likely (OR\u0026thinsp;=\u0026thinsp;0.35, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e\u003cp\u003eThese findings suggest that while formal organizational hierarchies matter for resource allocation, informal influence networks are far more predictive of actual funding flows. The preference for cross-level funding coordination indicates that cancer prevention governance operates through hierarchical mechanisms rather than horizontal collaboration within administrative levels.\u003c/p\u003e\u003cp\u003eModel diagnostics and quality assessment are available in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. All ERGM models converged successfully using maximum pseudolikelihood estimation (MPLE). Parameter stability tests across different random seeds (12345, 54321, 99999) showed coefficient variation below 0.1 for all parameters, indicating highly stable estimates. Standard errors were successfully computed for all coefficients, enabling reliable statistical inference. Model comparison revealed substantial improvement from the baseline to full specification. The full model (AIC\u0026thinsp;=\u0026thinsp;3,239.3) dramatically outperformed the baseline model (AIC\u0026thinsp;=\u0026thinsp;6,592.1), representing an improvement of 3,352.8 AIC points. This consistent improvement indicates that relationship-based predictors capture fundamental network formation processes not explained by baseline density alone.\u003c/p\u003e\u003cp\u003eSpecification sensitivity tests confirmed the necessity of both authority and influence predictors. Removing the influence network from the full model increased AIC by 3,243.2 points, indicating that this predictor is critical for model performance. Removing authority relationships increased AIC by 27.7 points, showing this predictor also contributes meaningfully to model fit, though to a lesser degree than influence relationships.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eERGM Model Diagnostics and Quality Assessment.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDiagnostic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eValue\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eInterpretation\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eDiagnostic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eValue\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eInterpretation\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003ctr\u003e\u003cth align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e\u003cp\u003eModel Convergence\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e\u003cp\u003eModel Specification Tests\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEstimation method\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMPLE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eReliable estimation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eFull vs. No Authority (ΔAIC)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e27.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eAuthority contributes\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMCMC convergence\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSuccessful\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNo convergence issues\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eFull vs. No Influence (ΔAIC)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e3,243.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eInfluence essential\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eParameter stability (CV)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.1 (Stable)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eHighly stable results\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eInfluence effect necessity\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eCritical\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eCannot omit influence\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eAuthority effect necessity\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eSignificant\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eAuthority adds value\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e\u003cp\u003e\u003cb\u003eModel Fit Comparison\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e\u003cp\u003e\u003cb\u003eNetwork Coverage\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBaseline model AIC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e6,592.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePoor baseline fit\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eObserved ties modeled\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e836/836 (100%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eComplete coverage\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFull model AIC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e3,239.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSubstantial improvement\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eDyads analyzed\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e16,256\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eFull network analyzed\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAIC improvement (ΔAIC)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e3,352.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eVery large improvement\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eMissing data\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eNone\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eComplete data\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBIC improvement (ΔBIC)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e3,329.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eStrong model preference\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e\u003cp\u003e\u003cb\u003eGoodness-of-Fit Assessment\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e\u003cp\u003e\u003cb\u003eEstimation Quality\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTotal GOF tests\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e635\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eComprehensive testing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eStandard errors available\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eRobust estimation\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSignificant deviations\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e468\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eMany significant tests\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eConfidence intervals computed\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eFull inference possible\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePercentage significant\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e73.7%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePoor overall fit\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eP-values\u0026thinsp;\u0026lt;\u0026thinsp;0.001 (all params)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eHighly significant effects\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eOverall fit assessment\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePoor structural fit\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eDyadic patterns reliable\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e\u003cp\u003e\u003cb\u003eParameter Precision\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAuthority coefficient SE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.305\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eModerate precision\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eInfluence coefficient SE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.106\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eHigh precision\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCross-level coefficient SE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.112\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eHigh precision\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAll z-values\u0026thinsp;\u0026gt;\u0026thinsp;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eStrong statistical power\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"7\"\u003eNote: Diagnostic assessment of ERGM estimation quality and model specification. All models converged successfully with stable parameters (CV\u0026thinsp;\u0026lt;\u0026thinsp;0.1). The full model outperforms baseline specifications (ΔAIC\u0026thinsp;=\u0026thinsp;3,352.8), with influence relationships essential (removing increases AIC by 3,243.2) and authority relationships significant (ΔAIC\u0026thinsp;=\u0026thinsp;27.7). Goodness-of-fit testing shows poor structural fit (73.7% of 635 tests are significant), common in sparse networks, but dyadic relationship patterns remain reliable. All parameters demonstrate high significance (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and adequate precision\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eGoodness-of-fit testing using degree distributions, shared partner statistics, and geodesic distance revealed poor overall structural fit, with 468 of 635 tests (73.7%) showing significant deviations from the observed network. This pattern is common in large, sparse networks where ERGMs struggle to capture all structural dependencies. However, this limitation does not invalidate the dyadic relationship patterns that are the primary focus of our analysis.\u003c/p\u003e\u003cp\u003eThe poor structural fit indicates that while our model successfully identifies which types of relationships predict funding ties, it may not fully capture higher-order network properties like clustering patterns or degree distributions. This is consistent with ERGM limitations in modeling complex network structures with simple dyadic predictors.\u003c/p\u003e\u003cp\u003eAll relationship parameters demonstrated high statistical significance (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) with adequate precision. The influence relationship parameter showed particularly high precision (SE\u0026thinsp;=\u0026thinsp;0.106), while the authority relationship parameter had moderate precision (SE\u0026thinsp;=\u0026thinsp;0.305). The cross-level preference parameter also showed high precision (SE\u0026thinsp;=\u0026thinsp;0.112), indicating reliable estimation of governance level effects.\u003c/p\u003e\u003cp\u003e The ERGM analysis reliably captures dyadic relationship patterns and their relative importance for funding tie formation. However, the poor goodness-of-fit suggests caution in interpreting results about global network properties. The models are best understood as identifying which dyadic characteristics predict funding relationships rather than explaining the full complexity of network structure. The cross-sectional design limits causal inference about network formation processes. The models identify associations between relationship types and funding patterns, but cannot determine whether influence relationships lead to funding or whether funding creates influence relationships.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eWe demonstrate the methodological application of Net-Map to map governance of primary cancer prevention in Romania and illustrate how funding relationships can be modeled using ERGMs. This methodological demonstration reveals three types of patterns researchers might expect when applying this approach. First, funding may be predominantly cross-level: in this illustrative case, 71.4% of ties span EU, national, and local levels. Second, influence ties may emerge as the strongest correlate of funding. In this demonstration, influence relationships are 29.3 times more predictive than authority relationships of the flow of money. Third, authority ties may matter but potentially less so (OR\u0026thinsp;=\u0026thinsp;5.76). Same-level funding may be disfavored (OR\u0026thinsp;=\u0026thinsp;0.35) while the funding network remains sparse (density 5.14%). These illustrative findings suggest that resources flow downward and follow informal influence more than formal hierarchy. As an illustrative case, this analysis shows how influence appears to organize coordination and access, acting as a practical mechanism for moving money across levels. The methodological demonstration indicates that authority contributes, but potentially far less. The cross-level pattern indicates reliance on vertical linkages rather than horizontal collaboration within levels.\u003c/p\u003e\u003cp\u003eA further pattern exemplifies insights relevant for system design: EU-Local flows (16.1% of possible dyads) are present but smaller than EU-national flows (17.8%). In this illustrative setting where influence strongly predicts funding, the relative paucity of direct EU\u0026ndash;local ties and the lack of upward connections suggest weak feedback from implementers to EU programs. This example provides grounds to strengthen direct EU\u0026ndash;local connectivity, not to bypass national roles, but to shorten paths, improve information flow, and align funding with delivery realities.\u003c/p\u003e\u003cp\u003eThis illustrative analysis suggests three types of implications that might emerge from our Net-Map study. First, the methodology reveals that formal mandates alone may not predict allocation. These findings suggest that policy teams should identify influential brokers early in program design. Doing so can shorten implementation chains and reduce coordination losses. Second, the analytical approach indicates the importance of strengthening vertical interfaces, with explicit EU\u0026ndash;local links. Most funding is cross-level. Current flows favor EU-national more than EU-local. These examples demonstrate how establishing clearer EU\u0026ndash;local contact points, co-design sessions, and direct technical channels can improve alignment while keeping national stewardship. Third, the methodology shows that authority networks do not map cleanly onto funding. The analytical approach suggests that budget processes should incorporate structured stakeholder checks before finalizing allocations, especially when new prevention actions are introduced.\u003c/p\u003e\u003cp\u003eThe primary contribution of this work is methodological. We demonstrate an end-to-end workflow linking participatory mapping to model-based inference. Net-Map generated a validated list of actors and ties. ERGMs then quantified which relations are associated with funding while accounting for interdependence. This illustration shows that the approach is transparent and reproducible and appears appropriate for decentralized systems where formal documents under-represent how decisions are made. Additionally, this methodological demonstration describes patterns of allocation rather than distributional fairness. We therefore refrain from equity claims. Future work should add explicit equity metrics (e.g., funding in-degree by level or sector; concentration indices) and relate them to prevention needs to test whether cross-level flows are also fair flows.\u003c/p\u003e\u003cp\u003eAs an illustrative methodology demonstration, several important limitations must be emphasized. Most critically, the network reflects the perceptions of a single, highly embedded NGO panel. This methodological choice prioritized demonstrating the complete research workflow over achieving representative findings. While this brings strong system knowledge for illustrative purposes, it creates significant risks of shared blind spots that limit generalizability. We mitigated this through scripted facilitation, consensus rules, triangulation with documents, and multi-session validation, yet residual bias is possible. The primary purpose of this analysis is to show how the methodology works rather than to provide definitive insights into Romanian health governance. The design is cross-sectional; therefore, causality cannot be inferred. Influence could enable funding, but funding could also create influence. ERGM structural goodness-of-fit is poor, which is common in large, sparse graphs. As stated, we interpret models as identifying dyadic associations, not reproducing higher-order structure. These choices keep the model interpretable but may omit additional dependencies.\u003c/p\u003e\u003cp\u003eSeveral steps would extend this methodological approach. First, broaden participation to include government, clinical, and local implementers, followed by cross-panel comparison to achieve more representative findings. Second, link administrative data (budgets, grants, contracts) to the mapped ties to validate reported funding flows. Third, track change over time using repeated mapping or longitudinal models to examine how reforms, budgets, or EU initiatives alter governance patterns. These extensions would transform the illustrative demonstration into a comprehensive empirical analysis suitable for policy conclusions.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis methodological demonstration illustrates that primary cancer prevention governance can be analyzed through cross-level relationships in which informal influence appears to align with funding flows. The illustrative findings suggest that formal authority helps but may not determine allocation. In this example, the lower EU-local density, coupled with the central role of influence, demonstrates the type of insights that might provide grounds to strengthen direct EU-local connectivity. This Net-Map–ERGM workflow demonstrates a practical diagnostic approach to reveal where coordination occurs and where to target improvements within decentralized health systems. Future applications of this methodology with broader, more representative samples would be needed to generate definitive policy insights about specific health governance systems.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe dataset analyzed in the current study and the R code are made openly available, as an additional file, in the Zenodo data repository as Mihăilă, B.-E., et al. Replication data and code for stakeholder mapping and Net-Map health governance analysis. \u0026nbsp;Zenodo https://doi.org/10.5281/zenodo.17329451 (2025).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eB.E.-M., M.-G.H., M.G., C.C., B.-A.V., A.T.-L., and I.O. were supported by the 4P-CAN project, HORIZON-MISS-2022-CANCER-01, project ID 101104432, programme HORIZON. Views and opinions expressed are those of the authors only and do not necessarily reflect those of the European Union or the granting authority. Neither the European Union nor the granting authority can be held responsible for them. J.L. was supported by Deutsche Forschungsgemeinschaft (DFG 555455503).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank Dr. Bianca Cucoş, Florin Găină and Simona-Elena Puncioiu for their valuable contributions to this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eSchiffer E. Manual: Net-Map toolbox: influence mapping of social networks. Presented at the Sunbelt Conference of the International Network for Social Network Analysis. 2007. https://netmap.wordpress.com/wp-content/uploads/2008/06/net-map-manual-long1.pdf. Accessed 16 May 2024.\u003c/li\u003e\n \u003cli\u003eSchiffer E, Waale D. Tracing power and influence in networks: Net-map as a tool for research and strategic network planning. 2008. Washington, DC. International Food Policy Research Institute.\u003c/li\u003e\n \u003cli\u003eAryeetey R, Harding K, Hromi-Fiedler A, P\u0026eacute;rez-Escamilla R. Analysis of stakeholder networks for breastfeeding policies and programs in Ghana. Int Breastfeed J. 2020;15(1):74. doi:10.1186/s13006-020-00311-x.\u003c/li\u003e\n \u003cli\u003eMahmood H, Suleman Y, Hazir T, Akram DS, Uddin S, Dibley MJ, et al. Overview of the infant and young child feeding policy environment in Pakistan: Federal, Sindh and Punjab context. BMC Public Health. 2017;17:474. doi:10.1186/s12889-017-4341-5.\u003c/li\u003e\n \u003cli\u003eMachado JG, Buccini G, Recine E. An analysis of key actor networks for scale-up strategies for childhood obesity prevention and the care of children with obesity in Brazil. Curr Dev Nutr. 2023;7(7):101961. doi:10.1016/j.cdnut.2023.101961.\u003c/li\u003e\n \u003cli\u003eLitwan K, Lara-Mej\u0026iacute;a V, Chahine T, Hern\u0026aacute;ndez-Cordero S, Vilar-Compte M, P\u0026eacute;rez-Escamilla R. An analysis of actors participating in the design and implementation of workplace breastfeeding interventions in Mexico using the NetMap analysis approach. Front Public Health. 2023;11:1192600. doi:10.3389/fpubh.2023.1192600.\u003c/li\u003e\n \u003cli\u003eHauck J. Managing social-ecological systems for resilience: fisheries in the small reservoirs of northern Ghana [dissertation]. Rheinische Friedrich-Wilhelms-Universit\u0026auml;t Bonn; 2010. https://nbn-resolving.org/urn:nbn:de:hbz:5N-23570. Accessed 26 Jun 2024.\u003c/li\u003e\n \u003cli\u003eBoyle E, \u0026Oacute; Gallach\u0026oacute;ir B, Mullally G. Participatory network mapping of an emergent social network for a regional transition to a low-carbon and just society on the Dingle Peninsula. Local Environ. 2021;27(12):1431-45. doi:10.1080/13549839.2021.1936472.\u003c/li\u003e\n \u003cli\u003eCort\u0026eacute;s-Calder\u0026oacute;n SV, L\u0026oacute;pez-Rodr\u0026iacute;guez MD, Jim\u0026eacute;nez-Aceituno A, Castro AJ, Mancilla-Garc\u0026iacute;a M. Contributions of Net-Map to sustainability action research. Curr Opin Environ Sustain. 2025;75:101542. doi:10.1016/j.cosust.2025.101542.\u003c/li\u003e\n \u003cli\u003eSchiffer E, Mustapha AY, Mustaph AL. Planning, budgeting and disbursing funds for newborn survival in Katsina State, Nigeria \u0026ndash; a Net-Map analysis. 2012.\u0026nbsp;https://netmap.files.wordpress.com/2007/11/schiffer_12_net-map_case_study_newborn_survival_nigeria.pdf.\u0026nbsp;Accessed 24 Jul 2024.\u003c/li\u003e\n \u003cli\u003eRosales Ap, Linnemann AR, Luning PA. A Net-Map analysis to understand the roles and influence of stakeholders in street food safety\u0026mdash;a study in Ecuador. Food Control. 2023;154:109966. doi:10.1016/j.foodcont.2023.109966.\u003c/li\u003e\n \u003cli\u003eKarn S, Devkota MD, Uddin S, Thow AM. Policy content and stakeholder network analysis for infant and young child feeding in Nepal. BMC Public Health.\u0026nbsp;2017;17:421.\u0026nbsp;doi:10.1186/s12889-017-4340-6.\u003c/li\u003e\n \u003cli\u003eSung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209-49. doi:10.3322/caac.21660.\u003c/li\u003e\n \u003cli\u003eEuropean Parliament. Fighting cancer in the EU: statistics and action (infographics). 2022.\u0026nbsp;https://www.europarl.europa.eu/topics/en/article/20200131STO71517/fighting-cancer-in-the-eu-statistics-and-action-infographics.\u0026nbsp;Accessed 29 Jul 2024.\u003c/li\u003e\n \u003cli\u003eEU Science Hub. Cancer cases and deaths on the rise in the EU. 2023.\u0026nbsp;https://joint-research-centre.ec.europa.eu/jrc-news-and-updates/cancer-cases-and-deaths-rise-eu-2023-10-02_en. Accessed 29 Jul 2024.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eOECD. Romania: country cancer profile. 2023. https://www.oecd.org/content/dam/oecd/en/publications/reports/2023/02/eu-country-cancer-profile-romania-2023_b7601b86/267467c6-en.pdf.\u0026nbsp;\u003c/strong\u003eAccessed 29 Jul 2024.\u003c/li\u003e\n \u003cli\u003eGlobal Cancer Observatory. Romania: fact sheet. 2022.\u0026nbsp;https://gco.iarc.who.int/media/globocan/factsheets/populations/642-romania-fact-sheet.pdf. Accessed 29 Jul 2024.\u003c/li\u003e\n \u003cli\u003eWorld Health Organization. Preventing cancer. 2025.\u0026nbsp;https://www.who.int/activities/preventing-cancer. Accessed 27 Sep 2024.\u003c/li\u003e\n \u003cli\u003eEuropean Commission. Cancer prevention. 2024.\u0026nbsp;https://knowledge4policy.ec.europa.eu/cancer/cancer-prevention-2024_en.\u0026nbsp;Accessed 27 Sep 2024.\u003c/li\u003e\n \u003cli\u003eEspina C, Ritchie D, Feliu A, Canelo-Aybar C, D\u0026rsquo;Souza E, Mitrou PN, et al. Developing evidence-based cancer prevention recommendations: methodology of the World Code Against Cancer Framework to create region-specific codes. Int J Cancer. 2025;1-10. doi:10.1002/ijc.70068.\u003c/li\u003e\n \u003cli\u003eHauck J, Schiffer E. Between intuition and indicators \u0026ndash; using Net-Map for visual and qualitative social network analysis. In: Gamper M, Reschke L, Sch\u0026ouml;nhuth M, editors. Knoten und Kanten 2.0: soziale Netzwerkanalyse in der Medienforschung und der Kulturanthropologie. Bielefeld: Transcript; 2012. p. 231-257.\u003c/li\u003e\n \u003cli\u003eYasnoff WA, O\u0026rsquo;Carroll PW, Friede A. Public health informatics and the health information infrastructure. In: Shortliffe EH, Cimino JJ, editors. Biomedical informatics: computer applications in health care and biomedicine. New York: Springer; 2006. p. 537-63. doi:10.1007/0-387-36278-9_15.\u003c/li\u003e\n \u003cli\u003eSch\u0026uuml;z J, Espina C, Villain P, Herrero R, Leon ME, Minozzi S, et al. European Code against Cancer 4th edition: 12 ways to reduce your cancer risk. Cancer Epidemiol. 2015;39 Suppl 1:S1-10. doi:10.1016/j.canep.2015.05.009.\u003c/li\u003e\n \u003cli\u003eOană I, H\u0026acirc;ncean MG, Perc M, Lerner J, Mihăilă BE, Geantă M, et al. Online media use and COVID-19 vaccination in real-world personal networks: quantitative study. J Med Internet Res. 2024 Oct 25;26:e58257. doi:10.2196/58257.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eMihăilă BE, H\u0026acirc;ncean MG, Perc M, Lerner J, Oană I, Geantă M, et al. Cross-sectional personal network analysis of adult smoking in rural areas. R Soc Open Sci. 2024;11:241459. doi:10.1098/rsos.241459.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eH\u0026acirc;ncean MG, Lerner J, Perc M, Molina JL, Geantă M, Oană I, et al. Processed food intake assortativity in the personal networks of older adults. Sci Rep. 2025;15:10459. doi:10.1038/s41598-025-94969-0.\u003c/li\u003e\n \u003cli\u003ePuska P, Vartiainen E, Nissinen A, Laatikainen T, Jousilahti P. Background, principles, implementation, and general experiences of the North Karelia Project. Glob Heart. 2016;11:173-8. doi:10.1016/j.gheart.2016.04.010.\u003c/li\u003e\n \u003cli\u003ePuska P, Jaini P. The North Karelia Project: prevention of cardiovascular disease in Finland through population-based lifestyle interventions. Am J Lifestyle Med. 2020;14:495-9. doi:10.1177/1559827620910981.\u003c/li\u003e\n \u003cli\u003eMcCarty C, Killworth PD, Rennell J. Impact of methods for reducing respondent burden on personal network structural measures. Soc Networks. 2007;29:300-15. doi:10.1016/j.socnet.2006.12.005.\u003c/li\u003e\n \u003cli\u003eStadel M, Stulp G. Balancing bias and burden in personal network studies. Soc Networks. 2022;70:16-24. doi:10.1016/j.socnet.2021.10.007.\u003c/li\u003e\n \u003cli\u003eAminullah A, Wusko AU. Pentahelix model to create shared value in empowering the people of Winong Village, Gempol District, Pasuruan Regency. Global Review of Tourism and Social Sciences. 2025;1:59-67. doi:10.53893/grtss.v1i2.348.\u003c/li\u003e\n \u003cli\u003eCassaniti J, Kumoji K, Rariewa F, Ohkubo S, Oyenubi O. Influence networks relating to health knowledge among Nairobi\u0026rsquo;s micro-retailers and their clients. Electron J Knowl Manag. 2021;18:302-24. doi:10.34190/ejkm.18.3.2068.\u003c/li\u003e\n \u003cli\u003eChoi BC, Pak AW. A catalog of biases in questionnaires. Prev Chronic Dis. 2005;2(1):A13.\u003c/li\u003e\n \u003cli\u003eLusher D, Koskinen J, Robins G, editors. Exponential random graph models for social networks: theory, methods, and applications. Cambridge: Cambridge University Press; 2013. doi:10.1017/cbo9780511894701.\u003c/li\u003e\n \u003cli\u003eMihăilă B-E, H\u0026acirc;ncean M-G, Geantă M, Cioroboiu C, Lerner J, Molina JL, et al. Replication data and code for stakeholder mapping and Net-Map health governance analysis. Zenodo; 2025. doi: 10.5281/zenodo.17329451 \u0026nbsp;\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[{"identity":"305df727-daf6-4540-920e-2a46adc8ff03","identifier":"10.13039/501100000780","name":"European Commission","awardNumber":"HORIZON-MISS-2022-CANCER-01, project ID 101104432, programme HORIZON","order_by":0},{"identity":"5babcd06-8701-4b8c-8bcc-e33f17c48af6","identifier":"10.13039/501100001659","name":"Deutsche Forschungsgemeinschaft","awardNumber":"555455503","order_by":1}],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"European Union","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Net-Map methodology, stakeholder mapping, primary cancer prevention, public health policy, social network analysis","lastPublishedDoi":"10.21203/rs.3.rs-7837431/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7837431/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e\u003cp\u003eCoordinating primary cancer prevention in Europe requires alignment across the EU, national, and local actors. Yet governance relations are challenging to observe. We used a participatory network approach to make relationships visible and test which ties are associated with funding. This study serves as a methodological demonstration to illustrate the complete Net-Map research cycle from data collection to statistical analysis.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eThree Net-Map sessions (January\u0026ndash;February 2024) with a national NGO expert panel elicited stakeholders and ties for authority, influence, and funding, followed by validation. Canvases were used to record the information, and they were digitized afterwards into edge and attribute files. We described network structure and estimated exponential random graph models (ERGMs) for the directed funding network, including overall tie propensity, governance level (EU, national, local), and dyadic influence and authority relations.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eThe final network comprises 128 organizations and 836 funding ties (density 5.14%). Funding is predominantly cross-level (71.4%). EU actors show the strongest outward activity (EU-national density 17.8%; EU-local 16.1%), while national-local is lower (6.2%). In ERGMs, influence is the strongest correlate of funding (OR 168.54, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), authority is positive but smaller (OR 5.76, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and same-level funding is less likely (OR 0.35, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Model comparisons favored the full specification; structural goodness-of-fit was poor, typical for large, sparse networks, so we interpret effects as dyadic associations.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e\u003cp\u003eThis illustrative analysis claims that participatory Net-Map workshops, combined with formal network analysis are helpful in diagnosing coordination patterns in decentralized health systems. The findings draw from a single embedded panel and cross-sectional design that can be extended in future interventions. Still, they are primarily serving as an example of methodological application rather than providing generalizable insights about Romanian cancer prevention governance.\u003c/p\u003e","manuscriptTitle":"From stakeholder mapping to statistical modeling: An illustrative demonstration of end-to-end Net-Map methodology for health governance analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-14 05:54:13","doi":"10.21203/rs.3.rs-7837431/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d03d3a44-839a-455f-ac91-80b9c258139b","owner":[],"postedDate":"October 14th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":56143629,"name":"Health Policy"},{"id":56143630,"name":"Preventive Medicine"}],"tags":[],"updatedAt":"2025-10-14T05:54:13+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-14 05:54:13","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7837431","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7837431","identity":"rs-7837431","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}