Integrating Multi-Source Qualitative Data Using Causal Loop Diagrams: A Methodological Framework for Developing Systems Thinking

Abstract

Health systems, as complex adaptive systems, are influenced by the intertwined interactions among various actors and policies interactions that often defy explanation through linear analytical approaches. This article presents the methodological framework that integrates multi-source qualitative data through the use of system dynamics causal loop diagrams, thereby enabling the researcher to progressively develop and refine a systems-thinking perspective. Qualitative data were obtained from four sources: a literature review, semi-structured interviews with managers and experts, analysis of official documents from Social Security Organization of Iran, and an expert panel. Each method made a distinct contribution to shaping the dimensions of systems thinking: The literature review contributed to identifying key variables and shaping the initial model; document analysis supported understanding the organizational structure and further contextualizing the model; interviews enabled the discovery of organizational behavior as well as nonlinear and delayed relationships; and the expert panel facilitated the uncovering of hidden drivers underlying system behavior.

The proposed “Integrating Qualitative Data for Systems Thinking” (IQDST) framework demonstrated that the convergence of these diverse data sources enabled a transition in the researcher’s analytical perspective, from descriptive analysis to structural and institutional analysis, while also identifying complex feedback loops. This article highlights that, structured integration of diverse qualitative data sources not only enriches conceptual models and enhances the ability to anticipate indirect and long-term consequences of policies but also serves as an effective mechanism for strengthening researchers’ systems thinking capacity during model development processes.

Introduction

Health systems in the contemporary world are increasingly recognized as complex adaptive systems. These are structures in which multiple actors -from policymakers to service providers and patients-interact through nonlinear dynamics in dynamic contexts characterized by uncertainty (Atkinson et al., 2015). This complexity is particularly evident in dimensions e.g. financing, resource generation, service delivery and governance, where conflicting interests, feedback loops, and policy interdependencies emerge, factors that cannot be explained through linear thinking (Nugraha, 2010).

Linear analyses in health policy and healthcare institutional management are often insufficient and even misleading, as they consider only direct cause effect relationships while failing to capture feedback dynamics and unforeseen consequences (Nyström et al., 2023; Squires et al., 2016).

To avoid such oversimplification, it is essential to design dynamic systems models that can represent real-world interactions. The development of systems thinking in the researcher’s mind is a prerequisite for constructing such models, as it enables the depiction of complex interrelationships among system components (Faezipour & Ferreira, 2011). At the policy level, this perspective transforms the decision-making process into a network of interlinked relationships with both micro- and macro-level consequences, rather than a collection of individual factors.

The first step in developing such models is the collection of adequate and appropriate data for building a dynamic conceptual framework. Gathering qualitative data from diverse sources can provide a robust foundation for the development of these conceptual models. Literature reviews, by extracting theoretical concepts and identifying existing causal relationships, help define key variables (Onwuegbuzie et al., 2011). Semi-structured interviews reveal stakeholders’ mental models, exposing nonlinear and lagged relationships as well as perceptual differences (Kallio et al., 2016). Analysis of official documents identifies critical and policy contradictions at the structural level (Tobin et al., 2022). Expert panels, by fostering consensus among specialists, not only validate feedback loops but also refine and strengthen the model’s causal structure (Cameron & Hynes, 2024).

Existing analytical frameworks categorize systems thinking processes into four dimensions thinking, decision-making, action, and interpretation and identify three levels of proficiency: skilled, competent, and novice (Jaiswal & Karabiyik, 2022). Systems thinking represents a crucial skill for researchers in simulation systems, and its efficacy maximized when researchers can develop appropriate qualitative models to represent system dynamics. Designing such models necessitates researchers, particularly novices, to move beyond linear, reductionist perspectives to grasp feedback loops, cumulative dynamics, and non-linear system behaviors (Stave et al., 2024).

CLDs serve as an effective tool for identifying critical feedback processes, competing goals within the system, and the indirect effects of policies (Uleman et al., 2024). The development of these diagrams integrating diverse qualitative data sources not only clarifies the system’s structure and dynamics but also functions as a continuous learning tool that organizes the researcher’s thinking throughout the study and facilitates coherent, in-depth data analysis.

The purpose of this study is to present and explain “Integrating Qualitative Data for Systems Thinking” (IQDST) methodological framework a framework developed during the “Designing an Organizational System Dynamics Model for Social Security” project through a structured, incremental, and iterative synthesis of qualitative data and simultaneously, facilitated the researcher’s transition from linear thinking to systems thinking. This framework demonstrates how step-by-step organization of qualitative data can activate and strengthen systems thinking patterns in the researcher’s mind, empowering them to design dynamic models adaptable to the complexities of the problem. Accordingly, IQDST is introduced as a methodological tool that can assist novice researchers in developing a systemic perspective and moving beyond linear approaches.

Theoretical Framework

Systems Thinking

Systems thinking is a cognitive paradigm that analyzes phenomena as sets of interconnected components, identifying multidimensional causal relationships, feedback patterns, and dynamic processes of change (Adam & de Savigny, 2012; Burato et al., 2023). Unlike linear analyses, this approach requires a holistic, dynamic, and multi-level perspective to understand the behavior of complex systems (Haraldsson, 2000).

Systems thinking originates from scientific movements of the late twentieth century, evolving from General Systems Theory and Cybernetics towards modeling dynamic system behaviors. This progression established the foundation for analyzing feedback structures and nonlinear dynamics (Forrester, 2007) and subsequently, it expanded into domains such as management, policy development, and education, as complex issues within these fields can only be effectively analyzed through understanding multifaceted causal relationships, accumulative dynamics, and feedback loops (Armenia, 2020).

Related studies have further solidified the role of systems thinking across domains including healthcare, environmental science, and governance. These fields are characterized by dynamic complexity, nonlinear shifts, and multi-layered analytical levels necessitating a holistic and temporal approach (Peters & systems, 2014). Concurrently, the development of modeling tools, such as system dynamics and agent-based simulation, has enabled researchers to more precisely represent both qualitative and quantitative aspects of systems structures (Macal, 2010). So, the literature demonstrates that the transition from linear to systems thinking is not merely a cognitive shift, but a fundamental prerequisite for developing robust qualitative models and analyzing the behavior of complex systems.

Systems thinking encompasses a set of fundamental principles that provide a conceptual foundation for understanding the behavior and dynamics of complex systems. These principles as below, guide us beyond reductionist perspectives, enabling a deeper recognition of interactions among components, feedback mechanisms, and internal system dynamics.

The first principle emphasizes holism and interdependence. It posits that every system, when viewed as a whole, exhibits emergent properties that transcend the sum of its parts. Therefore, understanding a phenomenon requires examining the entire system rather than focusing on its individual components. This holistic view facilitates recognition of internal interactions and mutual influences among system elements, providing an integrated picture of the whole (San Miguel, 2023).

The second principle pertains to feedback loops. It highlights that system components continuously influence one another over time. Feedback loops can be reinforcing, intensifying trends and fostering growth, or balancing, stabilizing the system and dampening fluctuations. Understanding these loops is essential for analyzing system behavior over time and for designing effective interventions (Paxton & Frost, 2018).

Another key principle is dynamic complexity. Complex systems rarely exhibit simple, direct cause–effect relationships. Interventions in such systems may generate unexpected or even contradictory outcomes, sometimes with considerable time delays. Recognizing this complexity enhances our ability to anticipate the potential consequences of new policies and actions (Adam & de Savigny, 2012).

The next principle is nonlinearity. In complex systems, small changes can lead to disproportionately large and sometimes unpredictable effects. This principle underscores that system behavior is not necessarily proportional or linear, necessitating analytical approaches different from traditional linear models (McAlister et al., 2022).

Finally, diverse perspectives and clearly defined analytical boundaries play a crucial role. Systems thinking emphasizes that a comprehensive understanding of a problem requires examining it from multiple viewpoints and continually questioning the boundaries of analysis. This approach fosters the discovery of new insights and supports more informed and adaptive decision-making (Khalil & Lakhani, 2022; Reynolds, 2024).

These principles have particular relevance and significant applicability in health systems. Health systems exemplify complex adaptive systems in which multilayered relationships and dynamic interactions among components and stakeholders coexist and evolve (Roxas et al., 2020).

Causal Loop Diagrams and Qualitative Data

CLDs are a key tool in systems thinking, graphically representing causal relationships and feedback loops (Cassidy et al., 2022; Schaffernicht, 2010). By integrating diverse perspectives, these diagrams reveal the structure of complex interactions and identify effective leverage points for interventions (Cavana & Mares, 2004; Ram & Irfan, 2021). Beyond analysis, CLDs serve communicative and educational functions, facilitating the translation of complex insights to stakeholders and enhancing their engagement in decision-making processes (Watz & Hallstedt, 2020). In this study, CLDs are employed as tools for the researcher’s progressive learning from qualitative data and for developing a systems thinking, an approach that has rarely been reported in an integrated manner.

Qualitative data comprise non-numerical information, including texts, speech, images, or documents, which focus on describing characteristics, contexts, and interactions (Adams, 2015; Imbert, 2010; Qu & Dumay, 2011). The primary aim of using this kind of data is to achieve a deep understanding of phenomena from the perspective of individuals or groups and to uncover the mental logic driving behaviors (Andruccioli et al., 2009; Mojtahed et al., 2014; Ruslin et al., 2022). Collected in naturalistic settings and emphasizing meaning and interpretation, qualitative data are particularly suitable for addressing complex issues and developing theoretical insights (Arksey & O’malley, 2005; Cacchione, 2016; Munn et al., 2018).

Qualitative data can be collected through various methods, including interviews, focus groups, observation, document analysis, expert panels, and literature reviews. Interviews can be conducted in three formats: structured interviews, which allow comparability and reduce bias (Dewi, 2021; Noble & Smith, 2014; Rambod, 2018); semi-structured interviews, which provide greater depth and facilitate the exploration of stakeholders’ mental models (Denny & Weckesser, 2022); and unstructured interviews, which offer high flexibility but are time-consuming and prone to bias (Heap & Waters, 2019). Focus groups generate rich data through group interactions, though their analysis can be challenging (Denny & Weckesser, 2022; Dewi, 2021). Observation captures behaviors in natural settings but entails ethical considerations and the risk of observer bias (Denny & Weckesser, 2022). Document analysis provides a cost-effective method for examining structured and historical data but requires careful evaluation of source credibility (Barrett & Twycross, 2018; Salminen, 2010). Expert panels offer diverse perspectives and enhance the credibility of findings, although reaching consensus can be difficult (Setia, 2017; Tong et al., 2007). Finally, literature reviews, by synthesizing multiple sources, consolidate existing knowledge and support more precise causal analyses (Adlini et al., 2022; Ferreira et al., 2018).

Study Context: “A System Dynamics Modeling Approach to Analyze the Impact of Policy Challenges on the Health of Insured Populations Covered by the Social Security Organization of Iran”

The study project, “A System Dynamics Modeling Approach to Analyze the Impact of Policy Challenges on the Health of Insured Populations Covered by the Social Security Organization of Iran”, aimed to develop a system dynamics model to analyze the impact of policy-related challenges on the health of insured individuals under the coverage of the Social Security Organization of Iran. Within this project, instabilities in policies related to healthcare service provision (whether production or purchasing of services) were systematically examined to identify the root causes affecting the health status of insured individuals. As a result, a comprehensive system dynamics model was developed, incorporating multiple feedback loops. Key leverage points for implementing effective policy interventions were identified, which have the potential to significantly improve patient health outcomes. This article presents the research methodology employed in the project, along with examples of identified causal loops to illustrate the researcher’s learning process and the development of systems thinking.

Methodology

This study was a mixed qualitative investigation conducted in four sequential stages:

Literature Review

A systematic review of scholarly sources was performed to design the preliminary model, aiming to answer the question: Which factors influence patient health? Using reputable databases and defined inclusion and exclusion criteria, theoretical concepts and initial causal relationships were extracted.

Document Analysis

The second stage entailed an analysis of organizational documents. In this phase, the researcher sought to identify the formal structure of the organization by addressing the question: What is the organization’s formal structure and its objectives? Relevant documents, including laws, regulations, guidelines, strategic plans, and organizational performance reports, were identified, screened for authenticity, validity, representativeness, and relevance, and subsequently analyzed.

Semi-Structured Interviews

Were conducted to uncover the underlying causes of unmet organizational objectives. The guiding question for this stage was: Given precise laws, policies, and guidelines, what are the SSO’s main challenges in achieving its goals? Interviews were carried out with current and former experts, managers, and policymakers, both within and outside the organization, selected through purposive sampling.

Expert Panel

Finally, an expert panel was convened, comprising specialists in health policy and health insurance. This stage aimed to review, refine, and validate the proposed causal loops and conceptual model. Panelists were selected based on their professional expertise and practical experience, and iterative feedback sessions were conducted to revise the model and reveal previously unrecognized relationships.

To analyze qualitative data and identify themes, the thematic analysis method (Braun & Clarke, 2006) was employed, alongside coding techniques developed by Kim and Andersen approach (Kim & Andersen, 2012) for constructing CLDs. This process involved not only the analysis of explicit and implicit themes within the text, but also the determination of causal linkages between these themes for the development of CLDs. Given the recommendation to avoid creating overly large and complex diagrams CLDs were developed iteratively and incrementally-accounting for the human cognitive limitation of short-term memory-thereby enhancing their comprehensibility, causal loops were sequentially constructed based on articulated causes and challenges, ultimately consolidated to develop the final dynamic conceptual model. The data were managed and analyzed using MAXQDA software, version 20.

Data Integration

In this study, data integration was conducted based on the IQDST framework (Figure 1). This framework comprises four sequential steps, beginning at the level of a static conceptual model and formal organizational structure, and progressing to the institutional level and more in-depth analyses. In the process of systems learning, distinguishing between the structural and institutional levels plays a key role in deepening data analysis and interpretation. The structural level refers to the formal and observable elements of the system, including resource allocation processes, operational guidelines, and financial mechanisms, as reflected in official documents and organizational plans (Lunenburg, 2017; Rapert & Wren, 1998). In contrast, the institutional level encompasses unwritten rules, conflicting interests, organizational culture, and power interactions among various actors (Wu et al., 2023). In the following, the data analysis and code integration steps are presented:

Figure 1.

The integrating qualitative data for systems thinking (IQDST)

In the first step, an extensive list of variables related to patient health was extracted based on a literature review. The propositions identified during this step were coded using MAXQDA, with codes including: organizational policies, management and leadership structures, hospital processes, staff performance, hospital financial resources, medical equipment, patient status, efficiency, quality, and access to services. An initial conceptual model in line with the main research objective was then developed. This static model comprised variables and causal relationships identified in the literature, depicting linear relationships among factors affecting patient health.

In step 2, involved analyzing official organizational documents to contextualize the model for the Social Security Organization (SSO). Through analyzing the organization’s structure, main goals, healthcare procurement and provision policies, and operational guidelines, codes were extracted. Some of these codes overlapped with those from step 1, while others were novel, such as financial support, financial sustainability, and legal mandates. New variables were incorporated into the initial model, aligning it more closely with the specific context of the Social Security Organization. In this step, the relationships between codes evolved, forming the basis for multiple feedback loops. For example, combining the codes “legal mandate” with “service quality”, “patient satisfaction,” and “patient health” formed Loop A (Figure 2), indicating that organizational legal requirements, across multiple dimensions, positively influence service quality, thereby enhancing patient health and satisfaction.

Figure 2.

Reactive decision-making feedback loop

In step 3,semi-structured interviews were conducted with experts to analyze their experiences and identify system behavior in response to internal and external challenges. This step involved mapping sequences of events and the system’s reactions to changes, resulting in the formation of causal loops between variables. These loops were iteratively refined with each interview, comprehensively covering various dimensions of the issue. Consequently, the model transitioned from a static state to a dynamic one incorporating feedback relationships.

Regarding the previously mentioned example (Loop A in Figure 2), new codes emerged during this step, including “service capacity” and “reactionism,” which represented contributing factors to the challenge of declining service quality. These codes were added to the existing variables within Loop A, and their relationships were determined based on interviewees’ perspectives. Illustrative quotes which influenced the direction and nature of these variable connections include:

- “The dissemination of news concerning SSO-owned healthcare centers and facilities efficiency in the media claiming higher service delivery costs compared to governmental hospitals affiliated by Ministry of Health despite insufficient evidence supporting this claim, heightened managerial and policymaker sensitivity regarding efficiency due to its social impact. This led to reactive interventions focused on cost reduction and increased efficiency.” (M2, PART A: Social expectations have a positive impact on policymakers’ reactionism.)

- “Service delivery development by SSO-owned healthcare centers has been driven both by adherence to health services regionalization protocols based on regional needs; and by political pressures for capacity expansion or social pressures exerted through the donation of a portion of the required capacity and subsequent demands for its completion and expansion-the latter development often lacked alignment with the mentioned protocols.” (M5, PART A: Social expectations led to physical capacity expansion through policymakers’ reactionism.)

- “Weak enforcement of certain regulations and inadequate management of political and social pressures related to service capacity development including facilities and high-technology medical equipment coupled with the failure to adequately leverage the organization’s potential as a member of the Supreme Health Insurance Council in defining and expanding health services packages and purchasing tariffs, are key factors impacting resources within the healthcare sector.” (M6, PART A: Increased developmental costs contribute to organizational financial instability and can negatively affect service quality.)

- “Now, the Board of Directors is encountering repeated and frequent proposals concerning facility development including expansion of existing health care centers, upgrading center levels, and the construction of new hospitals all while healthcare services regionalization protocols are not fully implemented. This presents a significant challenge in decision-making.” (M10: PART A: Social expectations have a positive impact on policymakers’ reactionism.)

In step 4, a panel of experts validated the analytical data and identified causal loops. This process addressed several researcher questions and fostered a deeper understanding of the system’s internal dynamics. Structural and institutional codes were examined collectively as evidence for “hidden causes of system behavior,” with codes such as “structural inconsistency” and “decision-making instability” identified as key factors explaining changes in system behavior. These codes represent leverage points within the system and formed the basis for designing proposed interventions to modify its behavior.

Researcher’s Learning Process

Based on the findings of the literature review, employee performance and the quality of healthcare service delivery have been reported as the most influential factors affecting patient health (Milosavljević et al., 2024). Moreover, the enhancement of service quality at the structural level of the organization is manifested through regulations, objectives, plans, and performance evaluation indicators. In interviews with organizational managers, the potential for improving service quality, resulting efforts to increase the efficiency of the healthcare centers affiliated to organization, was also emphasized. An examination of the underlying causes of declining service quality in these centers revealed that the system’s behavior, in the face of various situations, i.e. changes in government, subsequent shifts in senior leadership within the health system and the organization, and societal expectations, is complex and occasionally diverge from the organization’s structural rules. In this regard, several causal loops conceptually developed to explain the underlying mechanisms contributing to the decline in healthcare service quality. One such loop concerns policies aimed at expanding the physical resources for service provision, derived from the integration of organizational document analysis and interviews with policymakers.

In this loop (Figure 2), and in accordance with the Law on the Obligation of the Social Security Organization to Implement Clauses (a) and (b) of Article 3 of the Social Security Act, the organization is mandated to provide healthcare services through two primary mechanisms—direct provision by owned centers and purchasing of services from contracted centers. At the same time, ensuring equitable access, improving quality, enhancing the efficiency of owned healthcare centers, and providing financial protection for insured individuals are among the objectives of the organization’s healthcare sector (Islamic Consultative Assembly Research Center, n.d).

The growing awareness of the insured population regarding their legal rights has heightened public expectations and led to more assertive demands. Consequently, a range of political and social institutions, seeking to strengthen their positions and support public demands and expectations, exert pressure on the organization to expand healthcare services in terms of capacity and facilities across regions, the scope of covered services, and the level of financial protection.

Weaknesses in managing such pressures have led to reactive and hasty decision-making by organizational leaders. Over time, this has resulted in unsystematic expansion of healthcare facilities, thereby increasing organizational costs, creating financial sustainability challenges, and ultimately exerting a negative impact on the quality of healthcare services delivered. On the other hand, the decline in service quality has reduced patient health outcomes and satisfaction, which in turn intensifies media and institutional pressure on the organization once again.

Another feedback loop identified during the expert panel session reflected the organization’s disproportionate focus on access indicators rather than the quality of healthcare services. As previously discussed, the ultimate outcome of this trajectory was an increase in patient dissatisfaction. This loop, in turn, reconnected the organization to the cycle of institutional pressures, thereby reinforcing the previous loop. During the formation of this loop, the researcher observed a behavioral shift at the institutional level, whereby organizational responses, under the influence of media pressures and institutional pressures, sometimes deviated from established laws, regulations, and internal procedural guidelines.

In another loop (Figure 3), referred to the lack of stability in strategic policymaking at the macro level, the researcher identified, through the integration of multiple data sources, the system’s behavioral pattern in response to governmental changes, a dynamic that ultimately undermines patients’ health and satisfaction within the healthcare facilities affiliated to SSO. Findings from the literature review indicate that employee performance, and consequently the quality of healthcare services provided, are highly influenced by the performance monitoring and evaluation methods and their effectiveness across organizational levels (Danforth et al., 2023; Yasin et al., 2023). Despite the organization’s extensive body of policies, laws, and regulations, and its consistent emphasis on monitoring and accountability in its strategic and operational plans interviews with senior managers revealed that the quality of supervision, evaluation, and accountability mechanisms for both affiliated and contracted healthcare centers is perceived to be low. One of the key reasons identified for this structural and institutional instability was the instability of the organization’s strategic policies as the largest health insurance provider in the country, the organization is highly susceptible to country macro-level decisions and political shifts. Consequently, changes in government often result in the senior managers change within the organization. In the absence of a stable decision-making mechanism, such changes lead to fluctuations in the organization’s strategic direction, thereby weakening the consistency of managerial oversight and performance evaluation at lower administrative levels, including hospital and clinic management. Furthermore, the declining motivation among mid-level and operational managers to control costs, stemming from weaknesses in the organization’s traditional budgeting system, has gradually eroded managerial culture, increased expenditures, reduced service quality, and diminished staff productivity. Over time, these processes have collectively contributed to the deterioration of patient health outcomes and satisfaction.

Figure 3.

Strategic policy instability feedback loop at the macro level

The instability of the organization’s strategic direction gives rise to another loop (Figure 4), illustrating the failure to achieve all organizational objectives and the intended outcomes of plans. This challenge emerges when the implementation of previously adopted strategies and initiatives is halted or becomes ambiguous, despite the considerable investment of expert time and financial resources that went into their design and execution. Such discontinuities affect both the “service provision” and “service purchasing” policies, leading to inefficient expenditure within the organization. On one hand, delayed payments to contracted providers and their growing dissatisfaction undermine service delivery; on the other hand, resource constraints within the healthcare facilities affiliated to SSO, limit their operational capacity. Ultimately, these dynamics degrade the quality of healthcare services and negatively affect patient health and satisfaction. Within this loop, the researcher identified a clear inconsistency between the formal organizational documents and the system’s actual behavior under different circumstances. Although the organization’s strategic plans explicitly emphasize equitable access to healthcare, resource management, and quality improvement, senior managerial decisions sometimes disregard these goals in practice.

Figure 4.

Organizational goal failure feedback loop

By triangulating data from the literature review and interviews, it became evident that, beyond governmental transitions, several underlying factors contribute to this instability, including conflicts of interest, media pressures, institutional pressures, and the limited use of scientific evidence and research findings by policymakers (Arowoogun et al., 2024; Brems et al., 2021). Collectively, these elements reinforce the loop of strategic discontinuity and perpetuate inefficiency, ultimately undermining the sustainability and effectiveness of health policy implementation within the organization.

Through the expert panel sessions, the researcher was able to resolve several conceptual ambiguities that had emerged during the earlier stages of data analysis. These ambiguities stemmed from contradictions identified across multiple data sources, including the analysis of the organization’s financial documents, interviews with health economics managers and experts, and discussions with senior managers and policymakers.

Triangulated analysis revealed that the expansion of healthcare services through direct provision by establishing and developing the organization’s own medical facilities—entailed substantial financial expenditures. These costs were further intensified by inflation and fluctuations in the national exchange rate. In contrast, the findings of several feasibility studies suggested that purchasing some certain services from contracted providers would be more cost-effective and economically rational for the SSO. However, a key question remained: why did senior decision-makers continue to prioritize the expansion of direct service provision despite evidence indicating its inefficiency? The discussions within the expert panel provided the answer. According to the deliberations among health economists and mid-level managers in the health sector, the healthcare facilities owned by the organization—providing free services to insured patients—represent a substantial competitive advantage. These facilities have drawn considerable political and social attention to the organization.

Through the iterative review and refinement of causal loops during the panel sessions, experts not only validated the model’s internal logic but also identified critical leverage points within the system. This process enabled the researcher to develop a more holistic understanding of the system’s feedback structure and to propose balanced policy interventions targeting those leverage points, thereby enhancing the overall coherence and sustainability of the system dynamics model.

Discussion

Within the IQDST framework, four core steps were defined: initial conceptual model design, model localization, model dynamization, and model validation. This framework facilitates the integration of four data types through qualitative data collection methods: (1) theoretical concepts and linear causal relationships; (2) structural data and macro-policies; (3) institutional data derived from experiential narratives; and (4) institutional data informed by diverse expert perspectives. The integration of these data sources enabled the identification of key variables and relationships, discovery of non-linear and lagged connections, unveiling of structural and systemic contradictions, pinpointing of critical leverage points, and exploration of hidden drivers of system behavior. Consequently, a dynamic conceptual model was developed for analyzing the identified challenge, capable of representing the behavioral and structural complexities of the system.

This framework aligns with Thelen’s Systems Thinking for Health Actions (STHA) framework. STHA, through a systematic literature review and testing in Pakistan’s national COVID-19 response, identifies six key characteristics of systems thinking: recognizing interconnections and system structure, identifying feedback loops, identifying leverage points, understanding dynamic behavior, utilizing mental models, and creating simulation models for policy testing (Thelen et al., 2023). Furthermore, the article “Techniques to Enhance the Public Policy Impact of Qualitative System Dynamics Models” demonstrates that utilizing CLDs to explore multifaceted issues coupled with online group processes, leveraging quantitative data, and employing software-based visual analytics, can enhance the impact of qualitative system dynamics models on public policy (Veldhuis et al., 2024). This study, focused on the social consequences of the COVID-19 pandemic, indicated that such techniques strengthen all dimensions of systems thinking. In comparison, IQDST framework places greater emphasis on the researcher’s learning process and transition from linear to systemic perspectives through its focus on multi-source qualitative data integration and revealing institutional and structural relationships, which not only enhances analytical quality but also documents the researcher’s learning trajectory, thereby increasing policy impact potentials. This distinction positions IQDST as a complementary and developmental tool.

The methodological literature indicates that data integration or triangulation, is a cornerstone of qualitative and mixed-methods research. The article “Beyond Methods: Theoretical Underpinnings of Triangulation” demonstrates, through a systematic review of studies in the social sciences, that data integration is not merely a data collection technique but rather a theoretical approach for increasing analysis depth, validity, and reliability (Vivek et al., 2023). This article shows that simultaneous use of diverse sources such as in-depth interviews, focus groups, documents, texts, and even survey data can help represent the multifaceted nature of social reality and reduce the limitations of relying on a single data source. In this context, the IQDST framework, through its systematic integration of four data types theoretical, structural, institutional, and experiential represents an advanced application of triangulation. This framework not only leverages data source diversity to increase analytical validity and richness but also facilitates a deeper understanding of systemic dynamics by revealing hidden institutional and structural relationships. Thus, IQDST builds on the theoretical foundations of triangulation to enhance the quality of systems analyses and move beyond common limitations in single-source studies.

The educational literature also shows that novice researchers often experience confusion when confronted with the diversity of theories and qualitative frameworks. The article “Composing with Theory” emphasizes that strategies such as defining concepts precisely and practicing design and writing can streamline their learning process (Noble & Smith, 2014). Additionally, studies like “Understanding Qualitative Research Methodology” have shown that many students struggle to understand the theoretical and methodological foundations of qualitative research (Copple & Roulston, 2025). In this context, the IQDST framework can play an important supporting role for novice researchers. This framework facilitates the learning process both theoretically and analytically by documenting the researcher’s transition from linear to systemic perspectives and serving as a valuable complement to existing methodological guides in qualitative research.

Furthermore, the article “A Worked Example of Qualitative Descriptive Design” shows that the flexibility of qualitative descriptive design can lead to confusion unless study implementation steps are clearly defined (Villamin et al., 2025). In this regard, the IQDST framework provides a clear path for integrating multi-source data and can serve as a valuable complement to such practical guides for novice researchers in dynamic modeling. The utilization of CLDs in the development of a dynamic conceptual model not only assists researchers in identifying root causes of challenges but also plays a crucial role in fostering systems thinking. Specifically, through iterative loop construction, researchers gain insight into the consequences of challenges or interventions over time and appreciate the interconnectedness of system components.

The findings of this project revealed that many challenges in healthcare service delivery at contracted or owned facilities stem from hasty decision-making by managers and policymakers with a short-term focus. When these issues arise, more superficial and visible explanations such as staff demotivation, liquidity shortages, or equipment deficiencies are often cited; however, the root causes lie in past strategic decisions. This temporal distance between macro-decisions and their operational consequences alerted researchers to one characteristic of complex systems. As McCaskill notes, policymaking involves a significant spatial and temporal gap between cause and effect, making it difficult to accurately discern the relationship between policy inputs and outcomes in complex environments due to the influence of multiple feedback loops (McCaskill, 2014). Similarly, Brimble’s research indicates that root-cause analysis tools are useful for identifying challenges and potential solutions but cautions that unintended consequences and unforeseen factors must also be considered (Brimble & Jones, 2017) which CLDs fulfill this role effectively within the IQDST framework.

Conclusion

Systems thinking, as a problem-solving framework, significantly enhances the capacity for root-cause analysis by providing a holistic understanding of complex situations. This approach facilitates deeper comprehension and more effective decision-making when confronting complex issues through the identification of multiple causal factors and understanding the interactions among components within a dynamic system. In complex systems, change in one component can cascade to other components change, ultimately influencing itself recursively. The application of IQDST enabled the researcher to analyze, compare, and study diverse data types deeply, thereby gaining insight into the principles and mechanisms governing system behavior and tracing the ultimate consequences of policy decisions across multiple dimensions. Through the integration of varied collected data, the researcher’s perspective shifted from a linear to a systemic and comprehensive approach, allowing for analysis of component changes from diverse dimensions, identifying the root causes, and mapping the interconnected component influences through feedback loops. This process facilitated by IQDST revealed how the system behaves at the institutional level, discrepancies between this behavior and its structural levels, and the underlying reasons for these inconsistencies. The discovery and development of a dynamic conceptual model, presented to policymakers alongside technical explanations, fostered insights beyond personal experience, potentially enhancing the quality and effectiveness of policy making within health systems including the Social Security Organization.

Limitations

Despite the achievements demonstrated by this study, several notable limitations exist. First, the IQDST framework relies predominantly on qualitative data and multi-source integration; consequently, its findings are contingent upon the quality and diversity of collected data and may exhibit reduced validity in contexts characterized by limited or heterogeneous datasets. Second, this research was conducted within a specific institutional context, necessitating further investigation to ascertain the generalizability of the results to other healthcare systems or social policy domains. Third, the data integration process and conceptual model development were partially dependent on researcher judgment, which introduces the potential for interpretive biases in the analysis. Fourth, due to the qualitative nature of the data, comprehensive empirical testing of non-linear relationships and feedback loops was constrained, highlighting the need for complementary studies employing quantitative methodologies. While the IQDST framework successfully documented the researcher’s learning trajectory and enriched systemic analyses within this study, future research should explore its efficacy across diverse contexts and by independent researchers.