causal loop diagrams literature review

Diverse Approaches to Creating and Using Causal Loop Diagrams in Public Health Research: Recommendations From a Scoping Review

• Community Health and Epidemiology, University of Saskatchewan, Saskatoon, SK, Canada

Objectives: Complex systems thinking methods are increasingly called for and used as analytical lenses in public health research. The use of qualitative system mapping and in particular, causal loop diagrams (CLDs) is described as one promising method or tool. To our knowledge there are no published literature reviews that synthesize public health research regarding how CLDs are created and used.

Methods: We conducted a scoping review to address this gap in the public health literature. Inclusion criteria included: 1) focused on public health research, 2) peer reviewed journal article, 3) described and/or created a CLD, and 4) published in English from January 2018 to March 2021. Twenty-three articles were selected from the search strategy.

Results: CLDs were described as a new tool and were based upon primary and secondary data, researcher driven and group processes, and numerous data analysis methods and frameworks. Intended uses of CLDs ranged from illustrating complexity to informing policy and practice.

Conclusion: From our learnings we propose nine recommendations for building knowledge and skill in creating and using CLDs for future public health research.

Introduction

There is a trend in public health research for the application of complex systems thinking methods and tools [ 1 – 3 ]. We conceptualize public health research from this perspective in terms of examining systems that are complex webs of sectors, institutions, people, structures, and interventions that aspire to maintain and improve population health. Furthermore, we value public health research that is “based on the principles of social justice, attention to human rights and equity, evidence-informed policy and practice, and addressing the underlying determinants of health” [ 4 ].

There are published review articles regarding complex systems thinking methods used in public health research and together these paint a broad landscape [ 2 , 3 , 5 – 10 ]. In this literature, there is clear support for using qualitative system mapping and in particular, causal loop diagrams (CLDs) as analytical tools to embed complex systems thinking. The origins of the use of CLDs emanate from the system dynamics branch of systems science founded by Forrester [ 11 ] and CLDs are needed because “we live in a complex of nested feedback loops” [ 12 ]. One example of using a CLD in public health research is a study of factors that influenced health promotion policy and practice in a regional public health system [ 13 ]. Here, the CLD was useful because “feedback mechanisms can be seen as leverage points to strengthen systems” and to “identify potential opportunities to disrupt or slow down vicious feedback mechanisms or amplify those that are virtuous cycles.” At the time of this study (2018), there were few examples of CLDs in public health literature [ 14 – 21 ].

To our knowledge there are no published reviews that synthesize public health research in terms of how CLDs are created and used. We were motivated to conduct a literature review to determine how CLD methodology could be used to identify leverage points in local public health systems to strengthen the response to COVID-19 in Canada. The aim of this paper is to address this gap in the literature and synthesize knowledge from recent innovations for our research and contribute to knowledge development. We posed two research questions: 1) How are CLDs created and used in recent (>2018) public health research? 2) What recommendations emerge regarding how to create and use CLDs in public health research?

A scoping review was chosen for this study in order “to examine how research is conducted” and “to provide an overview or map of the evidence” [ 22 ]. A narrative synthesis approach was utilized as the topic required exploration more than explanation and human and time resources were limited [ 23 ]. Key issues identified by Byrne [ 24 ] to strengthen the review were addressed such as ensuring transparency in search strategy and data extraction, analysis and synthesis.

Search Strategy

Literature was searched using the Scopus and PubMed databases and used the following search terms: causal loop diagram*, complex*, system* thinking, method*, tool, approach, research, and public health. Inclusion criteria were 1) public health research, 2) peer reviewed journal article, 3) described or created a CLD as a research method, and 4) published in English from January 2018 to March 2021. The key objective was to find state-of-the-field examples of CLDs, therefore, extensive hand searches of references was completed. It is important to note that piloting this search strategy uncovered numerous articles that only mentioned CLDs and did not explicitly meet the criteria of “described or created a CLD as a research method.” While we set out to use PRISMA guidelines we deemed it unnecessary given the search strategy quickly became one of including all articles that meet our inclusion criteria.

Data Extraction and Analysis

Study selection was conducted by one author (LBL) while appraisal and duplicate independent data extraction and validation was conducted by two authors (LBL and CH). CN provided input throughout the study and facilitated discussion about any differences. Data extraction followed these six categories:

1) Research aim,

2) Description of complex systems thinking,

3) Why a CLD was selected as a method,

4) How the CLD was created,

5) How the CLD was used, and

6) Recommendations for future research using CLDs.

Two authors (LBL and CH) extracted verbatim text that aligned with the extraction categories and these were saved to a spreadsheet. Both authors reviewed the spreadsheet in its entirety, discussed individual articles to gain clarity, and wrote summary paragraphs to identify high level themes. Following this, for each article, summary statements were written for the six extraction categories and a table was created. The two authors reviewed each other’s summaries for accuracy and revisions were made. Finally, directed content analysis was used to interpret extracted data “through systematic classification of coding and identifying themes and patterns” [ 25 ].

We found 23 articles in total that met our inclusion criteria. A list of these articles and summary statements are provided in Table 1 . This section answers our first research question: How are CLDs created and used in recent (>2018) public health research? The organization of this section mirrors the six data extraction categories indicated above.

TABLE 1 . Summary statements of extracted data (Canada, 2021).

Research Aims

Although the literature addressed a range of public health topics, non-communicable disease prevention was most frequently addressed (15/23) and of those, seven were focused on obesity prevention. Table 2 provides a list of research topics.

TABLE 2 . Research topics of reviewed literature (Canada, 2021).

In terms of research aims found in the 23 articles, four themes emerged: 1) to examine the complexity of a public health topic and illustrate complex systems thinking [ 26 – 34 ]; 2) to discuss the complexity of a public health intervention [ 35 – 40 ]; 3) to describe study protocol and how CLDs were created [ 41 – 44 ]; and 4) to illustrate how CLDs can be used to monitor and track initiatives to improve population health or evaluate impact of interventions [ 45 – 48 ].

Complex Systems Thinking

Complex systems thinking was discussed in terms of systems, problems, interventions, and key concepts that drive this type of approach. Several articles indicated that the systems they were studying were complex, for example:

A complex system may be characterized by its heterogeneity (various actors and structures at different levels); its dynamic, interactive, and adaptive nature (its ability to respond to or resist external changes, or changes in the interacting parts); and its emergent properties (arising through interactions between processes or factors that alone do not exhibit such properties) [ 30 ].

Following on this, feedback loops in complex systems were explicitly discussed in all articles to some extent. Jalali et al. [ 38 ] described these in terms of “causal chains of multiple variables in which changes in each variable could be traced back to its historical values.” They go on to define the difference between reinforcing and balancing feedback loops.

Another way complex systems thinking was described was with respect to complex problems and interventions . Burrell et al. [ 36 ] discussed community violence in terms of embedded contexts and the lack of holistic understanding of such “dynamic complexity.” Complex problems and interventions were often discussed together. The need to move away from “isolated intervention thinking” to systemic interventions to study systems change was highlighted by Knai et al. [ 30 ].

All articles built upon the descriptions reported above in some manner when discussing complex systems thinking. Some articles described this as providing “the opportunity to understand, test, and revise our understanding of how the different components in a system work together” [ 31 ] and “to study complex problems as the manifestation of dynamic interactions among their constituent parts” [ 36 ]. Furthermore, a few articles expanded the discussion to include such concepts as boundary judgement [ 38 , 43 , 47 ], that is, “establishing boundaries to the system is a fundamental starting point to efforts to change systems” [ 47 ].

Why Causal Loop Diagrams?

CLDs were mostly seen as a means or a tool to examine feedback at play in public health issues. Some articles were explicit [ 28 , 32 , 33 , 40 , 43 , 44 ] while others implied this. Both Riley et al. [ 43 ] and Parmar et al. [ 40 ] labeled this as “causal loop analysis” and the resulting CLDs were a means to understand systems and potential “programming.” Using a CLDs was a new tool for some [ 42 , 46 ] and as one article related, “business as usual” was not working to address obesity [ 47 ]. CLDs were also considered a tool to help tell a story. For example, a CLD was thought to support the development of “a concise narrative about a particular problem” [ 42 ] and Brereton et al. [ 28 ] stated that “every causal loop tells a story that links cause and effect through feedback.”

How Were Causal Loop Diagrams Created?

There were many combinations of methods used to create CLDs. In this section we present this diversity in terms of 1) data sources, 2) processes, 3) data analysis, 4) frameworks, and 5) diagramming ( Table 3 ).

TABLE 3 . How causal loop diagrams were created (Canada, 2021).

Data Sources

Both primary and secondary data were used for creating CLDs ( Table 3 ). Most articles reported on primary data collection (18/23) and this included interviews [ 26 , 27 , 33 , 35 – 40 ], group model building with stakeholders and/or community members [ 32 , 41 , 43 , 44 , 46 , 48 ], behavioral data [ 42 , 47 ], fieldnotes [ 37 ], and workshops with experts [ 31 ]. Twelve articles used primary data only.

Secondary data was used in 10 articles [ 26 , 28 – 30 , 32 , 36 – 38 , 40 , 45 ] and this consisted of document and/or literature review ( Table 3 ). Of the eighteen articles that reported on primary data collection, six included document review [ 26 , 32 , 36 – 38 , 40 ]. Documents included policy briefings, reports, consultation papers, and evaluation reports [ 37 ], documentaries and ethnographies [ 36 ], program data [ 38 ], geographical information and government documents [ 32 ], and data from published databases [ 28 , 37 , 45 ]. Literature reviews were undertaken in four articles and these either supplemented primary data [ 26 ], secondary data [ 28 , 45 ], or both [ 36 ]. Document and literature review were utilized in four articles [ 28 – 30 , 45 ].

There were three processes used to create CLDs: group model building, researcher created only, and researcher created with stakeholder refinement ( Table 3 ). Group model building (GMB) was the most common process as reported in 11 articles [ 27 , 31 – 33 , 41 – 44 , 46 – 48 ]. Urwannachotima et al. [ 33 ] described GMB as “an established methodology for engaging stakeholders to gain mutual understanding of complex relationships and to collectively develop comprehensive systems models that represent the cause and effect relationships of a problem.” They go further to explain that “stakeholders are deeply and actively involved in the process of model construction through the exchange, assimilation, and integration of mental models into a holistic system description.” GMB was generally reported to be a process where participants brainstormed and named potential variables, drew connections and feedback loops between the identified variables, and then mapped these ideas onto a final CLD. However, there was a variety of GMB processes used and was often not clearly described in terms of session design and activities. Beyond GMB, Hassmiller Lich et al. [ 46 ] discussed group concept mapping and Gerritsen et al. [ 41 ] described graphing over time and cognitive mapping.

CLDs created by researchers only was the second most common process (10/23). Two articles reported that CLDs were presented to stakeholders for refinement [ 39 , 40 ]. The range of approaches included:

• Using coded interview data to map interactions between key variables [ 26 , 35 – 38 ],

• Conducting a literature review to compare causal links uncovered in interview data [ 26 ] or a document review [ 29 , 30 ],

• Completing both a literature review and a document review to identify variables [ 28 , 45 ],

• Building on an existing CLD [ 29 ], and

• Creating a CLD solely from researcher knowledge and expertise [ 34 ].

Data Analysis

Overall, we found that description was often lacking regarding qualitative data analysis methods used. However, some articles [ 35 , 37 , 39 ] that collected primary data discussed methods described by Kim and Anderson [ 49 ]. Others such as Owen et al. [ 39 ] created a table to demonstrate how they used coded interview transcript statements to inform their CLD. Steps in the analysis included 1) using coded text to show causal linkages, 2) translating these to cause-and-effect variables, and 3) creating word-and-arrow diagrams for CLD use. Similarly, Brereton and Jagals [ 28 ] presented a table to identify variables and describe influencing links.

Several articles applied specific frameworks to inform research. For example, Allender et al. [ 47 ] used Foster-Fishman’s [ 50 ] theoretical framework of six elements (i.e., systems norms, financial resources, human resources, social resources, regulations, and operations) to study root causes, system interactions, and levers for change. Similarly, Baugh Littlejohns and Wilson’s [ 5 ] framework of seven attributes of effective prevention systems (i.e., leadership, resources, health equity paradigm, information, implementation of desired actions, complex systems thinking, collaborative capacity) was used by Bensberg et al. [ 35 ] in their study design.

Diagramming

Many articles reported on the use of software for creating the actual diagram. Vensim [ 31 , 35 , 37 , 39 , 40 , 44 – 46 ], Stella Architect [ 28 ], and STICK-E [ 43 ] were the three diagrammatic programs used. Further to the actual diagram, there was a wide array of CLD types and degrees of diagram readability. We found that some CLDs were kept quite simple, with fewer variables, arrows, and loops, while others were very complicated. For example, Brereton et al. [ 28 ] created a tightly packed and dense color-coded main CLD and six diagrams of various feedback loops to highlight key variables, relationships, and potential leverage points. Overall, we found that key variables in blocks or shapes, labelled arrows and feedback loops, color coding, legends, and clear diagram interpretation descriptions were important aspects for readability.

Intended Uses of Causal Loop Diagrams

There were nine ways that CLDs were intended to be used and these are identified in Table 4 . The following provides examples of each intended use.

TABLE 4 . How causal loop diagrams were intended to be used (Canada, 2021).

Illustrate Complexity and Identify Leverage Points

Illustrating complexity was aligned with research aims in several articles ( Table 4 ) and was implicit in the other articles with respect to using CLDs. Identifying leverage points was explicitly discussed in twelve articles. Osman et al. [ 31 ] found that key variables and their interactions pointed to strategies to enhance leadership “through a reduction in bureaucracy in the health system.” Similarly, Bensberg et al. [ 35 ] identified leadership as a leverage point as well as knowledge and data, resources, workforce, and collaborative relationships that need to be “nudged in the desired direction.” One of the more detailed descriptions of leverage points was from Sahin et al. [ 32 ]. They adapted Meadows [ 51 ] framework of places to intervene in system to identify shallow or deep leverage points to address the “wicked complexity” of the COVID-19 pandemic.

Inform Policy and Practice

Informing policy was a reported intended use of CLDs in twelve articles ( Table 4 ). Some articles were detailed in offering policy directions while others simply stated that the CLD could inform policy. Clarke et al. [ 37 ] examined “key influences on policy processes, and to identify potential opportunities to increase the adoption of recommended policies” with respect to a state government obesity prevention initiative. Other examples include the need for policies to address population growth, family size, and family planning to improve child health [ 28 ], housing, energy and wellbeing [ 27 ], and sugar-sweetened beverage tax to reduce sugar consumption and dental caries [ 33 ].

Informing practice was also a frequently identified intended use of CLDs (13/23) ( Table 4 ). For example, Osman et al. [ 31 ] stated that their CLD could be used “to develop local action plans for implementation and consider strategies for mitigating possible future risks” and Parmar et al. [ 40 ] to develop “strategies to enhance capacities, services, and coordination to improve the health of refugees.”

For System Dynamics Modeling

Five articles created CLDs for use in system dynamics modeling [ 26 , 27 , 38 , 45 ] ( Table 3 ). This was defined by Araz et al as “a computer-aided approach to model and facilitate analysis of complex system behaviors over time” [ 45 ]. They further described the steps in system dynamic modeling, and this was very much in line with other articles:

We first constructed a causal loop diagram (CLD) informed by the existing literature to present the causal relationships between variables in drugged driving behaviors and traffic safety policies. A stock-flow diagram (SFD) was then used to convert these dynamic processes into quantitative expressions and a simulation tool [ 45 ].

Mirroring the above descriptions, Crielaard et al. [ 26 ] discussed the value of system dynamic modeling in terms of testing policy options from “studying ‘what if’ scenarios using computational modelling approaches.” It was notable that Urwannachotima et al. [ 33 ] and Swierad et al. [ 44 ] stated that the primary value of CLDs was in quantitative modelling.

Measure and Evaluate Systems Change

Table 4 identifies four articles that used CLDs to help measure and evaluate systems change [ 31 , 39 , 42 , 48 ]. For example, Owen et al. [ 39 ] reported that “the methods provide a technique to retrospectively evaluate community interventions from a systems perspective and understand the way successful and unsuccessful interventions addressed complexity.” They go further to explain that CLDs go beyond linear cause and effect logic models used in traditional evaluation and lessons regarding unintended consequences provide insights “to increase the chances of success for new prevention initiatives.”

Enhance Stakeholder and Community Participation

As discussed above, group model building (GMB) was a frequently reported process to create CLDs and inherent in these processes was the desire for stakeholder and/or community participation and shared understanding ( Table 4 ). Gerritsen et al. [ 41 ] stated what many others did, that is, GMB helped people develop an understanding of the system under study and that “participants learn to see causal connections and how these connections result in patterns of behaviour evolving over time.” They hypothesized that resulting plans for system change would be more successful with this fundamental level of participation and understanding. Another article highlighted that GMB brought diverse stakeholders “together to develop a system understanding of the problem, thus paving the way for further collaboration and community action” [ 44 ].

Inform Future Research and Enhance Theoretical Perspectives

The final two intended uses of CLDs were to inform future research and enhance theoretical perspectives ( Table 4 ). These intended uses were not widely discussed and if at all, they were mostly short aspirational statements. However, one example where future research was explicitly discussed was provided by Swierad et al. [ 44 ]. Here they reported that “hypotheses” from a CLD of childhood obesity could be used in future research such as “impact of food eaten at school influencing norms and acceptability of western/packaged food, elasticity of grandparents’ food norms, diversity of grandparents’ ideal body image for children, or beliefs in health of traditional foods.”

With respect to using CLDs to enhance theoretical perspectives, Clarke et al. [ 37 ] suggested that the CLD “enhanced previously published theoretical analyses of obesity prevention policy decision-making systems by making explicit how underlying feedback loops either spurred policy change or resistance.” Another example is from Burrell et al. [ 36 ]. They reported that creating a CLD resulted in “a testable ecologically oriented theory of violence” and “the resulting model conveys new theoretical insights on how racial and economic features of urban settings interact with intrapsychic dimensions to create a self-perpetuating system of violence.”

This section answers our second research question: What recommendations emerge regarding how to create and use CLDs in public health research? We offer nine learnings from the results above and interweave ideas from other research to support preliminary recommendations or possible directions to take forward in future research.

Boundary Judgements

We learned that some articles described in detail theoretical orientations with respect to complex systems thinking while others gave brief explanations. The most frequent concepts regarding complex systems were the inherent dynamic interactions among many entities, factors, variables that illustrate whole system structure and behavior. This is consistent with other public health literature on the topic [ 52 – 54 ]. The difference in descriptions was more a matter of comprehensiveness than definitions. For example, boundary judgement was not well articulated in the articles. According to Ulrich [ 55 ], drawing boundaries builds in selectivity and partiality and therefore transparency is important in study design. Therefore, we recommend that attention be given to defining boundaries to signal a specific endogenous perspective and a unique, snap-shot-in-time diagram of feedback loops of system behavior [ 56 ].

From Theory to Leverage Points

Some articles had strong theoretical coherence with respect to complex systems thinking that was demonstrated in discussions about the reasons for choosing, creating, and using CLDs. We learned that articles were most coherent when they first discussed feedback loops from a theoretical perspective and then carried this through to creating CLDs and to using them to identify leverage points for systems change (see for example 30). Overall, the descriptions of feedback in the articles were aligned with the idea that CLDs are “the applications of the loop concept underlying feedback and mutual causality” and that feedback loops are “powerful unifying notions that illuminate the structure of arguments, explanations, and causal views” [ 56 ]. Meadows [ 51 ] is well-known for explaining that disrupting or amplifying feedback loops can be effective leverage points in systems change. Therefore, we recommend that future research be designed with this theoretical coherence in mind.

Theoretical Frameworks

Lewin’s famous statement that “there is nothing so practical as good theory” was salient for what we learned [ 57 ]. Few articles used theoretical frameworks in research design or discussed the need to advance theory (i.e., complexity, systems) in public health research. The articles that used frameworks appeared to be more robust especially with respect to embedding theoretical constructs in the resultant CLD (see for example 35). While we appreciate that theory is emerging, we recommend that this be given more emphasis to help continue to build a solid foundation for furthering the application of CLDs in public health research.

Qualitative Data Analysis

Knai et al. [ 30 ] pointed out that current public health research “concentrates mainly on a system’s elements rather than the interconnections within it, and this is beginning to reveal its intrinsic limitations.” Some articles described data analysis methods to identify variables and examine interconnections to draw CLDs, however, others lacked clear descriptions of the often highly iterative methods and therefore it was difficult to follow a data trail and assess the resultant CLD. We recommend that more clarity be provided as to how researchers innovate in qualitative data analysis to further develop the art and science of creating CLDs.

Mixed Methods

We found a range of research methods used to create CLDs. Ozawa et al. [ 58 ] state that mixed methods research is important

because it allows researchers to view problems from multiple perspectives, contextualize information, develop a more complete understanding of a problem, triangulate results, quantify hard-to-measure constructs, provide illustrations of context for trends, examine processes/experiences along with outcomes and capture a macro picture of a system.

We hypothesize that mixed methods may produce more robust CLDs, however, this needs to be examined. We recommend that future research be undertaken to assess the strengths, limitations, and benefits of using mixed methods and determine what methods create greater confidence in the variables and feedback loops illustrated in CLDs.

Participatory Action Research

We found there was a wide range of who was involved in creating CLDs, from researchers only to multiple group model building sessions with stakeholders and community members. We see the latter methodology embedded in the traditions of action research [ 59 ] and/or community-based participatory research (CBPR) [ 60 ]. The CBPR approach involves “a commitment to conducting research that shares power with and engages community partners in the research process” and is intended “to increase knowledge and understanding of a given phenomena and integrate knowledge gained with interventions and policy and social change” [ 60 ]. There was little discussion of CBPR in the articles. We recommend that greater engagement with participatory action research literature be undertaken to embed the theory and philosophy of genuine participation and empowerment in research and action.

Knowledge Translation

There was limited discussion regarding how exactly CLDs were to be used to enhance evidence-informed policy and practice. Few articles explicitly discussed incorporating knowledge users or those able to use research results. As Sturmberg [ 61 ] relates, this requires users who are “deeply interested in understanding the highly interconnected and interdependent nature of the issues.” This led us to think about the importance of knowledge translation (KT) and how to strengthen the use of CLDs. Haynes et al. [ 6 ] state that KT needs to be conceptualized as not “a discrete piece of work within wider efforts to strengthen public health, but as integral to and in continual dialogue with those efforts.” We recommend that future public health research using CLDs should articulate KT plans that articulates knowledge user engagement in defining outcomes for strengthening public health policies and practices.

Health Equity

We conceptualize public health research to be guided by principles of social justice and human rights to address the goal of reducing health inequities through action on the determinants of health. Although many articles discussed determinants of health, the goal of reducing health inequities was largely absent. Baum et al. [ 62 ] discuss the concept of path dependency as “the tendency of institutions to retain policy directions and preferences rather than change or reform them.” They further suggest that disrupting “path dependency that exacerbates health inequities” is critical and we see how CLDs could uncover path dependencies. We recommend that CLDs in public health research should include the examination of leverage points for pro-equity policy and practice.

The Diagram

Senge [ 63 ] states that “reality is made up of circles” but often arguments and explanations are linear, therefore, CLDs can provide “a language of interrelationships” to uncover deep patterns in systems. Studying the interrelationships and explanations of each CLD was outside the scope of this paper, however, we learned about some basic elements of reader friendly CLDs. We recommend that the following questions could be used assess CLDs: Are established conventions [ 56 ] used effectively for drawing the CLD (e.g., labeling, positive and negative arrows, reinforcing and balancing loops)? Does the diagram illuminate the most significant variables, feedback loops or leverage points? How well does the diagram function as an effective medium for presenting findings to knowledge users? How well does the CLD tell a story of what’s going on in a system?

Strengths and Limitations

In terms of limitations, the 23 articles were not considered to be comprehensive. Since completing the study, we found that Mui and others [ 64 ] published an article on a community-based system dynamics approach and suggests solutions for improving healthy food access in a low-income urban environment. We also found Savona et al. [ 65 ] identified the views of adolescents regarding the causes of obesity and used CLDs. While this can be considered a limitation, we hope to see a continual building of knowledge and skill in using CLDs in public health research. A strength of this paper is that 23 recent articles were identified that used CLDs and the depth and breadth of discussion in the articles provided good representation. Having three authors conduct the literature review is also a strength because this afforded a high degree of confidence in reporting results and transparency in search strategy and data extraction, analysis and synthesis. Together the results and recommendations can contribute to informing global public health research by highlighting key considerations to help design research and address public health issues through complex systems thinking.

Author Contributions

LB designed the overall research aim and questions and CN provided input throughout the study. Study selection was conducted by LB. Appraisal and duplicate independent data extraction and validation was conducted by two authors (LB and CH). LB and CH completed data analysis and all authors (LB, CH, and CN) provided input into writing the final manuscript.

The authors declare that this study received funding from Canadian Institutes for Health Research (LB/Postdoctoral fellowship) and the College of Medicine, University of Saskatchewan (CH/Dean’s summer research project scholarship). Funding for the open access publication fee will be covered through the Canadian Institutes for Health Research award. The funders were not involved in the study design, collection, analysis, interpretation of data, the writing of this article or the decision to submit it for publication.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

1. Rutter, H, Savona, N, Glonti, K, Bibby, J, Cummins, S, Finegood, D, et al. The Need for a Complex Systems Model of Evidence for Public Health. The Lancet (2017) 390(10112):2602–4. doi:10.1016/s0140-6736(17)31267-9

CrossRef Full Text | Google Scholar

2. Carey, G, Malbon, E, Carey, N, Joyce, A, Crammond, B, and Carey, A. Systems Science and Systems Thinking for Public Health: a Systematic Review of the Field. BMJ Open (2015) 5(12):e009002. doi:10.1136/bmjopen-2015-009002

3. Chughtai, S, and Blanchet, K. Systems Thinking in Public Health: a Bibliographic Contribution to a Meta-Narrative Review. Health Policy Plan (2017) 32(4):585–94. doi:10.1093/heapol/czw159

PubMed Abstract | CrossRef Full Text | Google Scholar

4.Canadian Public Health Association. Public Health: A Conceptual Framework . Ottawa (2017).

Google Scholar

5. Baugh Littlejohns, L, and Wilson, A. Strengthening Complex Systems for Chronic Disease Prevention: a Systematic Review. BMC Public Health (2019) 19(1):729. doi:10.1186/s12889-019-7021-9

6. Haynes, A, Rychetnik, L, Finegood, D, Irving, M, Freebairn, L, and Hawe, P. Applying Systems Thinking to Knowledge Mobilisation in Public Health. Health Res Pol Syst (2020) 18(1):134. doi:10.1186/s12961-020-00600-1

7. Egan, M, McGill, E, Penney, T, Meier, PS, Savona, N, de Vocht, F, et al. Complex Systems for Evaluation of Public Health Interventions: a Critical Review. The Lancet (2018) 392:S31. doi:10.1016/s0140-6736(18)32053-1

8. McGill, E, Er, V, Penney, T, Egan, M, White, M, Meier, P, et al. Evaluation of Public Health Interventions from a Complex Systems Perspective: A Research Methods Review. Soc Sci Med (2021) 272:113697. doi:10.1016/j.socscimed.2021.113697

9. McGill, E, Marks, D, Er, V, Penney, T, Petticrew, M, and Egan, M. Qualitative Process Evaluation from a Complex Systems Perspective: A Systematic Review and Framework for Public Health Evaluators. Plos Med (2020) 17(11):e1003368. doi:10.1371/journal.pmed.1003368

10. Rusoja, E, Haynie, D, Sievers, J, Mustafee, N, Nelson, F, Reynolds, M, et al. Thinking about Complexity in Health: A Systematic Review of the Key Systems Thinking and Complexity Ideas in Health. J Eval Clin Pract (2018) 24(3):600–6. doi:10.1111/jep.12856

11. Forrester, JW. Learning through System Dynamics as Preparation for the 21st Century. Syst Dyn Rev (2016) 32(3-4):187–203. doi:10.1002/sdr.1571

12. Forrester, J. Some Basic Concepts in System Dynamics . Boston: Sloan School of Management, Massachusetts Institute of Technology (2009).

13. Baugh Littlejohns, L, Baum, F, Lawless, A, and Freeman, T. The Value of a Causal Loop Diagram in Exploring the Complex Interplay of Factors that Influence Health Promotion in a Multisectoral Health System in Australia. Health Res Pol Syst (2018) 16:126. doi:10.1186/s12961-018-0394-x

14. Paina, L, Bennett, S, Ssengooba, F, and Peters, DH. Advancing the Application of Systems Thinking in Health: Exploring Dual Practice and its Management in Kampala, Uganda. Health Res Pol Sys (2014) 12(1):41. doi:10.1186/1478-4505-12-41

15. Varghese, J, Kutty, VR, Paina, L, and Adam, T. Advancing the Application of Systems Thinking in Health: Understanding the Growing Complexity Governing Immunization Services in Kerala, India. Health Res Pol Sys (2014) 12(1):47. doi:10.1186/1478-4505-12-47

16. Agyepong, IA, Aryeetey, GC, Nonvignon, J, Asenso-Boadi, F, Dzikunu, H, Antwi, E, et al. Advancing the Application of Systems Thinking in Health: Provider Payment and Service Supply Behaviour and Incentives in the Ghana National Health Insurance Scheme - a Systems Approach. Health Res Pol Sys (2014) 12:35. doi:10.1186/1478-4505-12-35

17. Biroscak, BJ, Schneider, T, Panzera, AD, Bryant, CA, McDermott, RJ, Mayer, AB, et al. Applying Systems Science to Evaluate a Community-Based Social Marketing Innovation. Soc Marketing Q (2014) 20(4):247–67. doi:10.1177/1524500414556649

18. Allender, S, Owen, B, Kuhlberg, J, Lowe, J, Nagorcka-Smith, P, Whelan, J, et al. .A Community Based Systems Diagram of Obesity Causes. PLoS ONE (2015) 10:e0129683. doi:10.1371/journal.pone.0129683

19. Homer, J, Milstein, B, Wile, K, Pratibhu, P, Farris, R, and Orenstein, DR. Modeling the Local Dynamics of Cardiovascular Health: Risk Factors, Context, and Capacity. Prev Chronic Dis (2008) 5(2):A63.

PubMed Abstract | Google Scholar

20. Brennan, LK, Sabounchi, NS, Kemner, AL, and Hovmand, P. Systems Thinking in 49 Communities Related to Healthy Eating, Active Living, and Childhood Obesity. J Public Health Manag Pract (2015) 21:S55–69. doi:10.1097/PHH.0000000000000248

21. Kopelman, P, Jebb, SA, and Butland, B. Foresight. Tackling Obesities: Future Choices - Project Report . London: Government Office for Science (2007).

22. Munn, Z, Peters, MDJ, Stern, C, Tufanaru, C, McArthur, A, and Aromataris, E. Systematic Review or Scoping Review? Guidance for Authors when Choosing between a Systematic or Scoping Review Approach. BMC Med Res Methodol (2018) 18(1):143. doi:10.1186/s12874-018-0611-x

23. Haddaway, NR, Woodcock, P, Macura, B, and Collins, A. Making Literature Reviews More Reliable through Application of Lessons from Systematic Reviews. Conservation Biol (2015) 29(6):1596–605. doi:10.1111/cobi.12541

24. Byrne, JA. Improving the Peer Review of Narrative Literature Reviews. Res Integr Peer Rev (2016) 1:12. doi:10.1186/s41073-016-0019-2

25. Hsiu-Fang, H, and Shannon, SE. Three Approaches to Qualitative Content Analysis. Qual Health Res (2005) 15(9):1277. doi:10.1177/1049732305276687

26. Crielaard, L, Dutta, P, Quax, R, Nicolaou, M, Merabet, N, Stronks, K, et al. Social Norms and Obesity Prevalence: From Cohort to System Dynamics Models. Obes Rev (2020) 21(9):e13044. doi:10.1111/obr.13044

27. Eker, NZ, Carnohan, S, and Davies, M. Participatory System Dynamics Modelling for Housing, Energy and Wellbeing Interactions. Build Res Inf (2018) 46(7):738–54. doi:10.1080/09613218.2017.1362919

28. Brereton, CF, and Jagals, P. Applications of Systems Science to Understand and Manage Multiple Influences within Children's Environmental Health in Least Developed Countries: A Causal Loop Diagram Approach. Int J Environ Res Public Health (2021) 18(6). doi:10.3390/ijerph18063010

29. Klement, RJ. Systems Thinking about SARS-CoV-2. Front Public Health (2020) 8(650):585229. doi:10.3389/fpubh.2020.585229

30. Knai, C, Petticrew, M, Mays, N, Capewell, S, Cassidy, R, Cummins, S, et al. Systems Thinking as a Framework for Analyzing Commercial Determinants of Health. Milbank Q (2018) 96(3):472–98. doi:10.1111/1468-0009.12339

31. Osman, M, Karat, AS, Khan, M, Meehan, S-A, von Delft, A, Brey, Z, et al. Health System Determinants of Tuberculosis Mortality in South Africa: a Causal Loop Model. BMC Health Serv Res (2021) 21(1):388. doi:10.1186/s12913-021-06398-0

32. Sahin, O, Salim, H, Suprun, E, Richards, R, MacAskill, S, Heilgeist, S, et al. Developing a Preliminary Causal Loop Diagram for Understanding the Wicked Complexity of the COVID-19 Pandemic. Systems (2020) 8(2):20. doi:10.3390/systems8020020

33. Urwannachotima, N, Hanvoravongchai, P, and Ansah, JP. Sugar-sweetened Beverage Tax and Potential Impact on Dental Caries in Thai Adults: An Evaluation Using the Group Model Building Approach. Syst Res (2019) 36(1):87–99. doi:10.1002/sres.2546

34. Bradley, DT, Mansouri, MA, Kee, F, and Garcia, LMT. A Systems Approach to Preventing and Responding to COVID-19. EClinicalMedicine (2020) 21:100325. doi:10.1016/j.eclinm.2020.100325

35. Bensberg, M, Joyce, A, and Wilson, E. Building a Prevention System: Infrastructure to Strengthen Health Promotion Outcomes. Int J Environ Res Public Health (2021) 18(4):1618. doi:10.3390/ijerph18041618

36. Burrell, M, White, AM, Frerichs, L, Funchess, M, Cerulli, C, DiGiovanni, L, et al. Depicting "the System": How Structural Racism and Disenfranchisement in the United States Can Cause Dynamics in Community Violence Among Males in Urban Black Communities. Soc Sci Med (2021) 272:113469. doi:10.1016/j.socscimed.2020.113469

37. Clarke, B, Kwon, J, Swinburn, B, and Sacks, G. Understanding the Dynamics of Obesity Prevention Policy Decision-Making Using a Systems Perspective: A Case Study of Healthy Together Victoria. PLoS ONE (2021) 16(1):e0245535. doi:10.1371/journal.pone.0245535

38. Jalali, MS, Rahmandad, H, Bullock, SL, Lee-Kwan, SH, Gittelsohn, J, and Ammerman, A. Dynamics of Intervention Adoption, Implementation, and Maintenance inside Organizations: the Case of an Obesity Prevention Initiative. Soc Sci Med (2019) 224:67–76. doi:10.1016/j.socscimed.2018.12.021

39. Owen, B, Brown, AD, Kuhlberg, J, Millar, L, Nichols, M, Economos, C, et al. Understanding a Successful Obesity Prevention Initiative in Children under 5 from a Systems Perspective. PLoS One (2018) 13(3):e0195141. doi:10.1371/journal.pone.0195141

40. Parmar, PK, Rawashdah, F, Al-Ali, N, Abu Al Rub, R, Fawad, M, Al Amire, K, et al. Integrating Community Health Volunteers into Non-communicable Disease Management Among Syrian Refugees in Jordan: a Causal Loop Analysis. BMJ Open (2021) 11(4):e045455. doi:10.1136/bmjopen-2020-045455

41. Gerritsen, S, Harré, S, Rees, D, Renker-Darby, A, Bartos, AE, Waterlander, WE, et al. Community Group Model Building as a Method for Engaging Participants and Mobilising Action in Public Health. Int J Environ Res Public Health (2020) 17(10). doi:10.3390/ijerph17103457

42. Maitland, N, Williams, M, Jalaludin, B, Allender, S, Strugnell, C, Brown, A, et al. Campbelltown - Changing Our Future: Study Protocol for a Whole of System Approach to Childhood Obesity in South Western Sydney. BMC Public Health (2019) 19(1):1699. doi:10.1186/s12889-019-7936-1

43. Riley, T, Hopkins, L, Gomez, M, Davidson, S, Chamberlain, D, Jacob, J, et al. A Systems Thinking Methodology for Studying Prevention Efforts in Communities. Syst Pract Action Res (2020) 34 (5):555–73.

44. Swierad, E, Huang, TT, Ballard, E, Flórez, K, and Li, S. Developing a Socioculturally Nuanced Systems Model of Childhood Obesity in Manhattan's Chinese American Community via Group Model Building. J Obes (2020) 2020:4819143. doi:10.1155/2020/4819143

45. Araz, O, Wilson, F, and Stimpson, J. Complex Systems Modeling for Evaluating Potential Impact of Traffic Safety Policies: a Case on Drug-Involved Fatal Crashes. Ann Operations Res (2020) 291(1-2):37–58. doi:10.1007/s10479-018-2961-5

46. Hassmiller Lich, K, Urban, JB, Frerichs, L, and Dave, G. Extending Systems Thinking in Planning and Evaluation Using Group Concept Mapping and System Dynamics to Tackle Complex Problems. Eval Program Plann (2017) 60:254–64. doi:10.1016/j.evalprogplan.2016.10.008

47. Allender, S, Brown, AD, Bolton, KA, Fraser, P, Lowe, J, and Hovmand, P. Translating Systems Thinking into Practice for Community Action on Childhood Obesity. Obes Rev (2019) 20:179–84. doi:10.1111/obr.12865

48. Brown, A, Millar, L, Hovmand, PS, Kuhlberg, J, Love, P, Nagorcka-Smith, P, et al. Learning to Track Systems Change Using Causal Loop Diagrams. Obes Res Clin Pract (2019) 13:73–4. doi:10.1016/j.orcp.2016.10.210

49. Kim, H, and Andersen, DF. Building Confidence in Causal Maps Generated from Purposive Text Data: Mapping Transcripts of the Federal Reserve. Syst Dyn Rev (2012) 28(4):311–28. doi:10.1002/sdr.1480

50. Foster-Fishman, PG, Nowell, B, and Yang, H. Putting the System Back into Systems Change: a Framework for Understanding and Changing Organizational and Community Systems. Am J Community Psychol (2007) 39(3-4):197–215. doi:10.1007/s10464-007-9109-0

51. Meadows, D. Leverage Points: Places to Intervene in a System . Norwich, Vermont: Donella Meadows Institute (1999).

52. Mabry, PL, Marcus, SE, Clark, PI, Leischow, SJ, and Méndez, D. Systems Science: a Revolution in Public Health Policy Research. Am J Public Health (2010) 100(7):1161–3. doi:10.2105/AJPH.2010.198176

53. Finegood, D. The Complec Science of Obesity. In: J Cawley, editor. Handbook of the Social Science of Obesity . Oxford: Oxford University Press (2011).

54. Hawe, P, Shiell, A, and Riley, T. Theorising Interventions as Events in Systems. Am J Community Psychol (2009) 43(3-4):267–76. doi:10.1007/s10464-009-9229-9

55. Ulrich, W. Beyond Methodology Choice: Critical Systems Thinking as Critically Systemic Discourse. J Oper Res Soc (2003) 54(4):325–42. doi:10.1057/palgrave.jors.2601518

56. Richardson, G. Feedback Thought in Social Science and Systems Theory . Waltham, MA: Pegasus Communications, Inc. (1999).

57. Lewin, K. Action Research and Minority Problems. J Soc Issues (1946) 2:34–46. doi:10.1111/j.1540-4560.1946.tb02295.x

58. Ozawa, S, and Pongpirul, K. 10 Best Resources on … Mixed Methods Research in Health Systems. Health Policy Plan (2013) 29(3):323–7. doi:10.1093/heapol/czt019

59. Bradbury, H, Reason, P, Publishing, E, and Ebsco, V. The Sage Handbook of Action Research : Participative Inquiry and Practice . 2nd ed. London: SAGE (2012). p. ed2008.

60.B Israel, E Eng, A Schultz, and E Parker, editors. Methods in Community-Based Participatory Research in Health . San Fransico, CA: Jossey-Bass (2005).

61. Sturmberg, J. Without Systems and Complexity Thinking There Is No Progress - or Why Bureaucracy Needs to Become Curious. Int J Health Pol Manag (2021) 10(5):227–80. doi:10.34172/ijhpm.2020.45

62. Baum, F, Townsend, B, and Fisher, M. Creating Political Will for Action on Health Equity: Practical Lessons for Public Health Policy Actors. Int J Health Pol Manag (2020) x:1–14. doi:10.34172/ijhpm.2020.233

63. Senge, P. The Fifth Discipline . New York: Doubleday (2006).

64. Mui, Y, Ballard, E, Lopatin, E, Thornton, RLJ, Pollack Porter, KM, and Gittelsohn, J. A Community-Based System Dynamics Approach Suggests Solutions for Improving Healthy Food Access in a Low-Income Urban Environment. PLoS ONE (2019) 14:e0216985–13. doi:10.1371/journal.pone.0216985

65. Savona, N, Macauley, T, Aguiar, A, Banik, A, Boberska, M, Brock, J, et al. Identifying the Views of Adolescents in Five European Countries on the Drivers of Obesity Using Group Model Building. Eur J Public Health (2021) 31(2):391–6. doi:10.1093/eurpub/ckaa251

Keywords: scoping review, causal loop diagrams, public health research, methods, complex systems thinking

Citation: Baugh Littlejohns L, Hill C and Neudorf C (2021) Diverse Approaches to Creating and Using Causal Loop Diagrams in Public Health Research: Recommendations From a Scoping Review. Public Health Rev 42:1604352. doi: 10.3389/phrs.2021.1604352

Received: 16 July 2021; Accepted: 25 November 2021; Published: 14 December 2021.

Reviewed by:

Copyright © 2021 Baugh Littlejohns, Hill and Neudorf. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

PHR is edited by the Swiss School of Public Health (SSPH+) in a partnership with the Association of Schools of Public Health of the European Region (ASPHER)+

*Correspondence: Cory Neudorf, [email protected]

Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.

• Open access
• Published: 29 December 2018

The value of a causal loop diagram in exploring the complex interplay of factors that influence health promotion in a multisectoral health system in Australia

• Lori Baugh Littlejohns   ORCID: orcid.org/0000-0003-1766-4770 1 , 4 ,
• Fran Baum 2 ,
• Angela Lawless 3 &
• Toby Freeman 2  

Health Research Policy and Systems volume  16 , Article number:  126 ( 2018 ) Cite this article

19k Accesses

52 Citations

32 Altmetric

Metrics details

Despite calls for the application of complex systems science in empirical studies of health promotion, there are very few examples. The aim of this paper was to use a complex systems approach to examine the key factors that influenced health promotion (HP) policy and practice in a multisectoral health system in Australia.

Within a qualitative case study, a schema was developed that incorporated HP goals, actions and strategies with WHO building blocks (leadership and governance, financing, workforce, services and information). The case was a multisectoral health system bounded in terms of geographical and governance structures and a history of support for HP. A detailed analysis of 20 state government strategic documents and interviews with 53 stakeholders from multiple sectors were completed. Based upon key findings and dominants themes, causal pathways and feedback loops were established. Finally, a causal loop diagram was created to visualise the complex array of feedback loops in the multisectoral health system that influenced HP policy and practice.

The complexity of the multisectoral health system was clearly illustrated by the numerous feedback mechanisms that influenced HP policy and practice. The majority of feedback mechanisms in the causal loop diagram were vicious cycles that inhibited HP policy and practice, which need to be disrupted or changed for HP to thrive. There were some virtuous cycles that facilitated HP, which could be amplified to strengthen HP policy and practice. Leadership and governance at federal–state–local government levels figured prominently and this building block was interdependently linked to all others.

Creating a causal loop diagram enabled visualisation of the emergent properties of the case health system. It also highlighted specific leverage points at which HP policy and practice can be improved. This paper demonstrates the critical importance of leveraging leadership and governance for HP and adds urgency to the need for increased and strong advocacy efforts targeting all levels of government in multisectoral health systems.

Peer Review reports

The application of complex systems science to health promotion (HP) has much promise [ 1 ]. There are, however, few published empirical studies that discuss its application in order to study HP policy and practice and demonstrate its practical value. This paper reports on the application of a complex systems approach to study the key factors influencing HP policy and practice in an Australian multisectoral health system. First, an explanation of how HP and a complex health system are conceptualised followed by the gaps identified in the literature are provided.

• Health promotion

The WHO definition of HP is “ the process of enabling people to increase control over, and to improve their health. It moves beyond a focus on individual behaviour to consider a wide range of social and environmental interventions ” [ 2 ]. This definition points to the importance of multilevel (individual through to societal) and multisectoral (many sectors, including health) action on the social, economic and environmental determinants of health as central to the desired HP policy and practice. Evidence indicates that these structural drivers in society are pivotal determinants of health inequities [ 3 , 4 , 5 , 6 ].

This paper takes the goal of HP as promoting overall population health and reducing health inequities, that is, the preventable and unfair distribution of the determinants of health [ 4 ]. The conceptualisation of HP used in this paper is based on WHO’s Ottawa Charter for Health Promotion [ 2 ]. Reorienting health services towards HP [ 2 , 6 ], ensuring community participation in identifying and addressing priority determinants of health [ 2 , 7 ], and developing partnerships and intersectoral collaboration to take coordinated action [ 8 ] can be regarded as the fundamental processes or actions through which HP strategies need to be planned, implemented and evaluated (Table  1 ). Developing personal skills, creating supportive environments and building healthy public policies [ 2 ] represent three strategies to take action to address the goal of HP (Table 1 ).

Although international documents have long recommended the actions and strategies described above, there remain significant challenges for multisectoral health systems (further described below) to adopt policies and practice that are focused on reducing health inequities [ 3 , 4 , 9 , 10 ].

Complex health systems

HP is challenging, not only in terms of the range and interrelationships among determinants of health, but also the complex systems that shape HP policy and practice [ 11 ]. Health systems can be described in terms of the broad and numerous social systems that influence health and well-being as well as clinical healthcare services [ 12 ]. Multisectoral health systems are complex, primarily because of interactions, feedback and emergent order within systems [ 13 , 14 , 15 , 16 , 17 ]. Figure  1 illustrates these characteristics and their relationship to one another.

Three characteristics of complex systems [ 32 ] (used with permission from A. Strauss & Associates; http://maverickandboutique.com )

Interactions

Complex systems have numerous nested and heterogeneous system elements that exhibit considerable variation, with each element being a system in their own right [ 18 , 19 , 20 ]. Health systems are complex because they are comprised of multiple entities, organisations, agencies and sectors at local, regional, state, national and international levels, all of which vary in terms of their structure, function and interests. Each element has a unique relationship to and influence on the whole health system [ 21 ]. Key to understanding multisectoral health systems are the interactions among elements that influence the overall health system [ 13 , 18 , 22 ].

Complex systems are dynamic because of their continuous ability to change, adapt and reorganise in response to their environment [ 23 ]. Self-organisation is a concept used to describe the adaptation of systems to their environment and to study how systems organise, change and/or innovate [ 24 , 25 , 26 ]. Feedback loops are the interconnections that illustrate self-organisation in complex systems [ 25 , 27 ]. The behaviour of complex systems is in large part the accumulative effect of positive (reinforcing or self-enhancing) and negative (balancing or goal seeking) feedback mechanisms [ 24 , 28 ]. ‘Virtuous’ and ‘vicious’ are descriptors of feedback loops that are going in favourable or unfavourable directions.

Emergent order

Interactions and feedback mechanisms produce emergent order or properties of the whole system [ 29 ]. Emergent properties therefore cannot be inferred by the study of individual system elements or variables but rather through the study of relationships in the whole system [ 26 ]. Other factors that influence emergence include the history and context of the system [ 30 , 31 ]. Emergent order in health systems was described by Jayasinghe as follows: “ patterns of population health outcomes are an emergent property of the system. They arise from a web of causations that result from interactions among dynamic sets of interconnected systems ” ([ 33 ], p. 5).

Health system building blocks

One way to study health promotion in complex health systems is through the lens of WHO’s health system building block framework [ 34 ]. The framework describes the key capacities or building blocks needed for effective functioning, and these provide a way to study the complexity of health systems in terms of the interactions and feedback mechanisms among building blocks and the resultant emergent order. Harnessing the synergies created between interacting building blocks is considered instrumental to achieving health system goals or a desirable emergent order [ 35 ]. The adaptation of the framework to the study of a multisectoral health system for HP is described below.

Gaps in the literature

Despite the potential of using a complex systems approach that incorporates health system building blocks to study HP in multisectoral health systems, this appears not to have been done previously. Further to this, few studies focus on the interactions and feedback mechanisms that influence the emergent order in health systems with respect to HP policy and practice. The study of interactions and feedback could be very enlightening as Allender et al. [ 36 ] showed in terms of the causes of obesity in Australia. In a similar vein, the Foresight Centre [ 37 ] in the United Kingdom also illustrated interactions, feedback and emergent order with respect to causes or determinants of obesity within systems. In a Canadian HP policy study, Alvaro et al. [ 38 ] found a “ lopsided ” emphasis on individual lifestyle and behavioural approaches. They discussed positive or reinforcing feedback loops characterised as vicious cycles because they maintain the focus on HP strategies targeting individuals as opposed to a balance of strategies that also address societal or structural determinants of health. Building on these examples, a study of the interactions and feedback mechanisms among system elements and building blocks and the emergent order created in a multisectoral health system for HP appeared a promising way forward. This paper reports on research assessing the feedback mechanisms that appear to influence HP policy and practice in a multisectoral health system in Australia.

The first part of this section describes a case study approach and indicates the sectors, system elements and levels that bound the case and, thus, the multisectoral health system. Boundaries are the borders between complex systems and their environments and these are often “ fuzzy ” [ 39 ]. This ‘fuzziness’ applies to complex health systems and, by drawing boundaries of a multisectoral health system (i.e. delineating system elements, stakeholders and variables), it is possible to study feedback mechanisms [ 18 ]. Following this, data collection and analysis methods are explained, including document review, interviews, coding and how feedback mechanisms were identified.

This research was a single instrumental case study [ 40 ] and used qualitative methods. Luck et al. describe a case study as “ a detailed, intensive study of a particular contextual and bounded phenomena that is undertaken in real life situations ” ([ 41 ], p. 104). The case was a multisectoral health system in a region of South Australia (SA). The region is not identified at the request of stakeholders who were interviewed.

The multisectoral health system that formed the case study was selected based on the following attributes: (1) it was bounded in terms of geography and institutional governance structures of a Local Health Network and had co-terminus boundaries with four local governments (see Table  2 below for descriptions); (2) there were numerous and diverse sectors and system elements with roles in HP; and (3) there was a history of support for and action on HP, including intersectoral collaboration among sectors and subsystems [ 44 ]. At the time of the research (2013), the multisectoral health system was shaped by numerous federal, state, regional and local entities (i.e. sectors and system elements) and a range of governance structures. Table 2 provides a brief description of relevant sectors and system elements that were in place in the multisectoral health system.

Data collection and analysis

Table  3 provides an overview of data collection and analysis methods. To summarise, document analysis was used to first assess the extent to which the policy context – as formally articulated in policy and related strategic documents – supported the goal, actions and strategies conceptualised for HP and health system building blocks for HP. Interviews were then conducted with stakeholders in leadership roles in HP (Tables  2 and 3 ) to explore their perspectives of and experiences in the HP policy and practice environment. A semi-structured interview guide was used to ask questions in the following areas: details of individual and organisational roles in HP, descriptions of and changes in the HP policy and practice environment, and perspectives concerning the key factors that influence HP policy and practice. Interviews provided contextual information and explored the implementation of policy intentions.

Table  4 provides the unique coding schema for document review and interview data. Using NVivo software, documents and interview transcriptions were coded according to their reference to the HP goal, actions, and strategies and health system building blocks. Definitions of health system building blocks [ 34 ] were adapted to better reflect the capacities needed for HP in multisectoral health systems. Further, ‘medicines and technologies’ was not included in the schema as it relates mostly to clinical healthcare in the health sector as opposed to multisectoral health systems for HP.

Following coding and analysis, a summary of key findings was completed for both document analysis and interview data. A detailed discussion of the document review analysis as well as results can be found elsewhere [ 47 ]. Based on key findings, a complex system lens was applied to identify interactions, feedback and emergent properties in the multisectoral health system with respect to HP policy and practice. Kim and Andersen’s [ 48 ] process was adapted to link key findings from document review and interviews to feedback mechanisms through the identification of dominant themes. This involved five steps, as follows:

When a key finding was found in both data sets, it was labelled a dominant theme (Table 5 ).

Causal links were then identified among dominant themes and key findings. Several criteria described by Davidson [ 49 ] for inferring causality were used including temporal precedence (i.e. establishing A before B), constant conjunction (i.e. when A, always B), and contiguity of influence (i.e. plausible mechanisms for linking A and B). This process was intensely iterative and ended only when each causal link was clearly substantiated.

Following this, causal links were translated into words-and-arrows diagrams with each representing an interaction.

When a causal link demonstrated a reciprocal relationship, a feedback loop was created. Each feedback loop was assessed in terms of its polarity (positive polarity signifying a reinforcing relationship and negative polarity signifying a balancing relationship) thus establishing whether the loop was a facilitating or inhibiting factor for HP policy and practice [ 50 ].

All feedback loops were then assembled into a causal loop diagram to create a visual model [ 51 ]. Vensim PLE software was used to create word-and-arrow diagrams, feedback loops and the causal loop diagram. In the interest of providing a more reader-friendly diagram, facilitating (happy face) or inhibiting (sad face) influences on HP policy and practice in the case health system were used (i.e. the polarity of each feedback loop is not labelled).

First, an overview of the HP policy and practice context followed by key findings from the document analysis and interviews are presented. The next section interweaves reporting on dominant themes and the feedback mechanisms identified. Finally, the causal loop diagram portraying all feedback mechanisms in play in the case study health system with respect to HP policy and practice is described.

Overview of HP policy and practice context

The policy context changed from strong advocacy for HP in 2003 to its near abandonment in 2013. From 2003 to 2011 there was considerable support for HP but this support diminished significantly in 2013 following the Review of Non-hospital Based Services [ 52 ] (hereafter called the Review) and SA Health’s Response [ 53 ]. The government’s response to the Review resulted in substantial cuts to HP financing, workforce and services, which are essential health systems building blocks. Documents identify that cuts were made because of (1) the poor state economic environment, rising healthcare costs and the need for budgetary constraints, (2) unclear federal-state roles, governance structures and policy directions, and (3) the lack of evidence regarding HP effectiveness [ 52 ]. More positively, the SA Public Health Act provided a foundation for partnership, intersectoral collaboration and whole-of-government approaches to HP. All interviewees, except those from the state health department (5 of 53), described the HP policy and practice environment in very negative terms because of the heavy cuts to HP proposed by the Review and accepted in SA Health’s Response. Several participants said that HP was now a “ dirty word ” (NGO/Health Service/Professional Association, Local Government). Other descriptors included “ big void ”, “ devalued ”, “ devastating ”, “ dire ”, “ expendable ”, “ obliteration ”, and “ toxic ”. However, some state health department interviewees characterised the HP policy and practice environment as the “ glass is half full ” because of the implementation of the SA Public Health Act, which laid out governance structures for collaboration between state and local governments.

Dominant themes and feedback mechanisms

Table 5 provides a list of key findings and illustrates, through check marks, if they were found in document review and/or interview data. Dominant themes are those where key findings were found in both document review and interview data (two check marks). In the following section, dominant themes are reported and feedback mechanisms identified. All feedback mechanisms are illustrated in one causal loop diagram (Fig.  2 ) and dominant themes are indicated through bold font. A detailed explanation of each feedback mechanism can be found in Additional file  1 (Description of causal links and feedback mechanisms).

• Causal loop diagram

Goal of reducing health inequities

There was little support for or discussion of reducing health inequities in either the document review or the interview data. When reducing health inequities was discussed (NGO/Intersectoral Network interviewee; SA Health group interview) it was limited to addressing the needs of disadvantaged people and discussion did not address the social gradient in health [ 4 ]. The lack of strong leadership and governance by way of strategic policy frameworks for the goal of reducing health inequities forms a detrimental feedback loop or a vicious cycle that inhibited HP policy and practice (Fig.  2 ).

Community participation in HP

Although many documents identified and supported community participation in HP (2003–2013), the Review and SA Health’s Response greatly diminished this because of cuts to the financing, workforce and services. The SA Public Health Act included a participation principle; however, it is weak in comparison to empowerment approaches to strengthen community action [ 2 , 54 ]. Although community participation was discussed by several interviewees as being important (particularly in working with disadvantaged populations), it was reported to be “ old hat ”, “ not modern ” (#2/NGO/Health service). Interviewees reported a retreat from the strong history of using community development approaches in primary healthcare services in the case health system (#29 and #50/NGO/Intersectoral Network). The lack of leadership and policy directions for HP policy and practice to facilitate community participation also formed an inhibiting and vicious feedback mechanism (Fig.  2 ).

Federal-state-local roles, governance structures and policy directions

The three levels of government figured prominently in the data. The Review and SA Health’s Response altered governance structures significantly as HP leadership at the local and regional (case) health system level was conceded to (1) the federal government through the federally funded and regionally managed Medicare Local and to (2) local governments through the SA Public Health Act. The Review was frequently discussed in interviews with stakeholders from all sectors. Stakeholders reported that the lack of implementation of an original federal-state National Health Care Reform Agreement [ 55 ] was a key factor and, as a result, HP was a “ casualty ” of the politics between levels of government because no level of government accepted leadership responsibilities. One interviewee was particularly expressive with respect to this: “ what we’ve got is an ad hoc, politically influenced, double–dipping, cherry picking, State-Commonwealth split ” (#2/NGO/Health Service).

Federal-state level

The lack of delineation of federal-state roles, governance structures and policy direction played out with respect to the state cuts to HP financing, workforce and services in the case health system because of a false assumption that the federal government’s Medical Local initiative would be doing this work. Thus, inhibiting and vicious feedback loops were found (Fig.  2 ).

State level

Most documents discussed the important leadership and health governance role of the state government in reorienting health services toward HP to some extent. Examples include discussion of moving from “ an illness focused to a health focused system ” ([ 56 ], p. 14) and enhancing HP through primary healthcare services [ 57 ]. The most striking finding was the abdication of these functions in 2013 following cuts to HP recommended in the Review and SA Health’s Response. The perceived lack of information regarding HP effectiveness was one reason given for cuts to HP financing, workforce and services. Almost all stakeholders, except some from SA Health, discussed the Review’s perspective on the effectiveness of HP. Many interviewees shared the concern that the Review did not use an appropriate evaluation framework based upon HP principles and practices. Although many stakeholders reported that evaluating the effectiveness of HP was a great challenge, several suggested that expecting to generate evidence of effectiveness in the case health system was futile given that so little had been invested in HP initiatives. There were three vicious and inhibiting feedback mechanisms with respect to the lack of state leadership and health governance, the lack of information regarding evidence of HP effectiveness, and cuts to HP financing, workforce and services (Fig.  2 ).

Furthermore, the Review and SA Health’s Response identified the state’s economic circumstances and budgetary constraints due to rising healthcare expenditures as key factors influencing cuts to HP financing, workforce and services. Calls in earlier documents for strong leadership and health governance to ensure adequate and sustained funding for HP were unheeded. Stakeholders used phrases such as “ soft target ” and “ easy target ” to explain the HP cuts (e.g., #4/Local Council, #35/NGO/Professional Association). One interviewee voiced what others reported, namely that the primary concern of the newly appointed Minister of Health was “ the great hole in the Health budget ” and cuts to HP were a “ quick political win ” in an election year (#2/NGO/Health service). Others reported that the cuts were very abrupt and “ they’re cutting their nose off to spite their face because of their focus on a balanced budget ” (#9/Medicare Local) and “ some things seem to pass with little controversy like enormous new ovals [cricket stadiums] while small amount of money are cut ” (#46/NGO/Professional Association). In sum, stakeholders saw the cuts to HP financing, workforce and services to be part of an austerity agenda to put reducing budget deficits above HP policy and practice. One feedback mechanism links state roles, governance structures and policy directions with state economic circumstances and budgetary constraints and another links the latter with cuts to HP financing, workforce and services (Fig.  2 ). These are both inhibiting feedback loops that act to balance or stabilise the system to an undesirable state. That is, the feedback loops illustrate how healthcare costs are constrained through cuts to HP financing, workforce and services.

State-local level

State policy directions resulting from the Review, SA Health’s Response and the SA Public Health Act emphasised leadership and health governance for HP at the local or regional levels (local governments and the Medicare Local in the case health system). Cuts to HP financing, workforce and services in the Local Health Network were unveiled alongside a redirection of resources to chronic disease management. Many interviewees reported being demoralised because of HP’s decline. Further, interviewees commonly discussed the consequences of a policy that implied cost shifting from state to local governments for HP with no new HP initiatives being planned. For example, one interviewee reported: “ I see a lot of cost and expenses so no one is looking to really take it [HP] on board because they know it’s like a poisoned chalice ” (#4/Local Council). Medicare Local interviewees pointed out that they mostly worked from a biomedical or clinical model and had no dedicated funding or workforce for HP. From this, an inhibiting feedback loop was identified (Fig.  2 ).

Conversely, the policy context was somewhat favourable for state and local leadership in governance for health (Table 4 ) through developing partnerships and intersectoral collaboration. All documents discussed partnerships and intersectoral collaboration to some extent and the SA Public Health Act offered clear policy directions for partnership development between the state government departments, local government and other organisations. Furthermore, there is a historical richness in SA documents (2003–2013) regarding policy direction for governance for health, particularly whole-of-government or Health in All Policies approaches. For example, the Adelaide Statement on Health in All Policies [ 58 ] emphasises the need for new governance structures and processes for partnerships in order to join up efforts to improve population health. The intent to build healthy social, economic and environmental policies underlying this document carried forward to the SA Public Health Act.

Stakeholders from all sectors reported that the SA Public Health Act was the key policy driver for HP in 2013. While it provided state and local support for leadership and governance for health, sectors and system elements were reported to be fragmented and the structures and processes for partnership development and collaboration were in their infancy. There was knowledge and a certain pride among many interviewees that the whole-of-government approach was in play within the state government; however, local governments appeared to have minimal involvement in the case health system. Building healthy public policy was explained by SA Health interviewees in terms of the SA Public Health Act being “ a real drive for Health in All Policies ” at the state and local government levels. Figure  2 illustrates the relationship between governance for health, state-local roles and whole-of-government approaches as facilitating feedback loops or virtuous cycles that are favourable for HP.

Our use of a causal loop diagram enabled us to identify the complex interplay of factors that affect HP and explain why the case study health system no longer supported HP. We found a complex picture with numerous interactions and feedback mechanisms represented in the causal loop diagram. The approach used helped us understand the patterns in system behaviour. Doing this makes it possible to identify potential opportunities to disrupt or slow down vicious feedback mechanisms and/or amplify those that are virtuous cycles. The majority of feedback loops in the causal loop diagram were vicious cycles that would need to be disrupted or changed for HP to thrive in the case study heath system. Changing even one feedback loop could change the emergent order of the system because system behaviour is a product of how the parts fit together and not how they act separately. Thus, feedback mechanisms can be seen as leverage points to strengthen systems [ 59 ] and this section highlights potential implications and links to other literature.

Disrupt vicious feedback mechanisms that inhibit HP

Improving HP policy and practice requires changing the feedback loop that inhibited the system goal of reducing health inequities. Strong leadership and governance could ensure that strategic policy frameworks targeting health inequities exist and facilitate policy coherence between levels of government [ 60 ] to address populations experiencing disadvantage, closing the gap in inequities and flattening the social gradient [ 61 ]. This concurs with Kickbusch and Gleicher’s view that “ the actions needed to improve health and reduce health inequities require new systems-based governance and delivery mechanisms that take account of interdependencies, complexity and the need for whole-of-government and whole-of-society co-production of population health ” ([ 62 ], p. 19).

Similarly, changing the inhibiting feedback loop with respect to community participation in HP is a notable opportunity. Active community participation is essential to effective HP policy and practice [ 2 , 6 , 7 , 63 , 64 , 65 ]. A virtuous cycle to encourage this could be established through strong policy statements that embed community participation in all HP planning, implementation and evaluation.

A feedback mechanism illustrates the leadership and governance challenges between the federal and state governments. This feedback mechanism is an important leverage point and needs to be disrupted, yet actions are highly political. As Bennett [ 66 ] notes, there is an ongoing blame game between the federal and state levels of government in Australia and this feedback mechanism may prove hard to change unless a window of opportunity develops where both federal and state governments have a strong desire to improve HP practice.

At the state level, the feedback mechanisms that illustrate the dominance of budgetary constraints is clearly a challenge to HP given the resultant cuts to financing and workforce. Without disrupting these feedback loops, a void in HP policy and practice will remain. These feedback loops point to the vulnerability of HP financing and lend support to calls for political will and leadership and governance structures to leverage dedicated funding for HP in Australia [ 67 ]. Duckett and Willcox state that “ health expenditure and health financing policies are rarely off the policy agenda ” ([ 68 ], p. 42). They further report that health expenditures in Australia are “ what would be expected given its GDP ” ([ 68 ], p. 42), opening debate about assertions such as those in the Review. It is beyond the scope of this paper to enter into debate; however, it appears illogical for policy-makers to target HP when public health as a whole in Australia represents only 2% of all health expenditures and could lead to savings in healthcare [ 68 , 69 ].

There are other feedback mechanisms at the state level that require change in order to strengthen HP. Addressing the vicious feedback mechanism associated with the lack of evidence of HP effectiveness will require leadership and health governance to allocate sufficient resources to implement and evaluate sustained and promising HP actions and strategies [ 25 , 70 ]. Importantly, a systems approach focused on addressing the broad political and structural determinants of health is needed [ 3 ]. Rutter et al. [ 1 ] state that a complex system model of evidence is necessary: “ Although it is important for public health policy to be guided by evidence, if this evidence predominantly supports individual-level interventions that have minimal reach and effect across populations, the benefits of being informed by the existing evidence base might be illusory ”. Beyond this, the abdication of leadership and health governance for HP did nothing to address this challenge and opposed calls for health systems to address the paucity of intervention research [ 71 ].

Turning to the state-local level, the inhibiting feedback loop that links the lack of leadership and health governance for reorienting health services to HP produced a policy vacuum. In other words, without disrupting the feedback loop, system elements, such as local governments and the Medicare Local, will stabilise around the policy vacuum and nothing will change in HP services because of the lack of financing and workforce building blocks. International documents have called for the health sector to reorient health services and lead HP since at least the Declaration of Alma Ata [ 6 ] in 1978 and, more recently, in the Rio Political Declaration on Social Determinants of Health [ 8 ] in 2011. However, this has been a long-standing challenge primarily because of entrenched factors including powerful vested interests and the dominance of the biomedical model [ 10 , 72 ].

Amplify virtuous feedback mechanisms that facilitate HP

Virtuous cycles were identified with respect to governance for health through partnerships and intersectoral collaboration and the key implication is the need to amplify these cycles. Leadership and governance for health through partnership development and intersectoral collaboration is critically important to HP policy and practice because of the complex interactions between factors that contribute to population health that are beyond the influence of any one sector in society [ 4 , 6 , 73 , 74 ]. Amplification of these feedback loops would strengthen the implementation of the SA Public Health Act and whole-of-government or Health in All Policies approach at both state and local government levels. Whole-of-government approaches to HP have been called for many years [ 2 , 75 ] and this research identifies great opportunity to build upon the rich history in SA. Legislation can be a powerful driver for collaboration and the SA Public Health Act provides a platform for aligning policies at state and local government levels simultaneously. There is a note of caution, however, as the lack of health governance for HP in reorienting health services, as discussed above, has the potential to have a negative impact upon governance for health [ 76 ]. That is, if the state and federal government do not champion HP within their respective health sectors, then why would other sectors and partners champion HP?

Limitations

When applying complex systems approaches it is necessary to define what is within the boundary of the system and what is out. This inevitably means that elements important to the system may be defined as outside of it [ 77 ]. In this study, inclusion of stakeholders from sectors and system elements, such as social service agencies and schools, might have offered different and useful perspectives.

The WHO framework was a useful foundation to study the case health system. However, the adapted definitions of the building blocks for a multisectoral health system for HP, being novel, would benefit from further applications and testing with policy-makers and practitioners to assess their value.

Creating causal loop diagrams in conjunction with group model building processes with stakeholders is called for in the literature [ 78 ]. Time and resource constraints did not permit this step. Although the research team undertook extensive discussion and achieved consensus on the causal loop diagram, facilitating a group model building process would have been preferable to not only gain their perspectives but to engage in discussion about implications, priority leverage points and actions to strengthen HP in the case health system. Thus, future research could build upon this research and use participatory systemic inquiry methods [ 79 ].

Leadership and governance for HP were found to be central factors that influenced HP policy and practice confirming findings from other jurisdictions around the world [ 62 ]. This study demonstrates its critical importance and adds urgency to the need for increased and strong advocacy for HP. The application of a complex systems approach to HP policy and practice addressed a gap in the literature. Our new methods have made visible the complex web of factors that influenced HP in an Australian multisectoral health system. Our approach was pioneering in that we combined health system building blocks and feedback mechanisms as leverage points [ 59 ]. Our causal loop diagram offered a picture of the broad array of interdependent facilitating and inhibiting factors that can be targeted to improve HP policy and practice.

Rutter H, et al. The need for a complex systems model of evidence for public health. Lancet. 2017;390(10112):2602–4.

Article   Google Scholar  

World Health Organization. The Ottawa Charter for Health Promotion. Geneva: WHO; 1986.

Google Scholar  

Baum F. The New Public Health. Third ed. Melbourne: Oxford University Press; 2016.

Commission on Social Determinants of Health. Closing the Gap in a Generation: Healthy Equity through Action in the Social Determinants of Health. Final Report of the Commission on Social Determinants of Health. Geneva: WHO; 2008.

World Health Organization. Putting our own House in Order: Examples of Health-system Action on Socially Determined Health Inequalities. Copenhagen: WHO Regional Office for Europe; 2010.

World Health Organization. Primary Health Care: Report of the International Conference on Primary Health Care. Alma Ata: WHO; 1978.

World Health Organization. Sundsvall Statement on Supportive Environments for Health. Geneva: WHO; 1991.

World Health Organization. Rio Political Declaration on Social Determinants of Health. Geneva: WHO; 2011.

Solar O, Irwin A. A conceptual framework for action on the social determinants of health. Geneva: WHO; 2010.

Ziglio E, Simpson S, Tsouros A. Health promotion and health systems: some unfinished business. Health Promot Int. 2011;26(suppl 2):ii216–25.

Hawe P. Lessons from complex interventions to improve health. Annu Rev Public Health. 2015;36:307–23.

Martin C, Sturmberg J. Perturbing ongoing conversations about systems and complexity in health services and systems. J Eval Clin Pract. 2009;15(3):549–52.

Atun R. Health systems, systems thinking and innovation. Health Policy Plan. 2012;27(suppl 4):iv4–8.

Begun J, Zimmerman B, Dooley K. In: Theory SM, Wyttenback M, editors. Health Care Organizations as Complex Adaptive Systems, in Advances in Health Care Organization. San Francisco: Jossey-Bass; 2003. p. 253–88.

Greenhalgh T, et al. Response to 'The appropriation of complexity theory in health care'. J Health Serv Res Policy. 2010;15(2):115–6.

Plsek, P. and T. Wilson, Complexity, leadership, and management in healthcare organisations. BMJ. 2001. 323(7315):746–749.

Pourbohloul B, Lieny M. Complex systems analysis: towards holistic approaches to health systems planning and policy. Bull World Health Organ. 2011;89:242.

Finegood D. The complex systems science of obesity. In: Cawley J, editor. Handbook of the Social Science of Obestiy. Oxfrod: Oxford University Press; 2011.

Hawe P, Shiell A, Riley T. Theorising interventions as events in systems. Am J Community Psychol. 2009;43(3–4):267–76.

Keshavarz N, et al. Schools as social complex adaptive systems: a new way to understand the challenges of introducing the health promoting schools concept. Soc Sci Med. 2010;70(10):1467–74.

Leischow S, Milstein B. Systems thinking and modeling for public health practice. Am J Public Health. 2006;96(3):403–5.

Shiell A, Hawe P, Gold L. Complex interventions or complex systems? Implications for health economic evaluation. BMJ. 2008;336:1281–3.

Trochim WM, et al. Practical challenges of systems thinking and modeling in public health. Am J Public Health. 2006;96(3):538–46.

Meadows D. Thinking in Systems: A Primer. In: Wright D, editor. White River Junction. VT: Chelsea Green Publishing Co; 2008.

Best A. Systems thinking and health promotion. Am J Health Promot. 2011;25(4):eix–ex.

Rickles D, Hawe P, Shiell A. A simple guide to chaos and complexity. J Epidemiol Community Health. 2007;61(11):933–7.

Luke D, Stamatakis K. Systems science methods in public health: dynamics, networks, and agents. Annu Rev Public Health. 2012;33:357–76.

Homer J, Hirsch G. System dynamics modeling for public health: background and opportunities. Am J Public Health. 2006;96(3):452–8.

Checkland P, Scholes J. Soft Systems Methodology in Action. New York: John Wiley & Sons, Inc; 1990.

Sturmberg J, Martin C. Handbook of Systems and Complexity in Health. New York: Springer; 2013.

Book   Google Scholar  

Sturmberg J, O'Halloran D, Martin C. Understanding health system reform – a complex adaptive systems perspective. J Eval Clin Pract. 2012;18(1):202–8.

Interaction, Feedback and Emergent Properties. http://maverickandboutique.com . Accessed 5 May 2018.

Jayasinghe S. Conceptualizing population health: from mechanistic thinking to complexity science. Emerg Themes Epidemiol. 2011;8:2.

World Health Organization. Everybody's Business: Strengthening Health Systems to Improve Health Outcomes. Geneva: WHO; 2007.

de Savigny D, Adam T. Systems Thinking for Health Systems Strengthening: An Introduction. Geneva: Alliance for Health Policy and Systems Research and the World Health Organization; 2009.

Allender S, et al. A community based systems diagram of obesity causes. PLoS One. 2015;10(7):e0129683.

Vandenbroeck P, Goossens J, Clemens M. Foresight Tackling obesities: Future choices - Building the obesity system map. London: Government Office for Science, UK Governments's Foresight Programme; 2007.

Alvaro C, et al. Moving Canadian governmental policies beyond a focus on individual lifestyle: some insights from complexity and critical theories. Health Promot Int. 2011;26(1):91–9.

Article   CAS   Google Scholar  

Plsek P, Greenhalgh T. The challenge of complexity in health care. Br Med J. 2001;323(7313):625–8.

Stake R. Case studies. In: Denzin N, Lincoln Y, editors. Handbook for Qualitative Research. Thousand Oaks: SAGE Publications; 2000.

Luck L, Jackson D, Usher K. Case study: a bridge across the paradigms. Nurs Inq. 2006;13(2):103–9.

Patton MQ. Qualitative Research and Evaluation Methods. Newbury Park: Sage; 2002.

Hsiu-Fang H, Shannon S. Three approaches to qualitative content analysis. Qual Health Res. 2005;15(9):1277–88.

Baum F, editor. Health for All: The South Australian Experience. Kent Town, South Australia: Wakefield Press; 1995.

World Health Organization and Australian Department of Community Services & Health. Adelaide Recommendations on Healthy Public Policy. Geneva: WHO; 1988.

Kickbusch I, Gleicher D. Governance for Health in the 21st century. Geneva: WHO; 2012.

Baugh Littlejohns, L., Baum, F., Lawless, A., & Freeman, T. 2018. Disappearing health system building blocks in the health promotion policy context in South Australia (2003-2013). Critical Public Health. Published online: 03 Jan 2018, https://doi.org/10.1080/09581596.2017.1418501

Kim H, Andersen D. Building confidence in causal maps generated from purposive text data: mapping transcripts of the Federal Reserve. Syst Dyn Rev. 2012;28(4):311–28.

Davidson EJ. Ascertaining causality in theory-based evaluation. N Dir Eval. 2000;87:17–26 (Fall).

Richardson G. Feedback Thought in Social Science and Systems Theory. Waltham: Pegasus Communications, Inc.; 1999.

Richardson G. An introduction to systems dynamics. Canberra: First Global Conference Research Integration and Implementation; 2013.

McCann W. Review of Non-Hospital Based Health Services: A Report by Warren McCann, Internal Consultant. Adelaide: Office of Public Employment and Review; 2012.

Government of South Australia. SA Health's Response to the Review of Non-Hospital Based Health Services. Adelaide: Department of Health; 2013.

Labonte R. Community, community development, and the forming of authentic partnerships. In: Minkler M, editor. Community Organizing and Community Building for Health. New Brunswick: Rutgers University Press; 2005. p. 82–96.

Council of Australian Governments. National Health Reform Agreement 2011. http://www.nationalplanningcycles.org/sites/default/files/country_docs/Australia/6552_nphc_1205.pdf . Accessed 20 Feb 2016.

Government of South Australia. Better Choices Better Health: Final Report of the South Australian Generational Health Review. Adelaide: Department of Health; 2003.

Government of South Australia. First Steps Forward: South Australia's Health Reform. Adelaide: Department of Health; 2003.

World Health Organization and Government of South Australia. Adelaide Statement on Health in all Policies: Moving Towards a Shared Governance for Health and Well Being. Adelaide: WHO; 2010.

Meadows D. Leverage Points: Places to Intervene in a System. Norwich, Vermont: Donella Meadows Institute; 1999.

Brown C, Harrison D. Governance for Health Equity in the WHO European Region. Copenhagen: WHO Regional Office for Europe; 2013.

Mantoura P, Morrison V. Policy Approaches to Reducing Health Inequalities. Montreal, Quebec: National Collaborating Centre for Healthy Public Policy; 2016.

Kickbusch I, Gleicher D. Smart Governance for Health and Well-being: The Evidence. Geneva: WHO; 2014.

World Health Organization. Jakarta Declaration on Leading Health Promotion into the 21st Century. Geneva: WHO; 1997.

World Health Organization. Mexico Ministerial Statement for the Promotion of Health. Geneva: WHO; 2000.

World Health Organization. Belfast Declaration for Healthy Cities: The Power of Local Action. Geneva: WHO; 2003.

Bennett C. Are we there yet? A journey of health reform in Australia. Med J Australia (MJA). 2013;199(4):251–5.

Harris A, Mortimer D. Funding illness prevention and health promotion in Australia: a way forward. Aust N Z Health Policy. 2009;6(1):25.

Duckett S, Willcox S. The Australian Healthcare System. Melbourne: Oxford University Press; 2011.

Oldenburg B, Harper T. Investing in the future: prevention a priority at last. Med J Aust. 2008;189(5):267–8.

PubMed   Google Scholar  

Foster-Fishman P, et al. Using methods that matter: the impact of reflection, dialogue, and voice. Am J Community Psychol. 2005;36(3/4):275–82.

Hawe P, Potvin L. What is population health intervention research? Can J Public Health. 2009;100(1):I8–I14.

Bambra C, Fox D, Scott-Samuel A. Towards a politics of health. Health Promot Int. 2005;20(2):193.

Commonwealth of Australia. Tackling Wicked Problems: A Public Policy Perspective. Canberra: Australian Public Service Commission; 2007.

Public Health Agency of Canada and World Health Organization. Health Equity through Intersectoral Action: An Analysis of 18 Country Case Studies. Ottawa: Public Health Agency of Canada; 2008.

Lalonde M. A new perspective on health of Canadians: a working document. Department of National Health and Welfare: Ottawa; 1974.

Marmot M, Allen J. Social determinants of health equity. Am J Public Health. 2014;104(S4):S517–9.

Ulrich W. In: Daellenbach H, Flood R, editors. Boundary critique, in The Informed Student Guide to Management Science. London: Thomson Learning; 2002.

Vennix J. Group model-building: tackling messy problems. Syst Dyn Rev. 1999;15(4):379–401.

Burns D. Participatory systemic inquiry. IDS Bull. 2012;43(3):88–100.

Download references

Acknowledgements

The authors would like to acknowledge the 53 stakeholders who gave of their time to participate in an interview for this research.

LBL received an International Postgraduate Research Award from Flinders University of South Australia from 2012 to 2015. At the time of writing (2018), LBL was supported by The Australian Prevention Partnership Centre through the National Health and Medical Research Council (NHMRC) Partnership Centre grant scheme (Grant ID: GNT9100001) with the Australian Government Department of Health, the NSW Ministry of Health, ACT Health and the HCF Research Foundation.

Availability of data and materials

The datasets generated and analysed during this study are not publicly available due to ethics approved processes for storage and retainment but are available from corresponding author upon reasonable request.

Author information

Authors and affiliations.

Menzies Centre for Health Policy, The Australian Prevention Partnership Centre, D17 Charles Perkins Centre, University of Sydney, Sydney, NSW, 2006, Australia

Lori Baugh Littlejohns

Southgate Institute for Health, Society and Equity, Flinders University, South Australia, Bedford Park, Australia

Fran Baum & Toby Freeman

Discipline of Speech Pathology, Flinders University, South Australia, Bedford Park, Australia

Angela Lawless

Research Affiliate, University of Sydney, Sydney, Australia

You can also search for this author in PubMed   Google Scholar

Contributions

This paper reports on LBL’s doctoral research. FB, AL and TF were supervisors of LBL’s PhD and were therefore instrumentally involved in all aspects of the research. LBL wrote the first draft and all authors contributed to writing the manuscript, commenting on drafts and approving the final version.

Corresponding author

Correspondence to Lori Baugh Littlejohns .

Ethics declarations

Ethics approval and consent to participate.

Ethics approval was obtained through Flinders University of South Australia, Social and Behavioral Research Ethics Committee (SBREC#583) and the Government of South Australia, Department of Health, Human Research Ethics Committee (HREC/12/SAH/44). All interviewees received a Research Information Sheet and completed a Consent Form.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Additional file

Additional file 1:.

Descriptions of causal links and feedback mechanisms. (DOCX 21 kb)

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License ( http://creativecommons.org/licenses/by/4.0/ ), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated.

Reprints and permissions

About this article

Cite this article.

Baugh Littlejohns, L., Baum, F., Lawless, A. et al. The value of a causal loop diagram in exploring the complex interplay of factors that influence health promotion in a multisectoral health system in Australia. Health Res Policy Sys 16 , 126 (2018). https://doi.org/10.1186/s12961-018-0394-x

Download citation

Received : 12 May 2018

Accepted : 20 November 2018

Published : 29 December 2018

DOI : https://doi.org/10.1186/s12961-018-0394-x

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• System building blocks
• Complex systems thinking

Health Research Policy and Systems

ISSN: 1478-4505

• Submission enquiries: Access here and click Contact Us
• General enquiries: [email protected]

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Int J Environ Res Public Health

Applications of Systems Science to Understand and Manage Multiple Influences within Children’s Environmental Health in Least Developed Countries: A Causal Loop Diagram Approach

Associated data.

Publicly available datasets were analyzed in this study. The data can be found at http://www.healthdata.org/gbd/2019 (accessed on 25 January 2021).

Least developed countries (LDCs) are home to over a billion people throughout Africa, Asia-Pacific, and the Caribbean. The people who live in LDCs represent just 13% of the global population but 40% of its growth rate. Characterised by low incomes and low education levels, high proportions of the population practising subsistence living, inadequate infrastructure, and lack of economic diversity and resilience, LDCs face serious health, environmental, social, and economic challenges. Many communities in LDCs have very limited access to adequate sanitation, safe water, and clean cooking fuel. LDCs are environmentally vulnerable; facing depletion of natural resources, the effects of unsustainable urbanization, and the impacts of climate change, leaving them unable to safeguard their children’s lifetime health and wellbeing. This paper reviews and describes the complexity of the causal relationships between children’s health and its environmental, social, and economic influences in LDCs using a causal loop diagram (CLD). The results identify some critical feedbacks between poverty, family size, population growth, children’s and adults’ health, inadequate water, sanitation and hygiene (WASH), air pollution, and education levels in LDCs and suggest leverage points for potential interventions. A CLD can also be a starting point for quantitative systems science approaches in the field, which can predict and compare the effects of interventions.

1. Introduction

Children can be considered as least developed countries’ most valuable resources, but in least developed countries (LDCs), their health is threatened by ecological degradation, pervasive inequalities, climate change, migration, and urbanisation [ 1 , 2 , 3 , 4 ]. There are currently 46 LDCs, and these are home to over a billion of the world’s people [ 5 ]. LDCs are diverse in geography, topography, and climates, and include mountainous countries such as Nepal, tropical Pacific Island countries, and arid landlocked countries such as Mali. However, they share common characteristics of low per capita income, an economy dominated by subsistence activities, limited manufacturing, and an undiversified production structure, low education levels, high fertility rates, and inadequate infrastructure [ 6 ]. In six LDCs, more than 70% of the population live below the international poverty line [ 7 ].

LDCs account for 13% of the world’s population, but with birth rates averaging 4.2 children per woman, they will account for 45% of the global population growth by 2050. Whilst two-thirds of people in LDCs still have rural subsistence lifestyles, urbanisation rates are higher than the global average [ 8 ], with urban migration driven by rural poverty and climate change [ 9 ]. Whilst global under-five mortality rates have decreased by 59% since 1990 [ 10 ], morbidity related to early-life environmental exposures is increasing [ 11 ]. Direct and indirect effects of detrimental environmental exposures in childhood often persist through adulthood [ 12 , 13 , 14 , 15 , 16 ], affecting people’s lifetime health and wellbeing and their ability to contribute economically to their community and society. The future economic potential of LDCs is thus directly linked to the health of their children.

Children’s environmental health (CEH) is the study of how environmental exposures in early life influence health and development in childhood and the entire lifespan [ 17 ]. In two landmark reports, Preventing Disease through Healthy Environments and Healthy Environments for Healthy People, the environment is defined as all the physical, chemical, and biological factors external to a person and all the related behaviours [ 18 , 19 ]. Such definitions, whilst they recognise that social determinants are closely linked to vulnerability to environment, may have over time contributed to a general perception of environmental health (EH) as a discipline that focuses on modifiable physical, chemical, and biological environmental determinants of health within constructs such as water, sanitation and hygiene (WASH), air pollution, chemical use, etc. In this paper, we use the terms “environment” or “environmental” to refer to the physical, chemical, and biological environment and refer to it as a domain whilst also considering social and economic domains and their influences on children’s health outcomes in LDCs. A domain is used in the non-specialist sense to mean a sphere of activity or knowledge. For the purposes of this paper, EH is therefore defined by the social, economic, and technological influences that link environmental conditions to human health.

The science of systems thinking studies how component parts in a system connect, react, and interact and helps us to see the forest as well as the trees. It increases our capability to recognise that cause and effect are non-linear, that the outcome of an event can influence the cause, and that perceived problems can often be symptoms of other problems [ 20 , 21 , 22 ]. In systems thinking terminology, children’s health in LDCs and the environmental, social, and economic factors that influence it are the CEH system, the product of the interactions between a set of parts that influence and feed back into one another to function as a whole. Whilst systems science has been used extensively in fields such as environmental science and business, the application of its techniques has been limited in EH [ 23 , 24 , 25 ]. Systems thinking can be contrasted with linear thinking, which assumes that a cause leads to an effect with no feedbacks and that factors are independent. A major shortcoming of linear thinking is that interventions can have unintended consequences; for instance, the use of agri-supporting products like fertilisers and biocides leads to resistant pests and weeds as well as excessive nutrient enrichment of receiving environments. This leads to a drastic decline of natural ecosystems, accumulations of toxins in food chains, and pathogen resistance. This can be attributed to policies and practices not considering the feedback loops in our system, which may change the outcomes of what we try to achieve.

The CEH system in LDCs is complex and multifaceted. One common characteristic of complex problems is that the root problem that is causing the symptoms is not always apparent at first inspection, nor is the solution obvious once the problem has been defined [ 26 ]. A systems thinking approach towards understanding the feedback loops may provide new insights and help to determine root causes.

A causal loop diagram (CLD) is a qualitative systems science tool that shows the relationships between a set of variables (factors liable to change) operating in a system. It is a powerful tool for identifying the non-linear feedback loops that operate in the system to amplify or balance outcomes. It can help stakeholders to converge on a shared mental model of a system, a set of beliefs, values, and assumptions that underly why things work as they do [ 27 ]. This shared understanding about how something works and what is important can be used to enhance policy setting and decision-making. A CLD can also be the foundation for quantitative modelling techniques such as dynamic and agent-based modelling [ 28 ].

Many studies have reported on children’s health and the environment, but it appears that none have used a CLD (also known as an influence diagram) [ 28 ] to represent the big picture, the underlying feedback mechanisms and potential key leverage points. The objectives of this paper are firstly to represent the major feedback loops that link children’s health with the environmental, social, and economic domains in LDCs and secondly to seek insights into potential leverage points and interventions.

In the Results section, the CEH system is represented by a CLD that contains four interlinked sections; children’s health outcomes and the variables that influence them grouped into environmental, social, and economic domains. These domains align with the three pillars of sustainable development on which the UN Sustainable Development Goals (SDGs) are based; environmental, social, and economic [ 29 , 30 ].

2. Methodology

We used the Institute for Health Metrics and Evaluation (IHME) Global Burden of Disease results tool for 2019 [ 31 ] to identify the most significant estimated causes of child mortality and morbidity for countries in the World Bank’s least developed countries category and summarised the findings. We conducted a narrative review [ 32 ] of papers retrieved following a systematic search of current and past literature. The results were summarised, in a table for children’s health outcomes and three tables for influencing factors from environmental, social, and economic domains. These tables were used as the basis for constructing the CLD. The most important loops, based on their relative contribution to child morbidity and mortality, were then further investigated using the CLD and potential leverage points for solutions identified in Results Section 3.5 .

2.1. Literature Review

Information on CEH in LDCs is found in reports from studies focusing specifically on LDCs, also in low- and middle-income countries (LMIC) and global health studies. Reports and scientific papers published from January 2000 through December 2020 were searched, screened, and reviewed according to their relevance, based on the primary and secondary key terms.

Key search terms were developed to ensure that potentially relevant studies with content relating to children’s health in LDCs were identified. First, searches were run using a composite primary search term “least developed country” OR “least developed countries” OR “LDC” OR “LDCs” OR “low-income countries” OR “low- and middle-income countries” OR “LMIC” with secondary search terms of “children’s health”, “environmental health”, “children’s environmental health”, and “CEH”. The composite primary search term was next used with secondary search terms taken from the causes of child mortality and morbidity identified in Table 1 , e.g., “respiratory”. A further search was run using the term “global children’s health”. Databases searched were PubMed, Google Scholar, and the World Health Organization (WHO), United Nations International Children’s Emergency Fund (UNICEF), United Nations Environment Programme (UNEP), and World Bank publications databases. The same series of search terms were then used in the Google search engine to identify additional grey literature. For all sources, the first 100 results were checked, and a relevancy assessment approach [ 32 ] was used. References from identified publications were also searched.

Child mortality and morbidity for all least developed countries (LDCs) [ 31 ].

1 children aged under 15 2 children aged under 5 3 Years Lived with Disability 4 Neglected Tropical Diseases. Causes of mortality/morbidity from Level 2 ICD codes [ 31 ]: Morbidity is measured in years lived with disability. Neonatal includes maternal/neonatal disorders. Other infectious diseases include meningitis, measles. Other NCDs include congenital birth defects and sudden infant death syndrome. Enteric diseases include diarrhoea and typhoid. NTDs include dengue fever, yaws, trachoma, helminths including hookworm, ascariasis, and trichuriasis. Skin diseases include scabies and fungal skin diseases. Respiratory includes upper and lower respiratory infections, tuberculosis, and chronic respiratory disease. Mental disorders include intellectual disability. Italics: denote disease groups excluded from further analysis.

2.2. Causal Loop Diagram Principles

A CLD consists of variables and cause-effect links (also known as influencing links) that connect to form causal loops, also known as feedback loops. Causal loops are either reinforcing (vicious or virtuous circles) or balancing, where self-correction occurs within the system. Every causal loop tells a story that links cause and effect through feedback, e.g.,

• reinforcing—a dengue fever epidemic where the number of infected mosquitos drives up the number of infected humans, which in turn increases the number of infected mosquitos;
• balancing—where sweating is initiated in response to heat to regulate human body temperature.

The variables that represent the causal influences in the CLD are linked by directional arrows, which represent causal associations. Associations are either:

• reinforcing—denoted by a +, in which an increase in a variable causes an increase in the variable it influences and vice versa, or e.g., internal air pollution increases respiratory disease;
• opposing—denoted by a −, when an increase in a variable causes a decrease in the variable it influences and vice versa, e.g., a clean water supply decreases WASH-related disease [ 28 ].

An even number of negative polarities in a loop denotes a reinforcing loop; an odd number, a balancing loop. Hash marks on the connector arrows denote delays between cause and effect. Variables in a CLD are either endogenous, both influencing and influenced by other variables within the CLD, or exogenous, influencing but not being influenced [ 22 ]. A further explanation of the notation of causal loop diagrams with examples is given in Supplementary Material S1 .

2.3. Table and Causal Loop Diagram Creation

Children’s health outcomes and their influencers as identified by the literature review were all designated as variables for the CLD and grouped into four sections: children’s health outcomes, environmental, social, and economic domains. A table was created for each section, and each variable was mapped to:

• variables that it directly influences;
• variables that it is directly influenced by.

The mapping process identified two exogenous variables, remoteness and climate change, which are discussed in Section 3.3.2 .

These tables were used as the basis for constructing the CLD, which was created using Stella Architect software (iseesystems.com; Version 2.0.3). The CLD is a visual representation of the mapping shown in the tables, with many loops, both reinforcing and balancing, identified. The CLD and tables were then reviewed and refined using an iterative process. The most important loops in the CLD, based on their relative contribution to child morbidity and mortality as identified in Table 1 , were then further investigated and potential leverage points for solutions were identified.

3.1. Child Mortality and Morbidity in LDCs

Table 1 shows the disease groups that contribute the most to child mortality and lifetime morbidity as measured by deaths and years lived with disability (YLD) for diseases originating in childhood. The health data are categorised by Level 2 ICD codes [ 31 ]. Childhood is defined as ages 0–14 (inclusive) in line with the definition used in the SDGs [ 33 ].

The most common cause of childhood death in LDCs, in common with global rankings, is neonatal disorders, reflecting the high-risk 28-day post-natal period. Enteric disease followed by respiratory disease and neglected tropical diseases (NTDs) are the next highest ranked. The largest contributor to lifetime morbidity is nutrition-related disease, followed by skin and subcutaneous disease, which is prevalent in LDCs. The rankings are similar for under-five mortality and morbidity, with the largest discrepancy between under-fives and under-fifteens in lifetime morbidity caused by mental disorders, which are likely to be undiagnosed in under-fives.

3.2. Exclusions from CLD Scope

Table 1 shows that disease groups for HIV/AIDS, sense organ diseases, and neoplasms contributed 3% or less to only one of mortality or morbidity. They were excluded from the scope of the CLD because of their relatively small contribution to child health outcomes relative to other disease groups. Disease groups for other NCDs and mental disorders were excluded from further analysis because the literature review did not yield sufficient evidence for a relationship between the disease and the environmental, social, and economic domains to warrant their inclusion in the CLD.

3.3. Influencing Linkages

Table 2 , Table 3 , Table 4 and Table 5 show the linkages of variables with their influencers and influences. The tables show the cause and health effect pathways (e.g., internal air pollution leads to childhood respiratory disease) and also show links between variables such as economic development, poverty, infrastructure, clean water access, morbidity, and poverty, which are then represented visually in the CLD shown in Figure 1 . Table notation and logic are as follows:

• Variables are described in the shortest form possible, e.g., vehicles means number of vehicles, family size means number of people in the biological family, and clean water means the availability of a clean water supply;
• Polarities of links are shown, e.g., open defaecation increases WASH-related disease;
• Each endogenous variable influences other variables in the tables and is in turn influenced by other variables. As an example, WASH-related disease is the variable in the first row of Table 2 . It appears as an influencer (reinforcing or positive) of malnutrition/stunting in the second row of Table 2 and is shown as being influenced by open defaecation in Table 3 ;
• The relationship between the variable and the links appearing in the “influenced by” column is summarised in the table text with supporting references. Relationships for items in the influences column are summarised when they appear in the variable column, usually in another table;
• Only direct influencers are shown, e.g., the influence of improved sanitation on WASH-related disease is not shown in Table 2 , as improved sanitation is categorised as an influencer of open defaecation and can be found in Table 3 .

Variables in environmental domain with influencing links.

* Exogenous variable.

Variables in social domain with influencing links.

Variables in economic domain with influencing links.

Causal loop diagram for children’s environmental health system.

Children’s health outcome variables with influencing links.

1 Internal Air Pollution 2 Ambient Air Pollution.

3.3.1. Health Outcomes

Children’s health outcomes in LDCs are grouped in Table 2 according to the main environmental influences; thus, the WASH-related disease category includes enteric disease, skin disease, and parasitic diseases. Influences that generally improve child health outcomes, e.g., access to health services, are linked to a consolidated child morbidity/mortality outcome. Adult morbidity and premature mortality are also influenced by child morbidity.

3.3.2. Environmental Domain

The environment in LDCs is heterogeneous, with wide variation in geography, climate, and population density. LDCs contain both urban population concentrations and remote rural settlements. Endogenous variables in this domain that influence children’s health in LDCs are shown in Table 3 . Two exogenous variables were identified in the environment: climate change and remoteness. They can be seen in the “influenced by” column of Table 3 . In practical terms, exogenous variables cannot be influenced by other variables in the model. This means that in the context of this model, our focus for interventions should lie elsewhere.

• LDCs contribute only 0.5% of the annual carbon dioxide emissions that are driving climate change, producing 0.17 million kt of a global annual total of 34 m kt [ 57 ]. Climate change impacts are specific to individual countries and regions, but all LDCs are vulnerable to the effects of climate change, manifested in rising temperatures, changing landscapes, and increased magnitude and frequency of natural disasters [ 52 ]. Climate change has thus been treated as an exogenous variable, influencing but not being influenced by the other variables in the CLD [ 22 ].
• Remoteness is part of the economic vulnerability index for LDCs, calculated as an indicator of distance from world markets [ 59 ], and is a structural obstacle to the development of adequate infrastructure. Whilst an LDC can improve its infrastructure and services, it cannot change its geographical remoteness, which is thus an exogenous variable.

3.3.3. Social Domain

The social domain encompasses the social relationships and cultural constructs within which people function and interact. Components of the social domain include cultural and religious beliefs and practices, family structures, social and power relations, and inequalities. Social domain components function at multiple scales: households, extended kin networks, communities, and cities. Social domains are dynamic and change over time [ 72 ]. Variables in the social domain that influence children’s health in LDCs are family size, education levels, and culture, as shown in Table 4 .

The term “cultural norms” has been used in the tables to denote the set of beliefs and practices that influence many aspects of life in LDCs. Some examples are food preparation and cooking practices, acceptability of smoking, and views on the optimum number of children for women. Cultural norms may support or be detrimental to children’s health. They will change over time, driven by influences such as education. In Table 3 , cultural norms are represented as a force that resists and slows down positive change. In general, higher levels of adult education reduce the strength of detrimental habits and taboos and improve cultivating of health-supporting behaviours, for instance, by optimising sanitation and hygiene practices if water infrastructure and services are available. Without education, low adoption or declines in usage occur as communities revert to their traditions of open defaecation [ 23 ]. Cultural norms can also be seen influencing variables in both the environmental and economic domains shown in Table 3 and Table 5 .

3.3.4. Economic Domain

Children’s health is directly affected by their economic status, with clear evidence of influencing links between economic status and EH assets such as clean water, sanitation, clean fuel, and electricity [ 80 , 81 ]. Table 5 shows the variables in the economic domain that influence children’s health in LDCs, including the availability of health services and urban migration driven by rural poverty.

3.4. Overall Causal Loop Diagram

The CLD shown in Figure 1 represents the non-linear causal relationships in the children’s health system in LDCs based on the relationships identified in the literature review. It has been structured into four sections: health (grey), environment (green), social domain (pink), and economic domain (blue). Where variables can be categorised in more than one section, e.g., overcrowding, which could be viewed as both an environment and an economic variable, the colours overlap. Many causal loops can be identified, reflecting the complexity of the system. All of the loops that include children’s health outcomes include variables in at least two other sections, showing their interconnectedness. The majority of the many loops in this diagram show reinforcing cycles, causing accelerated growth or decline. Interventions discussed later can be used to change the direction of these causal loops.

The balancing and reinforcing loops considered to be most important are shown on the CLD, but its complexity makes it hard to trace the connections, so they are split out and discussed in Section 3.4 . Connections that are discussed are shown in colour or in black, e.g., blue from clean water to WASH-related disease. All others are shown in dark grey.

The term cultural norms as described in Section 3.3.3 describes a very broad range of human behaviours and customs. The table entries and the CLD show links from cultural norms to variables not only in the social domain but also to variables in the environmental and the economic domain. The CLD represents the links from cultural norms to variables in the environmental, social, and economic domains with dotted connectors to recognise that they are generalised and may not apply in all LDCs.

3.5. Analysis of Causal Loops

Areas of the CLD for more detailed analysis were chosen by referencing Table 1 and selecting the loops that include the disease groups, which cause the largest percentages of child mortality and lifelong morbidity. A loop focusing on the effects of population growth on children’s health was also added after linkages were noted in Section 3.5.1 and Section 3.5.5 . Reinforcing and balancing feedback loops are highlighted and discussed. The loops can all be traced in Figure 1 , but some positions have been rearranged for ease of reading.

3.5.1. Nutritional Deficiency Loops

The most significant cause of lifetime morbidity from diseases contracted in childhood is nutritional deficiencies, primarily including protein/energy malnutrition, with 26.6% of all morbidity caused by this disease group [ 31 ]. Loop R1, shown in red in Figure 2 , shows the reinforcing cycle of poverty, which reduces a family’s ability to provide adequate child nutrition. A reduction in adequate nutrition leads, with a cumulative and delayed effect, to malnutrition and/or stunting, which in its turn reinforces child morbidity [ 2 , 18 , 43 , 44 ]. As malnourished children develop into adults, the disease burden established in childhood remains with them, leading to adult morbidity and decreased life expectancy. This decreases the adult’s capacity to contribute economically to the family, reinforcing poverty and completing the loop. Note that this loop contains two negative and three positive polarities and is a reinforcing loop because the negatives counteract each other. Adequate child nutrition is also diminished by depletion of natural resources, particularly in rural settings where foraging or hunting provides food sources.

Loops influencing nutritional disease.

Another important loop exhibiting reinforcing behaviour is R2 (purple/red), which shows the effect of family size on malnutrition [ 45 ]. The new loop connectors are shown in purple; poverty reduces access to modern family planning, increasing family size, which decreases the likelihood of adequate child nutrition; the loop is then completed by tracing the red arrows around the common linkage through malnutrition/stunting > child morbidity > adult morbidity > poverty.

Loop R3 (dark green/red) shows how malnutrition/stunting reduces children’s education through impaired cognitive ability and school absences [ 73 ] which, as children grow into adults, has a detrimental effect on levels of adult education. Increasing adult education supports adequate child nutrition and vice versa. The reinforcing loop is completed by the link back to malnutrition/stunting shown by the red arrow. Loop R4 (green/dark green) directly connects adult educational attainment to children’s educational attainment. Better educated adults are more likely to value and prioritise the education of their children. If child education levels increase, a virtuous circle of education level improvement is created; if they decrease, the reverse happens. Loop R5 (starting with orange dotted connector) shows how cultural norms, which are generally challenged by increasing levels of education, reinforce both the expectation of and desire for larger families [ 7 ]. The loop continues with a connection to adequate child nutrition (purple) and can be traced through malnutrition/stunting to child education and adult education. In some LDCs, a cultural norm is a lack of food priority for children [ 77 ]; this too is challenged by education. Potential opportunities to reverse negative reinforcing loops are child nutrition and education interventions.

3.5.2. WASH-Related Disease Loops

Enteric disease, including diarrhoeal disease and typhoid, is estimated to be 95% attributable to inadequate WASH in LDCs [ 31 , 89 ]. Skin diseases, responsible for almost no child mortality but an estimated 9.9% of lifetime morbidity [ 31 ], are influenced by inadequate WASH and overcrowding. Loop R6 (blue/red) in Figure 3 shows the reinforcing loop connecting water access, clean water, and WASH-related disease, which includes enteric disease and skin diseases [ 3 , 18 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 ]. If an LDC’s government or external organisations do not provide a service, people and communities living in poverty, whether in informal settlements or a rural setting, do not have the resources to improve their own water supplies. Sufficient accessible clean water reduces all WASH-related disease. Improved water access is also a prerequisite for most improved sanitation services [ 60 , 61 , 62 ]. Loop R7 (brown joining red) shows the dependency of improved sanitation on improved water supplies. Improved sanitation is a prerequisite for the reduction or elimination of open defaecation, but poverty reduces the ability of communities to maintain sanitation facilities, and in some LDCs, there are powerful cultural traditions that impede a transition away from open defaecation. Education is required to support understanding of the benefits, as illustrated by the dotted lines linking adult education with cultural norms and open defaecation. Exposed faeces support the spread of pathogens, which increase WASH-related disease.

Loops influencing WASH-related disease.

Loop R8 (pink joining red) shows how poverty reduces access to soap through a lack of resources to purchase it [ 18 ]. Adequate hygiene, including hand hygiene after defaecation and before food preparation, requires soap but also improved water access as in most cases, a natural clean water source is not close enough to the household to provide adequate water for hygiene purposes. In common with the loops shown in Figure 2 , enteric disease increases child morbidity as repeated acute illness leads to chronic disease and later, either directly or through increasing malnutrition and stunting, to adult morbidity. Leverage points to reverse negative reinforcing loops are water and sanitation interventions combined with education, both general and specific.

3.5.3. Air Pollution-Related Disease Loops

Figure 4 shows loop R9 (dark red joining red), a reinforcing loop linking the use of biomass fuel and respiratory disease, responsible for 14.7% of childhood mortality and 6.7% of morbidity in LDCs It illustrates how poverty reduces a household’s ability to acquire clean fuel and equipment with which to use it [ 7 ], forcing households to create internal air pollution through the use of biomass fuel, which is often available with no financial outlay to rural families (although it depletes natural resources, which in itself has consequences for the health of the environment and for the ability of the environment to provide for children’s nutrition). The respiratory disease burden of both mortality and morbidity with its roots in childhood is possibly one of the most difficult to address; long lead times mean that policymakers do not necessarily relate adult health consequences to lung damage sustained in childhood. A move away from biomass fuel requires not only clean fuel availability and affordability but also a willingness to embrace new ways of cooking, supported by an understanding of the health implications [ 87 ]. As with WASH-related disease loops, adult morbidity and premature mortality reinforce poverty.

Loops influencing air pollution-related disease.

Loop B1 (dark blue joining dark red then red) describes a balancing loop between poverty and vehicle pollution. If poverty increases, the number of vehicles reduces. The converse of this is that reducing poverty increases vehicle pollution and increases child and adult respiratory disease [ 88 ]. In this case, the positive health effects of poverty reduction are offset by a negative health effect of unsustainable development. The assumption here is that vehicles produce pollution; in LDC, vehicles are likely to be old and heavily polluting and are often exported from higher-income countries whose stricter regulations they no longer meet. Leverage points are air pollution reduction and clean cooking interventions combined with education.

3.5.4. Vector-Related Disease and Skin Disease Loops

Vector-related disease accounts for an estimated 10% of child mortality and 9% of lifetime morbidity in LDCs [ 31 ]. Dengue fever is a growing urban problem whilst malaria still claims the greatest number of children’s lives in both rural and urban settings, particularly in African LDCs [ 18 , 51 ]. Loop R10 in Figure 5 (dark purple joining red) shows links between poverty, urban migration, overcrowding, numbers of vectors due to inadequate household waste management services, and vector-related disease. Overcrowding, particularly in informal settlements, is influenced in many LDCs by cultural obligations to house extended family and also contributes to skin disease transmission, WASH-related disease, and the spread of infectious diseases [ 54 ]. Loop R11 (brown joining dark purple and red) shows how poverty negatively influences improved sanitation, linking to open defaecation and increased numbers of vectors. Leverage points are in improving the built environment, waste management services and reduction of vector habitat as well as the WASH-related leverage points discussed earlier.

Loops influencing vector-related disease.

3.5.5. Neonatal Disease Loops

Loop R12 in Figure 6 is depicted in purple, linking poverty to family planning availability, family size, maternal health, neonatal disorders, and child morbidity. The loop continues in red through to adult morbidity and poverty. Neonatal disorders cause 28.7% of child mortality [ 31 ], with an estimated 20% attributed to the environment [ 89 ], but this percentage does not include the influence of maternal health and nutrition and access to health services, so one could argue that the total environmental attribution should be higher. CEH discussions do not generally include the influence of family size and maternal health on neonatal disorders or children’s health in general [ 45 ]; we believe that the CLD makes a case for doing so. A shorter reinforcing loop R13 also associates poverty with lower access to family planning availability, thus supporting increased family sizes.

Loops influencing neonatal disease.

Balancing loop B2 (dark purple/purple) depicts poverty driving migration from rural to urban or peri-urban areas. Access to modern family planning services is improved by moving to an urban area, with large variations between LDCs. This access acts to reduce family sizes, which helps to lift families out of poverty. However, urban migration in LDCs is increasingly forced by climate change and population growth, which leaves families unable to survive in rural areas, and another small reinforcing loop, R14, linking urban migration to increasing poverty, shows poverty and child malnutrition increasing in urban/peri-urban areas [ 83 ], counteracting the balancing effects of smaller families.

3.5.6. Population Growth Loops

Figure 7 shows that loop R15, starting in purple and moving through black and red, links household poverty with lower family planning availability and population growth. A larger population puts more demands on infrastructure, reducing the resources available per capita, whether for roads, improved water supplies, or health services. Improved health services reduce child morbidity and vice versa. Child morbidity leads to adult morbidity, which reinforces poverty, completing the loop. This vicious circle, which connects the situation of individual families to broad population growth and its socioeconomic impacts, and the situation in LDCs to the broader economic issues, is not, as far as we are aware, discussed in the CEH context even though lower per capita resources clearly have the potential to negatively affect children’s health outcomes.

Loop influencing population growth.

3.6. Leverage Points

A key environmental determinant of health is sufficient and available clean water supplied close to the household to deliver not only safe potable water but also water for hygiene and sanitation purposes. This is well known, but Figure 2 reinforces the need for this leverage point to be supported by education, both general and specific, to support the uptake of water and sanitation interventions, to support cultural change where needed to overcome traditions and taboos that work against uptake, and to support the maintenance of WASH infrastructure. Reductions in both WASH-related and vector-borne diseases are the potential results. An estimated 60% of WASH benefits come from the elimination of open defaecation in communities [ 60 ], but this can be hard to achieve because of the many prerequisites.

Air pollution, both internal air pollution (IAP) and ambient air pollution (AAP), is an environmental determinant of health in LDCs, and interventions that support the uptake of clean cooking are leverage points for children’s lifetime health but need to be supported by education and community engagement. The relative cost of fuel is important; in rural areas, as long as biomass fuel is free, interventions that include education and a supply of clean cookstoves are unlikely to deliver improvements to households living in poverty without ongoing financial support. Leverage points for the reduction of AAP, particularly in urban areas, are vehicle emission reduction and regulation and household waste management interventions to reduce burning.

The growth of urban and peri-urban informal settlements, reinforced by poverty-related and climate change-driven migration, threatens children’s health through overcrowding and lack of infrastructure. Interventions that improve services, including waste management to informal settlements, are leverage points for children’s health.

All the major causal loops that include child morbidity connect through to adult morbidity and reinforcement of poverty. It follows that poverty reduction in LDCs will improve health outcomes unless it is countered by negative health effects such as increases in air pollution. However, continuing population growth, which effectively reduces infrastructure and health services available at an individual level and puts pressure on natural resources, reinforces poverty. Leverage points to address population growth are interventions in education and family planning. One successful intervention in LDCs has been immunisation programmes; these have delivered significant reductions in child mortality globally but have only increased population growth in LDCs.

Family planning availability has not been considered in the CEH field as one of the tools for improving children’s health in LDCs, but the impacts on maternal health, neonatal outcomes, children’s nutritional health, and access to health services are insight from this CLD and support a case for its inclusion as a leverage point.

4. Discussion

4.1. application of the cld.

The CLD demonstrated in this paper is a system with many reinforcing causal loops that explain current behaviours of the CEH system through the interaction of a selected suite of endogenous variables [ 28 ]. Building a CLD based primarily on literature-derived variables is unusual; a more common approach is to collaboratively build or modify a CLD from participatory discussions with stakeholders [ 90 ]. This can either be done in community settings, with groups of policymakers/influencers or both, and can be a powerful tool for effective engagement in LDCs, lowering the risk of policy failure due to lack of cultural understanding [ 91 ].

Nevertheless, our work creates a better understanding of the often unsighted influencing loops that connect the environmental, social, and economic domains, highlighting and reinforcing known leverage points such as the need for education, cultural awareness, and community engagement if interventions are to yield positive results. It also shows how compounding delays reduce awareness of health outcomes; for example, the time delay between children’s respiratory clinic visits and adult respiratory-related mortality does not result in air quality policy interventions in LDCs because more immediate concerns dictate policy priorities.

These insights, particularly the links between population growth and children’s health, give us a big picture understanding of the issues facing LDCs and emphasise the need for support from more developed countries. There are many agencies active in LDCs delivering individual aid-based interventions, but the CLD highlights the need for collaboration across sectors to avoid suboptimal outcomes or unintended consequences. It also shows how LDCs, with external support, need to address poverty as a structural determinant of health, possibly in the context of the SDGs.

The selection of causal loops for further discussion in Section 3 does not mean that other disease groups, e.g., injuries, are unimportant and should not be analysed. One limitation of a CLD is that it is qualitative only; the choice of loops to discuss was based on quantitative information from Table 1 . Similarly, disease groups such as mental health without sufficient literature-based evidence for links to the environmental, social, and economic domains in LDCs should not be ignored and point to a research gap.

The discussion of leverage points in Section 3.5 shows some examples of how interventions could change health outcomes. From a CLD perspective, interventions can change the polarity of reinforcing loops from detrimental to beneficial outcomes. This CLD is necessarily high-level; using the tool to focus on a specific EH problem, environment, or socioeconomic setting enables greater depth and more specific insights, which can guide decision-making about policy setting and practice.

4.2. Potential Application for Systems Science in CEH

The complex problem of poor CEH in LDCs cannot be solved solely through linear cause and effect approaches to problem-solving, such as the provision of water infrastructure to reduce WASH-related disease. A systems science approach has real potential to support decision-making in research as well as policy setting and practice [ 26 ]. In a quantitative systems science approach, dynamic simulation models are developed from a CLD and explicitly quantify relationships between influencing variables and describe their rates of change over time. Such models can then be used to simulate and compare the impacts of different policies and practices over time and to identify potential feedbacks, limitations, and inhibitors. Specifically, interventions can be simulated, and the levels of investment required to change the polarity of reinforcing loops and the time delays can be estimated in combination with a range of different assumptions.

Using these methods, we could, for example, estimate how child mortality and morbidity over the last 30 years will change into the future considering multiple and heterogenous influence variables and not, for instance, just the size of the investment or the potential uptake by the community. Similarly, we could extend the modelling to link adverse child health outcomes in LDCs to subsequent adult health outcomes [ 92 ], gaining a greater understanding of how morbidity with its origins in childhood is influencing adult health expectations. Many more influences can be defined and modelled. For example, the extent to which reduced child mortality contributes to population growth, which then reinforces poverty. Aid-funded interventions in infrastructure, for example, health service facilities or improved water supplies, will have a positive impact on children’s health outcomes, but we can now model how population growth and urban densification stretch these resources to the limit when considering the influencers from all domains.

To be useful to decision-makers and give meaningful results, a model would need to be built for a specific country or area and stratified into urban and rural segments. Lack of data in LDCs may make this task more difficult, but systems science can overcome such limitations to a degree by incorporating expert opinion and data from similar domains (e.g., other LDCs and proxies) into models.

5. Conclusions

Understanding is merely a starting point; delivering scientific recommendations that lead to action and sustained progress is surely the most important goal of CEH research. Systems science, currently underutilised in this field, can make an important contribution in all CEH settings, including LDCs. A CLD is a powerful tool in its own right for exploring and recognising the many interconnections between the environmental, social, and economic domains and their influence on children’s health outcomes and creating a shared understanding, and is also a first step in a quantitative dynamic modelling process. Our CLD shows the need to include a policy on population growth, family size, and family planning availability as an influencer of children’s health.

CEH most often focuses on environmental influences on health, and our CLD approach demonstrates how these should be augmented with social and economic influences and shows the impacts of poverty, low levels of education, and inadequate infrastructure on children’s health in LDCs.

Acknowledgments

This research was supported by the University of Queensland.

Supplementary Materials

The following are available online at https://www.mdpi.com/1660-4601/18/6/3010/s1 , Supplementary S1: CLD creation and terminology.

Author Contributions

Conceptualization, C.F.B.; methodology, C.F.B. and P.J.; writing—original draft preparation, C.F.B.; writing—review and editing, C.F.B. and P.J. All authors have read and agreed to the published version of the manuscript.

This research received no external funding.

Institutional Review Board Statement

Ethical review and approval were waived for this study due to research only using existing collections of data that contain only non-identifiable data about human beings.

Informed Consent Statement

Not applicable.

Data Availability Statement

Conflicts of interest.

The authors declare no conflict of interest.

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.