Development and evaluation of system dynamics methods to engage with policy makers on the prevention and control of diabetes in a middle income region

Diabetes is a serious and growing problem globally. Evidence suggests diabetes disproportionately affects people in low- and middle-income countries, where 80% of people with diabetes live, both in terms of numbers of people affected as well as outcomes and deaths. Diabetes affects between 10 and 20% of the adult population in the Caribbean region with deaths due to diabetes estimated to be 35% higher than in the neighbouring United States. Not only are prevalence and mortality a large burden but also rates of complications such as lower-limb amputation are also high. Much of the high burden of diabetes can be attributed to major risk factors such as physical inactivity and obesity. Health systems with limited resources in these developing countries are struggling to meet the growing epidemic.
There is a strong political will in the region to tackle diabetes and other non-communicable diseases (NCDs). In 2007, the Heads of Government of the Caribbean Community put forth the Port of Spain Declaration (POSD) on NCDs, definitively challenging the high burden of these diseases in the region and pledging action through policies to strengthen prevention and treatment. This laid the groundwork for a global political movement to recognise NCDs on the public health agenda culminating in the United Nations High Level (UNHLM) meeting on NCDs in September 2011. Both the POSD and UNHLM strongly emphasise the importance of policy measures for reducing NCD risk factors and put forth policies and targets.

However, evidence on how to achieve a reduction in overweight/obesity and physical inactivity and subsequently reducing NCDs at the population level is scarce, particularly in developing countries. While the risk factors and determinants of NCDs like diabetes are well studied and established, research has not been able to conclusively demonstrate real-world interventions that can reduce their popburden or change the course of the epidemic.

Systems science, which combines multiple factors and complex interrelationships, may offer a solution to evaluating and testing policies for diabetes reduction. It does this by explicitly taking into account system behaviour that is non-linear and complex, with feedback loops and time delays. Within systems science, system dynamics modelling is a methodology incorporating input from experts and stakeholders and combining that with quantitative research evidence to produce a map of a system with the ability to simulate outcomes by changing parameters. The approach has been used effectively in a wide variety of fields including engineering, agriculture, energy planning, business dynamics, and health including diabetes. However, few models have been developed for use in middle-income countries.

This study will be the first to explicitly develop a model for diabetes in developing countries, drawing from work successfully conducted by the Centers for Disease Control in the United States and their model for diabetes. The study will apply the rigorous qualitative methods required by interviewing stakeholders, experts and policy makers in the region as well as gathering evidence from research published on risk factors, outcomes, and health system performance for diabetes in the Caribbean. The study will use the developed model to engage stakeholders and policy makers in time for the on going evaluation of the POSD as a tool for effective policy planning. It will also evaluate the utility of this method in the region in engaging policy makers to think in terms of systems and with long time horizons. The results of this development study will be used to build a larger model incorporating economics and costs, which can then be adapted and used in other low- and middle-income countries.

The study will target English speaking middle-income countries in the Caribbean Community (CARICOM) who participated in the Heads of Government Port of Spain Declaration on NCDs (POSD). An on-going data analysis on mortality, morbidity and risk factor trends across the 20 CARICOM member states is underway, alongside detailed new data collection to define current policy responses to NCDs. This work will be used to inform which countries have sufficient data and strong stakeholder contacts for the development of a system dynamics model. Jamaica will be included as well as two other countries.

The study will use mixed methods, gathering primary data through qualitative methods and secondary quantitative data. These data will then be applied to the design and implementation of a system dynamics model for policy evaluation.

The qualitative phase will use in-depth semi-structured interviews with stakeholders identified through work with the on going POSD evaluation. The data will then be used to develop the first draft of a conceptual model using the Centers for Disease Control diabetes model as a basis.

A quantitative data collection phase will collect evidence on diabetes prevalence, risk factors, determinants, outcomes, and health system response for the countries of interest. This review will provide the necessary information to quantify the relationships described by the conceptual model. Following the initial data collection phase, a model building workshop including all relevant stakeholders and investigators will work intensively to further refine the conceptual model until a consensus is reached. The final conceptual model will then be tested and validated.

The final phase will involve a series of simulations evaluating policies for the prevention and management of diabetes. The results will be used engage policy makers and to evaluate the utility of the system dynamics methodology in adoption of policy.

The impact of the work can be divided into three time periods: immediate benefits to policy makers and other stakeholders who participate in the system dynamics model development and use; mid-term benefits to other policy makers in the Caribbean as the findings from the modelling process are disseminated regionally; and finally, long-term benefits as the work from this development grant is used in a larger study and adapted elsewhere.

Immediate benefits
These will accrue to the policy makers and stakeholders who take part in the model building workshop and suggest policy scenarios for testing in the developed simulation model. A major benefit of participating in the model building workshop will be from collectively thinking about the systems that are driving obesity, physical inactivity and access to effective diabetes care. This process of itself is expect to provide 'policy insights' i.e. to suggest new ways of addressing these problems. One specific example that will help in thinking through the potential effectiveness of different policies in their settings is using the simulation model to examine the feasibility of achieving the relevant WHO global monitoring framework targets for NCDs.

Study participants will also be actively involved in an evaluation of the system dynamics methodology and its application to the prevention and management of diabetes. Their feedback and input will benefit the development of the broader model, into which, in studies made possible by this development grant, costing of the policy interventions will be included. Their feedback will also enable refinement of the actual process of engaging with policy makers/stakeholders in other settings - from in-depth interviews to inform the model structure, to how to run the model building workshop and interact on the results of model simulations.

Those working directly with the stakeholders may also learn directly from their colleagues who participated in the study, disseminating knowledge and expertise into the wider diabetes community.

Mid-term benefits
The findings from the model development and simulations will be distributed through the wider public health community in the Caribbean Community (CARICOM) countries through the Port of Spain Declaration Evaluation currently underway. The regional community will then be able to incorporate findings and learning from this study into the development of policies for diabetes and other non-communicable diseases. The policy makers in this sphere, comprising members of the Caribbean Public Health Agency, the Pan American Health Organization and the Healthy Caribbean Coalition, will be exposed to systems thinking and its applications. The interest generated from disseminating these results may result in the adaptation and use of the model in other countries in the region.

The wider diabetes community in the Caribbean region is expected to benefit from the evidence-based policies and informed interventions adopted from the POSD Evaluation using this development grant. The hope is that the model will help to inform policy makers into effective combinations of policies that will ultimately reduce the burden of diabetes.

Longer-term benefits
One of the outputs from this study will be a manual on the adaptation and implementation of the model in other settings. This manual will serve to open the possibility of using system dynamics in decision making for public health in other low and middle-income countries. In addition, findings from applications in other parts of the world may benefit in further refining and improving the Caribbean model in the future.