||Andrada Tomoaia-Cotisel1, Karl Blanchet2, Zaid Chalabi3, Samuel Allen 4, Victor Olsavsky 5, Cassandra Butu6, Michael Magill7, Bernd Rechel8
||1Health Services Research & Policy, London School of Hygiene and Tropical Medicine, Cluj-Napoca, United States, 2Department of Clinical Research, London School of Hygiene and Tropical Medicine, London, United Kingdom, 3Department of Social and Environmental Health Research, London School of Hygiene and Tropical Medicine, London, United Kingdom, 4Utah Medical Education Council, Utah Medical Education Council, Salt Lake City, United States, 5WHO Country Office Romania, WHO Country Office Romania, Bucharest, Romania, 6WHO Country Office Romania, WHO Country Office Romania, Bucharest, Romania, 7Department of Family & Preventive Medicine, University of Utah, Salt Lake City, United States, 8Department of Health Services Research and Policy, London School of Hygiene and Tropical Medicine, London, United Kingdom
|Country - ies of focus
||Romania, United States
|Relevant to the conference tracks
||Policy-makers are better able to identify and implement effective health system strengthening (HSS) efforts when they have an accurate understanding of the dynamic, emergent behavior of the system they are attempting to strengthen. Achieving such an understanding is difficult. Yet, without it, decisions can easily result in unintended consequences or policy resistance. This paper describes system dynamics methodologies employed in the context of a HSS effort in Utah, USA and explores ways of applying them in LMICs, based on a case study in Romania. We present differences in data needs, availability and quality; and discuss how methods can be modified in view of these constraints.
||Policy-makers are better able to identify and implement effective health system strengthening (HSS) efforts when they have an accurate understanding of the dynamic, emergent behavior of the system they are attempting to strengthen. Achieving such an understanding is difficult. Yet, without it, decisions can easily result in unintended consequences or policy resistance. In high-income countries, such understanding is increasingly obtained through the use of complex system modeling and detailed statistical analysis using large datasets. However, in low- and middle-income countries (LMICs) the data available are more limited, introducing higher levels of uncertainty in health system parameters. Despite this uncertainty, systems thinking and system dynamics supplies decision-makers with information needed in HSS efforts.“Systems thinking” provides a comprehensive framework for capturing, from diverse perspectives, how health systems function and how complex changes occur. System dynamics takes this approach to the next level by developing quantitative computer-based simulation models that can analyze system behavior and simulate how systems respond to policy measures and other changes over time.
||To describe system dynamics methodologies employed in the context of a HSS effort in Utah, USA. Methodologies used are explained and ways of applying them in low and middle income countries are explored, based on a case study in Romania. The World Health Organization projects the burden of non-communicable diseases (NCDs) in LMICs to grow from half of total disability-adjusted life years in 2004 to three quarters by 2030. As LMIC health systems are already strained, this awareness necessitates that LMIC policy-makers anticipate and prepare for the consequences of this shift. As many NCDs are best managed in primary care settings, many HSS efforts aim to enhance primary care. System dynamics provides methods for creating custom-tailored tools to do this.HSS efforts in Romania, as in other former communist countries, focus on overcoming a previous neglect of primary health care, while redesigning the provision and financing of primary care at the same time. The goal being to facilitate patient centered care with a whole person orientation, providing all key elements of primary care.
||System dynamics methodology will be presented as used in a high-income country setting and as modified for implementation in a middle-income country setting. In both contexts, the core methodology progresses as follows: 1) develop a conceptual model of the health system, 2) transpose the conceptual model to a dynamic quantitative model of the system, 3) develop and run scenarios simulating the policies and interventions under consideration. This methodology is couched within a participatory action research approach. Methodological tools employed included: Causal Loop Diagrams (CLDs) identifying key system structures such as feedback loops and time delays; statistical analyses and literature review identifying relationships among system variables; model validation techniques and key informant discussions with a diverse set of stakeholders. Decision-makers are involved throughout the project, participating in model development and critique, providing key informant expertise, designing scenarios to be tested, and discussing scenario results.We present differences in high and middle income country data needs, availability and quality. We also discuss how methods can be modified in view of these data constraints. These modifications impact the model produced and the lessons obtained from it. Strengths and limitations of these modifications are discussed.
||We found that applying a SD methodology in LMICs is possible, but that the level of uncertainty in the model developed depends on the type and amount of available data. CLDs can be developed on the basis of interviews with key stakeholders, as well as using information in the literature. Quantifying the relationship between the identified system variables should ideally use context-specific data to increase model validity. However, model validation techniques can be performed using less data, for example via key informant discussions to elucidate a relationship’s potential behaviour. A health system model can be operationalized using less than ideal datasets. Existing data sources include qualitative and quantitative data on primary care in Romania and nationwide hospital diagnosis-related groups (DRGs) data. Additional low-cost resources would be required to conduct key stakeholder interviews to verify model structure and to design policy scenarios.
||Applying system dynamics in HSS requires the creative use of mixed methods within the constraints of data availability, transdisciplinary research teams and multi-level stakeholder involvement (of patients, providers, administrators and policy-makers). In particular, in LMICs’ HSS efforts, policy-makers need to know how to adapt innovations to their specific context and health system. System dynamics methodology promises to allow for this kind of tailoring; it also provides a framework for conceptualizing and simulating system behavior. Its design, tools and required parameterization can draw on experiences from elsewhere, while at the same time be adapted to local contexts.