Analysing “the Workings” of Health Systems as Complex Adaptive Systems

Abstract

Complex adaptive systems can be analysed in two very different ways.

• Looking backwards asking: which of its structures and behaviours allowed its current state to emerge (the what and why questions)

• Looking forward asking: changes to which of its structures and behaviours may most likely shape dynamics that achieve a future desired outcome (the how questions)

Looking backwards decomposes the system. Decomposition of systems can produce:

• Health atlases that describe the distribution of various health services like the location of hospitals, community practices, specialist medical and allied services, etc. in a geographic area

• Geospatial distribution maps superimpose a variety of different data onto a map and visually highlight their linkages. For example, researchers have found a strong relationship between socioeconomic characteristics, the distribution of fast food outlets, and obesity rates across different suburbs

• Creating geospatial maps for different time periods can show progress in combating particular issues of concern like the improvements in the battle against sleeping sickness in the Democratic Republic of the Congo

Looking forward takes account of the fact that the system as a whole, rather than its discrete entities, produce the behaviours and outcomes one observes and desires. Looking forward is the realm of modelling. A model is a “simplified version of reality”, it contains those “agreed upon” variables regarded as “responsible to cause” the behaviours and outcomes of the system.

Modelling in the first instance is a learning tool—it allows all involved in model building to gain a deep insight into the system’s structure and function. In the second instance it is a decision-making tool—it invites the exploration of “what-if” scenarios to help find the “best possible” approach to solving a problem. Modelling provides decision makers a “safe space” to explore the long-term effects of potential solutions on the system’s behaviours and outcomes.

However, modelling is not a panacea; as Rittel emphasised, every solution to a wicked problem is a “one shot solution”; modelling helps decision makers to “give it the best shot possible”.

Addendum 1

Three Types of System Diagrams

Rich picture diagram. Image courtesy of Ramage M and Shipp K. Expanding the Concept of ‘Model’: The Transfer from Technological to Human Domains within Systems Thinking

Rich pictures are unstructured pictures and usually drawn by hand. They depict all the major aspects of a problem of interest.

The aim of a rich picture diagram is to capture the full extent of the issues without giving any thought to their groupings or relationships, however, some related aspects of relationships between different issues are frequently illustrated through relative positions or simple lines connecting those issues.

Multiple cause diagrams are used to explore why changes or events happen.

The purpose of the diagrams is to examine the multiple causes behind particular events and processes. Here one looks primarily at the causal chain behind processes. Developing a multiple cause diagram may identify important feedback loops that can provide insights into the multiple causes of a system’s behaviour and how to make undesirable behaviour less likely.

Influence diagrams are a type of systems map that show the relationships between the components within the system. They highlight the relationships and influences between the system’s components.

An extension of influence diagrams are sign graph diagrams (not shown). Analysing the relationships between the components can tell if the component at the beginning of an error will cause a change in the same (indicated by a “+”) or opposite (indicated by a “−”) direction in the component at the tip of the arrow. Furthermore influence loops—technically known as feedback loops—become apparent and provide further insight into the dynamics of the system’s behaviour.

Multiple cause diagram. Image courtesy of Ramage M and Shipp K. Expanding the Concept of ‘Model’: The Transfer from Technological to Human Domains within Systems Thinking

Influence diagram. Image courtesy of Ramage M and Shipp K. Expanding the Concept of ‘Model’: The Transfer from Technological to Human Domains within Systems Thinking