‘Simplicity, simplicity, simplicity!’ The mantra may have worked for Henry David Thoreau as he sat around Walden pond, but there’s a growing recognition within the health systems, development and humanitarian relief communities that ‘complexity, complexity, complexity’ is more appropriate these days. Complexity science isn’t new, but applying it in these fields is relatively recent. My new paper with Ligia Paina, ‘Understanding pathways for scaling up health services through the lens of complex adaptive systems’, begins to unpack the implications for health systems if we take a complex adaptive system (CAS) lens to understand initiatives and scale up health services.
And while this blog looks at how a CAS approach can help us design and deliver better programs, Ben Ramalingam (a visiting fellow at the Institute of Development Studies and an expert on complexity science) and I have also recently sat down with Jeff Knezovich from FHS to produce a podcast looking at the issue in more depth. You can listen to the podcast below.
Complex adaptive systems are described as such because, in addition to being comprised of many interacting components and agents, they have the capability to self-organize, adapt or learn from experience – what are sometimes known as emergent properties. Most social, biological and economic systems can be considered CAS, as well as many complex physical systems, such as those related to weather. The interactions of system components are non-linear, and are not easily controlled or predictable in detail.
Whereas scientific enquiry attempts to simplify understanding and create simple and elegant solutions, the CAS approach is important, as often our simpler models just aren’t good predictors of behavior. X doesn’t necessarily lead to Y, and indeed it might not even lead to one specific point. This can be a big problem when planning interventions.
Keeping a few CAS concepts in mind while framing projects and programs can certainly help improve them. In the paper, we look at several of these concepts and how they can be applied to health systems. In particular we look at: emergent behavior, path dependency, feedback loops, scale-free networks, and phase transitions. I encourage you to read the paper or see my presentation for more information about these concepts.
Focusing on ‘emergent behavior’, the first phase of FHS has shown us why this is important. The ‘Safe Deliveries’ intervention, which was led by the FHS Uganda team at Makerere University, worked both on the demand-side and supply-side to improve access to institutional deliveries in rural areas of Eastern Uganda. On the supply side, it led trainings for health workers and provided essential equipment, drugs and supplies. On the demand side, the program organized a significant voucher scheme for both maternal and newborn services (including antenatal screenings, delivery and newborn care) and for transport to clinics via boda boda (motorcycle taxis), as transport to facilities was a big factor preventing institutional deliveries.
One of the interesting – if unexpected – things to happen, was that the boda boda drivers actually organised themselves in such a way that they started to keep track of and encourage pregnant women to go for care. Obviously there was a built in financial incentive, but this level of ‘enforcement’ had not been planned for – it emerged from the complex system. More importantly, it ended up playing a significant part in tripling the average monthly number of births in facilities.
All of these phenomena have an implication for what it means to create change. What a CAS approach tells us is that, when designing and delivering programs, we need to:
- Plan differently: Don’t expect to control change. Expect complexity. Expect emergent behavior (and feedback loops, and all those concepts I mentioned above) and expect the unexpected. This means looking beyond where a typical research program might be shining its light. In a complex adaptive system, a specific intervention is likely to cause shifts in other parts of the system, whether it’s through displacement or new actors or something else. In order to keep an understanding of what is happening in other parts of the system, this may also mean involving a wide number of stakeholders in both the planning process and the implementation phase.
- Plan to re-plan: We need to avoid too much of an emphasis on the first planning cycle and get away from blueprints. Detailed planning almost never works, and if it does, it probably didn’t need to be planned for in the first place. Allow for course correction by creating mechanisms that allow ‘learning by doing’. One approach to this might be to create more iterative, rapid learning cycles. This might include changing an intervention or, in research, changing what is being measured.
- Use mixed-methods research approaches: Having multiple perspectives and multiple methods helps to better identify changes in a system. Working with complexity is a data rich process. Having good access to information, and from multiple sources, makes it easier to make relevant decisions. Don’t just ask whether or not something works, look at how and look at why – both of which might be important for scaling up the interventions.
These are principles that we’re trying to embed into the next phase of the Future Health Systems project. Already we’ve had a training workshop with our FHS China team in Beijing to orient them to the approach and to help them make sure their research design incorporates some of these ideas.
If you’re interested in finding out more about our growing body of work on complex adaptive systems, visit our theme page for more resources.