An adaptive approach to influencing complex systems

There are many and varied complex problems in society. Government health and social service systems are often overwhelmed and there are a growing number of people who are unable to afford the basics of food and housing and are excluded from the workforce. Given this it can be hard to know what to do to improve things, and even if we do know, where to start.

The traditional approach to solving problems involves coming up with a possible solution, bringing together a team and associated resources, developing a plan and implementing the plan, followed by evaluation of how we went. This approach requires us to create a complete plan before we start to tackle the problem, using the information we already have to take action. The risks with this approach are that we generally do not have all the required knowledge or information to take the best action, and most problems we deal with are not caused through simple linear pathways, but rather part of a complex set of interacting systems. This traditional approach can be very slow or delayed in implementation and we invariably are required to make several assumptions when producing such a plan which may lead to misalignment and failure of the intervention.  

We also know we cannot necessarily predict how the complex systems that affect our health will respond to an intervention. For example, change in one part of the system can lead to a counter change in another part of the system leading to no net benefit. This is like banning junk food advertising within a school only for new advertising billboards to be erected just outside of the school grounds.

Is there another way? Public health approaches have always borrowed from the disciplines of others. In the mix of public health professionals are town planners, and professionals from environmental health, communicable disease control, marketing and communications. There are lawyers, medical staff, nutritionists and developers (to name only a few). Rarely however have we learnt from the entrepreneurial sector.

My colleague Katherine Pontifex and I co-authored a paper which explored in detail what public health can learn from Lean start-up approaches (D'Onise and Pontifex, 2023). In summary, an intervention is implemented early, while the intervention is only partly formed, a so called minimum viable product. Alongside there is an explicit process of testing the assumptions we have made about the intervention through the use of deliberate micro trials and collection of data. This cycle is repeated allowing refinement of the intervention over time within the real-world systems that we are trying to shift.

There are many tools in the public health toolkit. This is another approach we can consider when change is being introduced into an already crowded space, with many stakeholders, significant complexity and uncertainty. These conditions suggest an adaptive approach is warranted, enabling us to start to tackle the problem, and to learn about possible solutions in a systematic way through the collection of meaningful data as we go. Over time, solutions are tailored to the context and end user, improving both efficiency and effectiveness.