Archived: When theory follows practice: learning from complex interventions
Penny Hawe happily admits to fitting a popular joke definition of an academic – a person who sees something working in practice and wonders if it could work in theory.
“When you figure out the theory or pattern that fits people’s experience, you have the key for replicating and strengthening effects elsewhere,” she says. “Unfortunately, lots of people equate theory with abstraction, and they equate abstraction with distance or irrelevance, but nothing could be further from the truth. Understanding the theory can make practice more powerful.”
Professor Hawe, Professor of Public Health with the Menzies Centre for Health Policy at the University of Sydney, is leading a program of work on complex intervention theory for the Prevention Centre. She says the purpose is to gain stronger insights about mechanisms of action.
Papers she has published this year tell the story. Her review of complex interventions in the Annual Review of Public Health demonstrates that interventions that harness complexity have stronger health impacts than those that impose rigid rules. Harnessing complexity involves adapting to context, taking advantage of local relationships and resources, and allowing lessons from implementation to be incorporated into ongoing improvement.
Context affects outcome
The team has also observed how the context or system of an intervention affects the outcome. “In our case studies the system is a school,” Professor Hawe says. “But it could be a workplace or something else.”
They studied the transfer of Australia’s Gatehouse Project into Canadian schools and from a younger age group to an older one. The Gatehouse Project involved participatory action research to create organisational change and a more socially inclusive ethos in the school. The project achieved relative reductions in smoking, drinking and drug use of around 20-40%. These effects were much larger than those obtained by highly packaged, curriculum approaches.
Professor Hawe’s replication in Canada showed a similar type of large jump in a bundle of risk factors as well as a relative reduction in unsafe sex of 61% among girls.
How interventions activate systems
The team believes the findings show that interventions activate the systems they are placed in to a greater or lesser extent. They affect different risk factors differently according to the prevalence, salience and distribution of the risk behaviours and the interconnectivity of relationships between the people within that system. “The interconnectivity of the people in the system can be targeted by the intervention and this could radically alter intervention effects,” Professor Hawe says.
The team has gained insights about these system-activating interventions from some unusual places. In a cluster randomised trial of a whole community intervention to reduce postnatal depression in Victoria, the lessons came from the comparison communities that did not get the intervention.
“The maternal and child health teams in the comparison communities were faithful to the trial conditions and did not reproduce the activities that they knew the intervention communities were getting. But they may have invented something better instead,” she says. “Essentially, in complex systems language, it seems that the comparison communities self-organised.”
Professor Hawe says the results might explain previous evidence from elsewhere that overly stipulated preventive interventions have weaker sustainability than those where local people have autonomy.
The program of work is being noticed. Professor Hawe has an invited commentary in Social Science and Medicine on minimal, negligible and negligent interventions. It questions the ethics of not employing complexity theory and methods in the face of entrenched patterns of health inequity.
The importance of using a complexity approach when considering health inequity is also being championed in another Prevention Centre supported project, led by Professor Sharon Friel, who with her team is taking a systems approach to developing a healthy and equitable food system.
But system-activating interventions have a downside, Professor Hawe says. “We can observe big changes in a bundle of risk factors, but there is no guarantee or predictability about any particular one. This can be off putting for funders and managers accustomed to separate funding streams for smoking, alcohol, diabetes, obesity and so on. But new models of project management and performance monitoring are being developed.
“After years of seeing little change in many risk behaviours, and in some cases things getting worse, we need to rethink the approach,” Professor Hawe says. “That is what the Prevention Centre is for.”