The Health Foundation is working with Dr Harry Rutter from the London School of Hygiene and Tropical Medicine to develop a new model of evidence that will inform the policy and action needed to make our population healthier. We spoke to Dr Rutter about the challenges inherent in generating and using evidence in new ways, and how traditional measures alone don’t capture the complexity of work in this area.
Where do the greatest challenges lie for improving the health of the UK?
We are faced with many big health challenges in our society: things like obesity (which has been my main area of work over the last decade), health inequalities, mental health, and non-communicable diseases all have a major impact on our health and wellbeing.
Why do we need to develop a new model of evidence?
These issues, and the factors that drive them, are highly complex and overlapping, and traditional scientific methods are not always the best tools to test the effectiveness of our responses. When dealing with complex problems isolated interventions may not have direct impacts on outcomes through linear cause and effect, and should instead be seen as creating changes within complex systems.
You can’t generally ‘solve’ or ‘fix’ a complex system problem in the same way that you can solve a simple problem, but you can influence the system within which it sits to achieve a more beneficial set of outcomes: you reconfigure the system, rather than ‘solve’ the problem.
This is what my work with the Health Foundation has been looking at. We’ve a paper due to be published in the Lancet shortly, which pulls together a broad range of perspectives on complex systems thinking and how to apply it. Over the next couple of years we’ll start to come up with guidance and practical support to help people turn these ideas into action.
What will be the benefits of changing how evaluations happen?
At the moment many public health problems seem intractable. That’s partly because we have an evidence base that is quite good at answering questions about specific interventions that affect individuals, but much less good at answering questions about the effects of actions on complex systems. So the evidence is limited, and in some cases it may even be unhelpful if it pushes us to continue focusing at the individual level to the detriment of more appropriate measures that could have large scale positive impacts across populations.
Take obesity for example. Bariatric surgery is an extremely effective procedure. It’s not for everyone, and it’s a difficult operation to live with, but in some people it can even reverse type II diabetes within a week or so. Randomised controlled trials have shown its effectiveness as an intervention, and we now do around 6,000 operations a year. But that is in the context of well over a million people in this country with severe obesity. Even without getting into the practicalities or ethics of scaling up the number of operations it is clear that surgery is not a feasible or appropriate population level response to the problem.
At the other extreme, we have interventions such as the National Cycle Network. This is a potentially important contributor to physical activity among a large proportion of the population. But you can’t conduct a randomised control trial of the National Cycle Network. It’s not an intervention that can be tested in isolation in the way that a pill or a surgical procedure can.
If we can work out ways to conduct scientifically robust evaluations of those kinds of interventions, identifying if and how they can help improve health, then I think we’ll stand a better chance of being able to put together effective responses to some of those seemingly intractable problems.
What would a different approach to evaluating interventions involve?
There’s no simple answer to that, but some work that we are doing with the proposed tax on sugar sweetened drinks in this country provides a good example.
The obvious thing to evaluate is how the tax affects people’s purchasing and consumption patterns. But to some extent we already know how this is likely to play out. There’s plenty of good existing evidence that shows that, all other things being equal, if you put up the price of something, in general people will consume less of it.
Let’s think about the other likely responses to a levy on sugar. The drinks industry may well respond by changing advertising, marketing, and the way drinks are displayed in shops. There might be other drinks that manufacturers see an opportunity to promote. All these responses may lead to behavioural changes by individuals. So the change in the pricing of sugary drinks won’t be occurring in isolation. The wider food and drink system will adapt in response to the introduction of the levy and find ways around it.
In order to ensure we understand this broader picture, we’ve suggested that an evaluation of the sugar tax needs not only to look at the impact on consumption, but also at wider contextual factors. Does the tax change public attitudes towards sugar sweetened drinks? Does it have any impact on the way the media portray obesity, and so on? Because those things can all affect behaviours - of individuals, politicians, and the industry.
Who will find a new approach to building evidence useful?
I hope a wide range of people will find this work useful. For researchers, a new approach will require the development of different skills and methods, but it also opens up opportunities for collaborations across disciplines and sectors. For funders of research projects, it will show the value of supporting research at population level, even where that’s methodologically difficult. And of course we hope that the evidence that emerges will help policy makers and practitioners across sectors to find new and more effective ways of responding to challenges.
This isn’t going to happen overnight. It involves adopting a fundamentally different mindset from the one most of us have been trained to use.
We need to be willing to take risks if we are going to innovate, but we shouldn’t ignore the even bigger risks attached to the status quo. Until we are prepared to try out some different approaches we are unlikely to make significant progress on the biggest health challenges we face, in this country and around the world.