In this piece, Dr Aveek Bhattacharya, Social Market Foundation, discusses individual-level and population-wide public health interventions, providing social science evidence for their efficacy and recommendations to the next UK government.
‘Following the social science’? What politicians can do to improve public health
Britain is not in a healthy place. In the 100 years to 2011, life expectancy increased by three years per decade. In the subsequent decade, life expectancy fell. This was primarily because of COVID-19, but progress had stalled before that, with male life expectancy rising just 0.8 years and female life expectancy rising 0.6 years between 2011 and 2019 in England.
To a substantial extent, those negative trends are due to ‘behavioural risk factors’ – things like smoking, drinking, diet and exercise. It is estimated that 40% of premature mortality is due to behavioural patterns and 50% of the health status is due to the social and economic environment.
If the next government is to take decisive action to address those problems, it ought to acquaint itself with the social science evidence on public health policies. In fact, public health is a particularly interesting and challenging area for social scientists seeking to influence policy because evidential expectations are often higher, anchored to the level of medical trials. That tension was evident during the pandemic, when policymakers unable to draw on randomised trials as they could for vaccines, had to make use of more pragmatic techniques (observational studies, natural experiments, modelling) for non-pharmaceutical interventions like social distancing and masking.
Yet those drawing up policies for the election and beyond should broaden their approach, or risk missing the most effective interventions. Consider the ACE-Obesity study, a priority-setting project conducted by researchers at Deakin University. Most of the interventions it examined had a ‘low certainty of effect’ because they are applied at a population level, and so are not amenable to the sorts of experimentation that sit atop the evidence hierarchy. By contrast, individual-level interventions – like incentive payments for losing weight and weight management programmes – have the strongest evidence behind them because they can be applied randomly to different populations.
Yet that limited scope also means that individual-level interventions tend to have more modest effects. Modelling conducted as part of the ACE-Obesity study found that on average, population-wide regulatory interventions saved 1.7 times as many health adjusted life years as programme-based measures. This should not be surprising: we can reach many more people at a time with population-wide interventions. Levying taxes or regulating marketing affects tens of millions of people at a stroke. By contrast, with individual interventions, we have to engage people one at a time.
In general, in my review of the evidence of policies to address smoking, harmful drinking, problem gambling and obesity, I found that more interventionist policies tend to be more effective. In particular, taxing harmful commodities, reducing their availability (for example, through licensing restrictions) and regulating their marketing have the greatest effect. By contrast, educational and informational campaigns tend to be less effective.
The trouble is that these more interventionist policies tend to be less palatable for politicians (though not, typically, unpopular – resistance tends to come more from industry and media). That is why, for example, the government’s obesity strategy has been so weakened and delayed. There are promising signs – for example, the current administration’s smokefree generation policy and the opposition’s willingness to embrace the ”nanny state” as it implements a 9pm watershed for junk food advertising on TV.
The case of minimum unit pricing (MUP) for alcohol illustrates many of these issues. The policy eliminates the cheapest drinks associated with the greatest harms by setting a ‘floor price’ below which it is illegal to sell alcohol: 50p per unit in Wales, €1 in Ireland and soon to be 65p in Scotland. Yet policymakers at Westminster have resisted calls to introduce it in England. Keir Starmer has objected due to his concerns around the cost of living, the Government continues to “monitor emerging evidence”, remaining in the same holding pattern it’s kept for over a decade.
This is peculiar because Scotland’s official evaluation of MUP, conducted by Public Health Scotland, is already complete. It was, in the words of one researcher, “notable for its ambition and comprehensiveness”. On top of that, there have been independent analyses conducted by other researchers.
The evidence is, on the face of it, positive. Public Health Scotland estimates that deaths wholly attributable to alcohol fell by 13% as a result of the policy. Alcohol consumption is believed to have been cut by 3-3.5%. Yet some have found this hard to square with the recent sharp rise in alcohol deaths in Scotland – notably the Scottish Conservative health spokesman, Sandesh Gulhane, who accused the Scottish Government of “grossly misleading” the public by claiming MUP has been effective.
Though there is some statistical complexity in the methods used by evaluations, the apparent paradox is easily explained. In the first year and a half following the introduction of MUP in May 2018, alcohol-specific deaths fell in Scotland. Then the pandemic happened. It increased harmful drinking just about everywhere, including in Scotland. Even so, Scotland’s death rate rose less than England’s, where MUP was not in force. On that basis, Public Health Scotland concludes that alcohol deaths, high as they currently are in Scotland, would have risen further without MUP.
How far you accept their conclusion depends on how plausible you think trends in alcohol harm in England are as a counterfactual for what would have happened in Scotland without the policy. That is the nature of causal inference in social science, and it’s the sort of question policymakers should be more comfortable engaging with.
For my own part, I find the evidence from the official evaluation compelling, particularly when triangulated with the broader evidence that higher alcohol prices (mostly on alcohol taxes) reduce harmful drinking and save lives. But as social scientists, we have to admit that there is an element of judgement, and one in which policymakers should share.
Policies like minimum pricing – and other more experimental ones like taxes on unhealthy foods – are likely the best tools we have to address behavioural health risks. They should be on the agenda for those writing party manifestos.
About the author
Dr Aveek Bhattacharya is the Interim Director of the Social Market Foundation, having joined as Chief Economist in September 2020. Prior to that, he was Senior Policy Analyst at the Institute of Alcohol Studies, researching and advocating for policies to reduce alcohol-related harm. He has also previously worked for OC&C Strategy Consultants, advising clients across a range of sectors including retail, consumer goods, software and services.
Aveek studied Philosophy, Politics and Economics at undergraduate level, and has Master’s degrees in Politics (from the University of Oxford) and Social Policy Research (from the London School of Economics). He holds a PhD in Social Policy from the London School of Economics, where his thesis compared secondary school choice in England and Scotland. Aveek is co-editor of the book Political Philosophy in a Pandemic: Routes to a More Just Future.
Image credit: Ross Sneddon, Unsplash