Ebola, as I write, is ravaging several African countries, with the most optimistic predictions hoping that the rate of increase of horrible deaths will reach zero early in 2015. A recently discovered virus meets inadequate medical infrastructure and we have a classic public health problem on the scale of the cholera outbreaks that spurred the Victorians to build sewers, if not of the repeated plagues that swept London two centuries earlier.
Both cholera and the earlier plague outbreaks spurred the beginnings of statistical approaches to public health, and that's why today's researchers can predict how long it could take us to get the disease under control, how many more hospital beds Sierra Leone needs, and so on.
You could argue that it's already obvious how lacking Sierra Leone and its neighbours were in hospital beds, doctors and most of the things we take for granted in the UK. Certainly, a country that has to turn away highly infectious, dying patients because its clinics are full, is not going to able to contain Ebola. When the disease struck this year, Liberia had one doctor per 100,000 people. Nigeria, in contrast, has so far successfully controlled Ebola within its borders.
Nevertheless, some ability to model epidemics may direct help to where it can be most effective. And certainly, from epidemiology to genetics, much of medical science relies on data and statistics. In the UK, death by Ebola remains vanishingly unlikely.
Here, each of us will probably live to retirement age and eventually succumb to either cancer or some cardiovascular disease. So public health here tends to focus, not on the big infectious and parasitic killers, but on the factors that make small changes to an individual's odds, but look likely to make a difference across whole populations.
Many of these factors seem to be our own habits – for example: what, where and how much we eat, drink and smoke. So this week's proposal to improve the public health of Britain's population (a proposal backed by England's Chief Medical Officer, Sally Davies) is to ban smoking in London's parks and public squares.
Health policy, in particular, likes to be evidence-based. So what is the evidence on which this proposal is based?
New York City has already imposed a similar ban, and the London Health Commission report calling for a smoking ban in London's parks cites the success of this policy (though the graphic portrayal of the fall in smoking rates from 21% to 15% over 10 years shows more correlation between tax hikes and smoking cessation than any effect of smoking bans).
No other evidence is cited in the report. It would clearly be nonsensical to invoke the health effects of secondary smoke, as the report is proud to claim that parks make up nearly 40% of London's area. The city's 1.2 million smokers would have their work cut out to fill that open air with harmful air pollution. No, the proposed ban is justified thus:
'Smoke free London will be better for us all: a better example for children; fewer opportunities for smokers to smoke; less litter; more green and more pleasant places for us to come together for better health. It would be a powerful message for the iconic centre of our city and the political heart of our country to become smoke free. What better way to show our city’s ambition to be the healthiest major global city?'
It's not about the effects of smoking, it's about the effects of seeing people smoke.
The effects of smoking are bad for the smoker's health, no doubt about that. The link between smoking and lung cancer is the iconic success story of medical statistics, revealing a correlation so strong that the researchers who discovered it quit the habit without waiting until the causal mechanism was found.
Nevertheless, despite the ill effects being so widely known today, almost one in five adults continues to smoke. So public health research turned its attention to how to change human behaviour. What works best in reducing the smoking rate? Information? Legislation? A range of behavioural techniques often called ‘nudge’?
But people, their habits and their minds, are much harder to turn into data than the numbers tracking their illnesses and deaths. Seeing (or smelling) others smoke may be a temptation to a would-be quitter – but it turns out that seeing no-smoking signs also provokes the desire to smoke.
Predicting the effects of public policy is notoriously hard, even when you use both mathematical modelling and ‘ecological studies’ (looking at what other countries did). And this kind of predictive research is usually commissioned with an eye to policy initiatives already in somebody's mind, laying it open to the accusation of being ‘policy-based evidence’.
The problem for public health campaigners is that the white coat of science still looks much cleaner than the gravy-stained tie of the politician (of whatever party). Statistics or data carry more authority than any statement of principle.
So assertions which are based more on philosophical beliefs about how human beings work and how we ought to relate to one another, are portrayed as ‘evidence’ with a veneer of objectivity.
But there are questions for which statistics and data don't work. Even if research could show that seeing people smoke in parks correlates with more young people starting to smoke, would the ban be justified?
A park as a ‘beacon of healthy living’ is a profound redefinition of public space. The 40% of London in which its citizens can interact freely, without harming one another, becomes instead a space in which to model officially sanctioned healthy behaviour.
Which is why, when it comes to public health, data and statistics are not enough. We also need philosophers, political animals and yes, we the unruly public.
