Revisiting the paradox of preventive medicine

By Dr Stephen John

I once went to a talk by a prominent epidemiologist which involved discussing the work of the great British epidemiologist, Geoffrey Rose. In particular, the speaker was taken by Rose’s insight that often the best way to improve overall health outcomes is through “population strategies” – which reduce the small risks suffered by a large number of people – rather than “high-risk strategies” – which reduce the large risks suffered by a small number of people. This was stirring stuff, providing a sophisticated justification for a range of controversial policies, from “minimum unit pricing” of alcohol to routine prescription of statins; although such policies may do little for each individual – reducing individual risk by a tiny amount – they can, cumulatively, end up greatly improving health outcomes. After the talk, I ended up chatting to the speaker. He was enraged that a friend had recently been fined for driving “just over 80” on the motorway: why, he asked, bother with people like his friend rather than boy-racers?

I resisted the temptation to remind the speaker that his own talk answered his question, instead tucking away the anecdote to use as my go-to illustration of an important issue: although Rose’s insight is logical and powerful, it is also extremely difficult to grasp. Intuitively, it seems obvious that we should “punish” those who impose high risks on others, and “help” those at the highest risk of harm. To take a contemporary example, many commentators seem to find it “obvious” that the COVID-19 vaccination programme should target the “vulnerable”, even though Rose-style reasoning suggests that it is far from obvious that this policy will have the greatest effect on overall morbidity and mortality.

One way of thinking about this phenomenon is to suggest that humans are just irrational – that, for whatever reason, we just cannot internalise the mathematical logic of Rose’s work. Another response, which I have explored in my work, is that our resistance to Rose’s insights points to a deep tension in our ethical thinking: that we have ethical reasons both to care about helping those at the greatest risk of harm and to care about improving overall population health outcomes. What Rose’s work vividly illustrates is that we can’t always have all of the good things at the same time. Much as medical practice might involve tensions and conflicts between independently important ethical principles – as, for example, when we weigh a concern to improve a patient’s welfare against respect for autonomy – so, too, public health policy might involve a clash between independently important ethical concerns – to protect the vulnerable and to do as much good as possible.

One advantage of this way of thinking about public health policy is that it captures some of Rose’s own concerns that his results seem to create a “prevention paradox”: there is something paradoxical in the idea that a policy can simultaneously be brilliant for the population, but of only marginal utility for any individual. 

However, if we think that we should care not only about the outcomes of public health, but about individuals’ risks, we face some problems. Some of these problems are conceptual: what are we doing when we calculate individuals’ risks? Are we measuring some independent quantity, like when we measure blood pressure? Some of the problems are sociological: how do notions of risk relate to concerns about medicalisation and disease creep? Finally, some are ethical: it is one thing to say that there is a tension, but another to say how to resolve that tension.

So, the senior epidemiologist was right: we should and shouldn’t adopt population strategies.