The United Kingdom’s National Institute for Clinical Excellence (NICE) has proven its worth as an independent arbiter of value-for-money when it comes to medical interventions. People may question the details of the methodology they employ to compare effectiveness of drugs and procedures, but few doubt the principle of its existence.
Unfortunately, there is no equivalent framework for assessing public health interventions. A swathe of unreliable cross0-secvtional studies have highlighted potential harms from a diverse array of sources, but particularly various components of our diet. A vocal minority is calling for action to limit these perceived “harms”. But how real is the risk? Even where the risk is proven, how much of a contributor to our overall health are these factors? There is a very real risk governments will start to intervene (through legislation or tax measures to disincentives certain behaviours) on the basis of public perception rather than a quantitive, rational and transparent assessment of the costs and benefits of such interventions.
The solution, according to DrugBaron, is to adopt the NICE methodology to estimate value-for-money of proposed public health interventions, in exactly the same way we do for medical interventions. This would be the mission of the National Academy for the Study of Things that harm You (NASTY).
Although a small number of the individual decisions made by the UK’s National Institute of Clinical Excellence (universally known by its disarming acronym NICE) attract vocal criticism, particularly over high-priced cancer drugs that offer ‘only’ a modest extension of life, there is nevertheless a widespread recognition of the important role it plays.
NICE plays the role of gatekeeper, determining which medicines and procedures are available on the publicly-funded National Health Service – a role not dissimilar to the Pharmacy Benefit Managers (PBMs) in the US. The challenging objective for such organisations is to select those interventions that deliver at least a threshold value for money. This requires a complex calculation to somehow compare life-saving cancer drugs and hearing aids; psychotherapy and antiviral drugs; fertility treatments and hip replacement operations.
It is a thankless task, and one which is perhaps inevitably doomed to at least a degree of failure because different people value different outcomes (quality versus quantity of life, for example) very differently, and these subjective judgments cloud any rational assessment framework.
But the alternatives are worse: as we struggle to afford spiraling healthcare costs even with the limitations placed on free prescribing by NICE and the PBMs its manifestly clear that having no such guards in place is not an option. Leaving such decisions to local bodies seems similarly flawed: whatever the best estimates are for where the available dollars are best spent, they should be applied globally.
Interestingly, though, while governments (and the population at large) have largely come to accept the NICE methodology as the appropriate way to quantify cost-effectiveness for healthcare, the same rationality is not afforded to quantifying the cost of harms.
Exemplified most clearly by the UK government in the 1980s, who decided to ban beef-on-the-bone at the height of the mad cow disease outbreak while still affording its citizens the free will to decide whether to smoke tobacco, it is clear that political expediency rather than any kind of rational analysis determine what we ban, tolerate or encourage.
Where public health interventions are likely to be unpopular, its even more important to be confident that in return for such pain there is actually some benefit
As the debate rages as to whether to tax “unhealthy” foods, such as those with high saturated fat, the science often takes a back seat. Politicians think they know that such foods are harmful, despite limited interventional study data to support that view.
Part of the problem lies in the sheer difficulty of determining conclusively what IS harmful. Over the past few decades there has been an explosion in the number of cross-sectional studies, which examine a population of individuals dividing them into groups depending on a particular exposure (exercise, eating habits, alcohol consumption or whatever) and then following them to observe health outcomes (such as disease or death). A simple statistical test can tell you whether those with the highest exposure had a higher or lower risk of disease or death.
But it cannot tell you whether the exposure caused the outcome
The central issue with these cross-sectional studies is that unhealthy behaviours are often highly clustered within the population (people who eat unhealthily also tend to smoke more, drink more, exercise less and so forth). As DrugBaron noted several years ago, as a result cross-sectional studies tend to dramatically over-estimate the importance of any given risk factor.
Take saturated fat for example, demonized by the media on the basis of these cross-sectional studies. How much does it contribute to the development of heart disease? Or five portions of fruit and vegetables a day. Is that a panacea for all ills, as you might imagine given government-backed advertising campaigns promoting it?
The gold-standard test is the interventional study. Instead of just observing people with different exposures to the risk factor of interest, you actually intervene to change behaviours. And then watch the impact the changes have made. Such studies are much smaller in number than cross-sectional studies (because they are orders of magnitude more expensive, more challenging from an ethical perspective and longer to perform, often taking years or even decades to deliver their verdict). As a result, the information they give us is often drowned out by the inferior but bulky weight of evidence from the cross-sectional studies.
However, the answers they do give us are chastening
Reducing saturated fat intake has very little (none at all, in some very large studies) impact on heart disease, cancer or indeed all-cause mortality. Increasing intake of fruit and vegetables to five portions a day does improve outcomes, but by a tiny margin (a few percentage points reduction in cancer and heart disease – much less than the 30-50% lower rates of disease among those eating fruit and vegetables of their own volition in cross-sectional studies). The reason for these discrepancies is obvious enough: in the interventional study, the people eating less saturated fat are still engaged in other “risky” behaviours that dwarf the impact of the one small change made to their diet.
Looked at dispassionately, the evidence-base would not support government legislation to curb saturated fat content of food, or to tax high-fat foods to discourage their consumption. You could make a case for promoting fruit and vegetables but is the public health gain large enough to justify the cost of promoting it? And the effort (and cost) individuals have to go to in order to comply?
The methodology to answer that question may be imperfect, but it’s the same methodology that NICE uses to decide whether medical interventions are cost-effective. Do they deliver enough benefit to justify their costs? Exactly the same analytical framework can be applied to environmental risk-factors to determine which ones should be the priority for interventions.
In the 1980s the UK government banned beef-on-the-bone while still affording its citizens the free will to decide whether to smoke tobacco
Today, there is no consistent framework for translating the science base into coherent policy decisions. Instead, political expediency is patchily fused with ad hoc academic publications, usually selected to match the agenda of the authorities performing the analysis. The solution would be the formation of an equivalent body to NICE to perform a rational analysis that, although based on imperfect methodology, would nevertheless provide a level playing-field for comparing different public health initiatives – a body DrugBaron has termed the National Academy for Study of Things that harm You (or NASTY for short).
On the face of it, politicians may feel uneasy about the rulings such a body might make. Tobacco smoking is clearly disproportionately harmful and the current regime of high taxes may be insufficient to mitigate that risk, yet with 20% or more of voters smoking tobacco regularly politicians may shudder at racking up the disincentives against tobacco smoking any further.
Yet without such a body, there is a very real risk that politicians will come under increasing pressures to act on things where the evidence is somewhere between sketchy and non-existent. A vocal minority calling for taxes on saturated fats or sugar in foods, for example, risks leaving the politicians looking like they don’t care about public health unless they act. A body such as NASTY resolving that the evidence against, for example, trans-fats does not merit regulation or other public health interventions would provide politicians with the backbone necessary to resist these calls.
Indeed, it was a similar rationale that led to the creation of NICE in the first place. Ministers feared the negative public reaction to any refusal to provide high-priced medical treatments even if all the data suggested they were only marginally-effective or provided very poor value for money. By pointing to an independent assessment of ‘value’ politicians hoped for (and to an extent achieved) the ability to deflect those criticisms.
But the best reason for setting up NASTY would surely be the desire for science-driven policy. Politicians presumably want to enact interventions that really do deliver the best outcomes for their populations. Where those interventions are likely to be unpopular (its hard to imagine a tax on perceived unhealthy foods, for example, could be anything other than a vote-loser on average) its even more important to be confident that in return for such pain there is actually some measurable benefit.
As the BBC prepares to broadcast an edition of its well-respected Horizon series asking whether eating meat is safe (since very many cross-sectional studies show that eating large amounts of processed meat is associated with a wide range of negative health outcomes – although no interventional study data is available to determine if such an association is causal), we face the very real spectre of governments intervening in our lives in all kinds of well-meaning but ultimately utterly fruitless ways.
At heart a “small government” supporter, DrugBaron does nevertheless accept that government interventions such as banning smoking in public places and taxing alcohol and tobacco are necessary and appropriate. But if the trend is for more widespread intervention to protect public health, DrugBaron for one would like to see some kind of independent framework in place to filter the science and provide transparent recommendations for intervention that pass some kind of agreed threshold for benefit versus cost.