BMI: the GDP of health

I've railed many times about the inadequacies, and worse, of using Gross Domestic Product as the de facto indicator of social well-being. Which it has become, in the absence of any serious thinking about the social goals we want to achieve and the priorities we give them.

A similar phenomenon seems to have occurred in the world of health, where the body-mass index (BMI) is a long-used measure of obesity. BMI is body mass in kilograms divided by the square of the body height in metres. Apparently, as measured by BMI, obesity in the US plateaued around the year 2004 at 35 percent of the population. But, just as GDP ignores such things as leisure time, state of the environment, income distribution; and counts useless or anti-social economic activities as positives, so too does BMI have its flaws:

It does not consider distribution of fat, type of fat, muscle tone, age, sex, or even big bones. In spite of these flaws, healthcare professionals continue to use BMI as a guideline. A BMI of 20-25 is considered ‘normal’, and anyone larger or smaller is automatically counselled to achieve a healthier weight. .... Obesity is generally understood as a risk factor for heart disease, stroke, cancer, and diabetes, as well as an increase in overall mortality. Excess body weight also increases stress on joints and internal organs. Given these concerns, it’s easy to understand why so many people have celebrated the plateau in BMI. Unfortunately...BMI may not be the best measure of obesity. Your percentage of body fat and waist or abdominal circumference are far more reliable personal indicators of health outcomes than BMI. For example, central obesity, measured by waist circumference, is a more accurate determinant of personal risk and shows an even stronger correlation with poor health outcomes. Caroline Weinberg, Fat but fit?,  'Aeon', 27 August 

There is no clarity about goals. If our intention is to improve the health of people, why not target indicators of people's health, instead of easy-to-calculate but flawed measures such as BMI? If our intention is to prevent disasters arising from adverse climatic events, why not target for reduction the negative impacts of such climatic events, instead of atmospheric carbon dioxide levels or temperatures recorded in weather stations? If our intention is to improve social well-being, why not target indicators of social well-being instead of those economic activities captured by GDP?

BMI, greenhouse gas emissions, GDP per capita: they might answer certain specific questions, but their use as policy instruments is inadequate at best, dangerous at worst. Policymakers need some humility here: neither they nor their advisors know the best ways of improving health, preventing climatic disasters or improving social well-being. Even if they did, at one particular point in space and time, circumstances vary with region and our knowledge grows with time. No government, no single organization can hope to use fossilised knowledge on a one-size-fits-all basis and achieve meaningful results. Instead, policymakers should set broad goals, and let a motivated coalition of interests explore diverse, adaptive approaches aimed at achieving society's broad social and environmental goals. Government can set these goals and, indeed, it is probably the organization best suited to doing so. And, if it concentrated on that, it would come up with a better array of target outcomes than the flawed indicators it now uses, whether explicitly or not.

Government can also raise the revenue to reward the people who achieve these outcomes. But it has no business dictating how they shall be achieved, nor who shall achieve them. We need diverse, adaptive approaches, and those are exactly the sort of approaches that government discourages.

Smoking and obesity; anxiety and sirens

The UK has seen an onslaught against tobacco smoking. At the same time cases of diabetes are soaring, such that 'diabetes medication now accounts for 10% of the NHS [National Health Service] drugs bill'. Research appears to show that people who give up smoking put on weight.

By doing everything possible to suppress smoking has the UK Government unwittingly encouraged obesity and diabetes?  Do the social costs of more obesity and diabetes outweigh the benefits of less smoking? I have no idea, but the important point is how little it is in anybody's interests to answer these questions and use their answers to influence government policy. With smoking the government has had an easy ride: 'everybody knows' that smoking is bad for you, just like 'everybody knows' that taking illegal drugs is bad for you, as is drinking alcohol. You see where I am going here: road traffic kills 1.24 million people annually worldwide, but there are benefits to it as well as costs, as there are for drinking, taking illegal drugs and, yes, smoking, especially, but not only, insofar as people who are denied the opportunity of smoking then may be more likely to be become obese and diabetic. These costs aren't easy to calculate of course, but government has created an environment in which nobody has an interest in doing those calculations. Instead, seeing that smoking directly and obviously causes some diseases, it reacts in the Pavlovian, short-term, one-size-fits-all manner that it, in common with other governments, adopts when they encounter the symptom of a problem. So now, in England: 'work smoking rooms and areas are no longer permitted. All smokers must take their smoke breaks outside.' I've no doubt that rates of lung cancer and other diseases directly related to smoking have fallen as a result. But, as well as the costs to freedoms of the campaign against smoking, there are also the indirect costs to physical health, possibly taking the shape of increased rates of obesity and diabetes. The cancer specialists, and the well-meaning (though perhaps hysterical and self-righteous) anti-smoking lobby have no incentive or capacity to see whether smoking bans help or damage the overall health of people. Nor, under the current policymaking regime, are there any incentives for others to do so. And smokers are an easy target. Car drivers not so much.

We see the same in the area of mental health. The small city in which I currently live is blighted, maybe 20 or 30 times in every 24-hour period, by emergency vehicle sirens. Designed to create alarm and panic, that is what they do, to thousands of people, day and night. I have no doubt that these sirens shave a few seconds off the average journey time of the police, fire and ambulance vehicles. And those few seconds, might, on occasion, make the difference between life and death. But has anybody looked at the costs in terms of mental health of these sirens? It's no surprise that urban living is 'found to raise the risk of anxiety disorders and mood disorders by 21% and 39% respectively'. Physical health too: we may well be at the point where, as well as their reducing the quality of life of thousands of citizens every day, these sirens create more accidents than they help ameliorate by disturbing sleep patterns and inducing panicky responses in other road users and members of the public. Again, under the current policymaking regime, it's in nobody's interests to find out.

If government is to intervene in matters of health, it must look at the overall physical and mental health of its citizens. There have been, and no doubt still are, areas in which relationships between cause and effect are easy to identify. Provision of sanitation for instance, is clearly beneficial. I'd also support bans on smoking in all areas where there will be children and adults who don't choose to be exposed to the fumes. (That would be on aesthetic as well as health grounds.) But society is complex, as are the human body and mind. Most scientific relationships aren't easy to identify; and they vary over space and they change with time. We need policies that allow for diverse, adaptive approaches and that target broad mental and physical health, rather than particular maladies.

I offer my suggestion in this essay, which applies the Social Policy Bond principle to health care. Briefly: governments would target for improvement the health of the population, as measured in Quality (or Disability) Adjusted Life Years. Bonds would be redeemed only after sustained periods of improved health. A bond regime would reward the most efficient ways of improving health by channelling society's scarce resources into the areas where they could do the most good. Unlike today's healthcare systems, it wouldn't assume that a one-size-fits-all approach, based on fossilised science, is good enough for everybody, for all time.

Where we're at

How are Social Policy Bonds doing?

It's about 27 years since they first entered the public arena (see here). In that time the Social Policy Bond idea has won praise from distinguished economists (point 8 here), but no Social Policy Bonds have actually been issued.

It's not all doom and gloom though. There is widespread, though belated, recognition now that rewarding better performance in the public sector is a good thing, and non-tradeable variants of Social Policy Bonds are being issued on a trial basis in the UK, US, Australia and Israel. They are also being considered in New Zealand (see here for a short video discussion).These bonds have various names including Social Impact Bonds, Social Benefit Bonds and Pay for Success Bonds.

I have my reservations about them, which I've expressed here and here. Essentially, their being non-tradeable drastically reduces the scope - in breadth and time horizon - of the goals that can be considered. They favour existing service providers, and their administrative costs are likely to be relatively higher than Social Policy Bonds. As well, because they aren't openly traded, they generate no price information that could be extremely useful to policymakers. Nevertheless, these bonds do reward greater efficiency in achieving their limited objectives, and they might well improve on current policy where that is particularly inefficient. They could therefore be a handy (and though I am hesitant to say so, necessary) first step toward a fully-functioning Social Policy Bond.

That said, I'd disappointed that, as far as I am aware, the backers of all the Social Impact Bonds being issued have been, or will be, governments. I'd have much preferred the private sector, in the form of non-governmental organizations or philanthropists, take the lead. The bonds being issued are intended to help vulnerable or disadvantaged people and it seems regrettable in today's climate that the other beneficiaries of taxpayer funds will be financial intermediaries who will take their cut of the transaction costs. Indeed, it seems these bond issuers will benefit whether or not the bonds work as intended. If the bonds do work as intended, that's fine, but we should remember that they are an experiment and, it would be a shame if they come to be seen as a means by which the financial services sector syphons off yet more cash from taxpayers while contributing nothing (at best) to wider society. The other danger is that if Social Impact Bonds fail - and especially if they do so while the brokers benefit - they could discredit the Social Policy Bond concept. For many reasons, therefore, I hope they succeed....

Hidden metrics

Hidden metrics, not to be confused with the hidden variables of quantum mechanics, are one possible way of preventing the gaming that could otherwise result from targeting specific quantitative outcomes.

Let me explain. Say we are targeting female literacy in Pakistan. We issue bonds that will be redeemed when the literacy level of 15-year old girls in Pakistan exceeds 95 percent each year over a ten-year period. Campbell's Law tells us, rightly I think, that:

The more any quantitative social indicator (or even some qualitative indicator) is used for social decision-making, the more subject it will be to corruption pressures and the more apt it will be to distort and corrupt the social processes it is intended to monitor. Source

What metrics could we deploy that measure, accurately and inevitably, that which we want to achieve: namely, near-universal literacy for 15-year old Pakistani girls?  We could subject every girl of that age to a standard reading test, but we know that academic tests and their results, even in developed countries, are manipulated. Girls could be taught to the test (by, for instance, being given the test in advance and taught to memorise it by rote). Or anyone up the ladder from the administrators of the test to the collators of the results could be bribed to alter the data at any stage.

So instead we could take a random sample of, say, 100 schoolgirls from 100 districts around Pakistan at ten yearly intervals, and test them in non-standard ways for their literacy. We don't reveal which districts, let alone which girls, will be chosen; nor do we release exact details of the test. That makes it much harder to game the outcome.

More complex is the goal of peace. Now, nuclear peace is relatively easy to target: we can issue bonds that will not be redeemed, say, until a thirty-year period has elapsed during which there has been no detonation (accidental or not) of a nuclear weapon that kills more than 100 people. But what about peace more generally? Hot wars are fairly easy to recognise and define and so would be correspondingly easy to deter using a bond regime. But there are more nebulous ways of fomenting conflict between states (Ukraine), or of ratcheting up tension to levels that severly curtail quality of life.  In these instances we could use an array of metrics such as: trans-border movement of weapons, numbers of people killed or fleeing their homes. Even the lack of information could also be an indicator of conflict that could find its way into our calculations. We could also use survey data, such as attitudes about potential enemies, or expressions of fear. Other potential indicators are disruptions to food, water or electricity supplies and other results of damage to infrastructure. 
 
The important point is that in this as in other goals, we need not and indeed should not specify in advance exactly which combination of metrics and indicators will be used to determine whether or not the bonds' redemption terms shall be deemed satisfied. In general, we try, as far as possible, to target metrics that are, or that are inextricably linked to, exactly what we want to achieve. Where that is difficult, we try to prevent gaming by not specifying too far in advance the exact redemption terms of the bonds. The aim, at all times, is for bondholders to comply with the spirit as well as the letter of the goals set by the backers of the bonds.

Irrational health funding

This is irrational, but not surprising:

The [US] National Institutes of Health last month published a startling analysis of how it allocates its funding: in 2010, HIV research received nearly $3.1 billion in funding, while a deadly lung disease that has more than six times the health toll in the United States got only $118 million. Two diseases with a similar health burden, breast cancer and chronic liver disease, received wildly different levels of support: $763 million for the cancer best known for iconic pink ribbon awareness efforts, versus $284 million for a disease commonly caused by alcohol abuse. Autism receives more than five times the funding of eating disorders, but their impacts on health, measured in years of disability and premature death, are quite close. Why the diseases that cause the most harm don’t always get the most research money, Carolyn Johnson, washingpost.com, 17 July

I've inveighed against this sort of bias in government policy for years. In health, as in other policy areas, I am sure that government is well meaning and hard working. But it suffers from its inherently uniform, top-down approach. Government can also be short term in its thinking, reactive rather than proactive, and disdainful of innovation while favouring tried, tested but failed approaches. It has to make its resource allocation decisions on the basis of data that are necessarily incomplete. How can it know in detail the effect that spending on, say, cancer diagnostic machinery will have on the overall health of the nation, as compared with subsidising the cost of nicotine chewing gum?

So, by default, health expenditure is influenced by groups of medical specialists with little incentive or capacity to see improvements in the overall health of a large population as an objective. Funding decisions are also heavily influenced by the public profile of a disease or its victims, rather than on what would best meet the needs of society. It’s also a question of diet, exercise, transport, and culture. Recent research shows, for instance, the beneficial effects on health of green spaces in our cities (see here (pdf) for instance). The way government is structured, with its discrete funding bodies, makes it unlikely that such benefits will influence funding decisions.

We cannot expect a government nor any single organization to identify the huge numbers of variables, with all their time lags and interactions, that influence the nation’s health. We can, though, devise a system that rewards people who explore and implement the most cost-effective health solutions, even when circumstances and knowledge are changing continuously. I have tried to do this with my essay on Health Bonds, which would aim to distribute scarce government funds to where they would do most good, as measured by such indicators as Disability Adjusted Life Years. One small caveat though: I'm assuming that most of us - at least in our most rational moments - favour such a distribution of health resources. But there is a possibility that the current (mis)allocation of resources originates in unvoiced but widely held preferences. Ms Johnson quotes one expert: "...we tend to underfund things where we blame the victim". It's unlikely that the large disparities in health funding outlined above do reflect such deliberate choices. But even if they do, it would be better to be explicit about it.