by Carl V Phillips
I recently had the opportunity to give a talk at what was basically the wake for the end of the quarter-century run of the wonderful Robert Wood Johnson Foundation Scholars in Health Policy Research program at the University of Michigan. I chose to put together some themes from my work as a tribute to one of the goals of that program, bringing the thinking of serious social scientists into health policy arenas where it is desperately lacking. Alas, most of my fellow alumni focus on engineering a better medical system or medical financing, with few choosing to try to deal with public health (let alone “public health”). Medical practice is obviously extremely important, but not so desperately in need of imported thinkers. Well, at least you have me.
I got some great feedback on this talk making that alone well worth my effort. (Thanks to all my colleagues. And it was great seeing you. We’ll be in touch.) But I wanted to also share what I created more broadly here. The following are my slides from the talk, with some text to explain what is not fully contained in the slides, along with a bit of extra material that was not in the talk.
You may recall a few months ago the kerfuffle that was created when CDC told all women who were capable of getting pregnant that they should not let a drop of alcohol pass their lips, and instructed medics to badger them about this. Not only was this a far more stringent recommendation than is supported by the science, but it clearly demonstrated the lack of recognition of any human preference other than maximizing health outcomes.
This generated a lot of bad press, but really only because it was politically incorrect. If the target of this aggressive shaming and disregard for humanity had been men — let alone tobacco product users — it probably would have passed without a peep of pushback from the chatter press. As it usually does…
While the bullying is concentrated in the “public health” faction, the failure to recognize real human preferences is present throughout public health. Consider the Zika recommendations:
“So how long should we delay pregnancy? Is there an end to this in sight? Should we go out and contract Zika now so that we can get pregnant later, after we are mostly immune to a new acute infection?”
“We have no idea. We’ll get back to you. Just delay pregnancy until you hear from us. What’s the big deal?”
That last bit is critical, so it is worth emphasizing. It unethical to use government and quasi-governmental force in pursuit of a special-interest goal that is not shared by most people. It is insane to assume that everyone shares that goal and base your “science” on that assumption.
For those not familiar, the latter point is a true story. Public health types genuinely insisted that people would not change their behavior as a result of HIV treatment improvements, and expressed shock when the inevitable change occurred. And these were not the “public health” people whose rhetoric might be explained by pure special-interest posturing. The innumeracy is not limited to them. These are the same infectious disease people, or their mentors from the previous generation, who are now suggesting delaying pregnancy in the face of Zika because technological change will (we hope) lower the risk soon, at which point the behavior can be more safely resumed. The mind boggles at the contradictions. The analogy to lowering the health costs of tobacco use is nearly perfect.
Readers of this blog will recall the previous post, about the direct harm inflicted by “denormalizing” campaigns and other efforts to punish smokers and other tobacco product users. Though I did not address the point there, the blithe acceptance of such abusive policies can be seen as an outgrowth of this same innumeracy. Recall also my analysis, in the “Why is there anti-THR?” series, of how the dream of a tobacco-free world is a major motivator for anti-THR efforts. While the goal is motivated by a special-interest political preferences, the innumerate failure to recognize that it is an absurd fantasy is critical for perpetuating the dishonesty, wasted resources, and massive loss of welfare that result from that goal.
I have discussed extensively on this page how the FDA’s “MRTP” rules, if taken at face value, mean that no product or claim can ever be approved. Simple economics tells us that of course there is going to be someone out there in our huge population who is right below the margin for choosing to use tobacco products, but will consume a new lower-risk product or an existing product upon learning it is low risk. Thus, even setting aside the fact that both a time machine and unimaginably precise research methods would be needed to support the mandated claim, simple economics tells us the condition will never be met. I have speculated that this might be tactics rather than innumeracy on the part of its architects: FDA’s decisions about tobacco products have all been arbitrary rather than process-driven, and this condition lets them refuse any application they want to refuse. But even if this is true, the fact that they can get away with such a patent absurdity is still a product of the pervasive innumeracy.
For those who may not know, the first point refers to the economic principle of “revealed preference”, in which we (for good reason) believe that the best evidence of preferences can be found in actual choices. Talk is cheap. Public health research methods that ostensibly translate cheap talk into assessments of preferences are utter jokes.
I have developed the “demonic possession” concept over time on this page and elsewhere. For those wanting a specific reference, see here.
Live link to our paper that explores the second-order preferences concepts and some of the other points in the slide.
To expand in the concept in the footnote: There is nothing in epidemiology that is not specific to a particular population and time. Any reference to an epidemiology results needs to be presented in the context of those variables, and any suggestion that a result is a constant across people and time is ludicrous, as Burstyn and I noted in this lesson. Any assessment of behavior or preference is even more subject to change across time and space, as well as based on exactly how the question is posed. Yet public health authors make claims as if one survey result is a universal constant (including when, as in this case, it is clearly not an accurate measure). This is not economic innumeracy specifically, but rather an innumeracy within public health of even its own core sciences.
The first point there is quite important, and I perhaps should have given it a full slide. Undoubtedly some smokers have a preference pattern such that smoking is their preferred choice (as demonstrated by their actual choice) and abstinence is next-best, beating out even their most preferred low-risk alternative. It might simply be they really dislike all of the alternatives, or it might be there is a dynamic such that while they prefer to continue to smoke, if the continuity of their behavior were broken they would be happiest not consuming any product. But for most preference patterns, whatever someone’s tradeoff among health, the consumption benefits, and other factors, if smoking (with its great health costs) is preferred to abstinence, abstinence will be inferior to one or more low-risk alternatives, since their health costs are indistinguishable from abstinence, but they keep some of the smoker’s benefits from smoking.
The first example is common, though it may not be obvious it is a case of economic innumeracy; just because flavor X is demonstrably preferred to flavor Y, it obviously does not mean that abstinence is preferred to flavor Y (which is implicit in the claim that flavor X is causing consumption). The second example is one of the dominant myths of tobacco control, and is frequently repeated by vaping advocates. Smoking clearly saves the rest of society resources (money) despite the rationalization for punitive taxes that it is costly. For medical care alone it is so close to a wash that it is impossible to distinguish from the null (dying from a smoking-caused disease means not consuming any subsequent medical care), but if you add in other foregone consumption the health impacts clearly create a net cost savings. The third example will be familiar to those who follow the UK’s nanny state follies, because this is currently being touted by government officials and their sock puppets as a reason for avoiding moderate drinking.
The fourth example requires some explanation if you have not read the paper or the popular press coverage. This is one of my favorite recent examples. The typical attitude from public health types is that poor people are just too stupid to make good decisions for themselves. So when they feed their children less-than-healthy food, it is obviously due to their ignorance or irrationality. After all, insist the public health types, a healthy diet is just as affordable as an unhealthy one, and so the plebes just need some convincing at the hands of Whole Foods shoppers. Enter Ms. Daniel, who actually did some research on real human motivations and discovered that people facing genuine food insecurity — who are desperately worried about what little food they can afford going to waste — do not dare try something new and different. They know intuitively what research also shows, that getting a kid to try a new food typically takes a half-dozen tries, during which a lot of food is wasted. All parents know that, but most of us do not give the waste a second thought. Some people do not have that luxury, so they stick with the chicken nuggets that they know will be eaten.
There has been a decades-long running joke about tensions between the economists and sociologists in the RWJF Scholars program. I noted in the talk, only half jokingly, that when a sociology student (Ms. Daniel) can point out that you are being economically innumerate — as with public health assuming utter irrationality for a repeated and very important decision, rather than seeking a rational explanation for the behavior — you know you are really doing something wrong. (See also: demonic possession and tobacco use.)
Yes, the talk was given on April 1, giving me a perfect opportunity to make this point, which I have made here dozens of times (the slide was animated so only the title appeared at first). “Public health” people consistently write whatever political conclusion they personally support and attach it to every research paper, even if it is no way supported by (or even flatly contradicted by) the study results. But even apart from that, all manner of public health people almost always tack on a call for some half-assed policy intervention to every research report. This demonstrates a very deep economic innumeracy, failing to recognize that every action has costs that must be identified weighed against the benefits. Moreover, many proposed actions will not even produce the benefits (i.e., solve the problem the research purportedly demonstrates) that they kinda sorta sound like they will produce. Recall my policy analysis (or the shorter blog post covering some of the points) of the FDA e-cigarette deeming regulation, which showed the policy will probably not further any of its own stated goals, and will clearly be counterproductive for many of them.
Of course “teach public health people some economics” is an intervention, though obviously not one that I just thoughtlessly concocted. It has no apparent downsides when measured in terms of bettering the world. I both thank and curse the Scholars program, and my mentors and colleagues from it, for helping me become the one trying to do that.