Why They Trust

Wednesday, November 18, 2009

Stephen Bourque makes some incisive commentary about a recent change in the government's recommendations on the timing and frequency of breast cancer screening. His post deserves a full read, but one paragraph in particular caught my eye:

It has always amazed me how much trust the general public puts in government recommendations of this sort. The group in this case, the United States Preventive Services Task Force, is characterized as an "independent panel of experts in prevention and private care appointed by the federal Department of Health and Human Services." But what exactly is this group independent of? The implication is that they are independent of individuals and corporations that have a vested interest in the guidelines. However, what the panel is entirely dependent upon for its existence is the federal government, an institution that has absolutely no incentive to meet consumer demands. The panel is independent of responsibility and accountability. [minor format edits]
I suspect that the amazement may be at least partially rhetorical, but it is worth noting where such blind trust originated and considering its full ramifications. The above paragraph reminded me of the following warning from an essay by Alan Greenspan in his better, younger days:
To paraphrase Gresham's Law: bad "protection" drives out good. The attempt to protect the consumer by force undercuts the protection he gets from incentive. First, it undercuts the value of reputation by placing the reputable company on the same basis as the unknown, the newcomer, or the fly-by-nighter. It declares, in effect, that all are equally suspect... Second it grants an automatic (though, in fact, unachievable) guarantee of safety to the products of any company that complies with its arbitrarily set minimum standards. The value of a reputation rested on the fact that it was necessary for the consumers to exercise judgment in the choice of the goods and services they purchased. ... [bold added] ("The Assault on Integrity" in Capitalism: The Unknown Ideal, pp. 119-120)
In other words, over the past few decades, people have become less and less accustomed to acting as their own "consumer watchdogs" even as the government slowly gobbled up larger and larger swaths of the medical and scientific sectors, slowly making the independent advice of scientists to the government less so. Consider this the informational equivalent of the illusory "access" to medical care the Democrats are promising us.

This affects everyone and even confounds the efforts of those of us who are inclined to distrust the government to establish our own opinions on medical matters. For example, at the site Quack Watch is a list of "Twenty-Five Ways to Spot Quacks and Vitamin Pushers." Reading through the list, I noticed that ten of the items on the list (1, 2, 9, 10, 16, 18, 21, 23, 24, and 25) included reliance on the government in some way: e.g., mentioning government nutritional guidelines, alluding to the role of the government in regulating the practice of medicine, or linking to a government web site.

Of course, since the government funds so much research and "educates" so many people about science, the truth is that every single item on the list is affected in some way by government interference in the economy: Both the information under consideration as well as the ability of an average person to evaluate it have often been undermined from the start.

The immediately possible debacle of a government takeover of the medical sector would be a bad enough pit for America to have to have to climb out of, but the truth is that we need only turn around for a moment to see that we are already in another.

-- CAV


Realist Theorist said...

There is another problem here which is very fundamental: there are some things that no other body can decide, on principle. This is not just about a moral usurpation, but about an epistemological impossibility. Let me explain...

These types of recommendations consist of two major steps:

Step 1 lays out the options, their costs and the pros and cons of each

Step 2 choose from among them

Step 2 is impossible without reference to a particular valuer. In absence of a particular valuer, the people making the recommendation assume what they consider an "average" valuer, or they simply decide what a rational valuer would do.

Here is an example: when a woman is pregnant doctors will sometimes recommend a test where amniotic fluid is drawn and tested for certain things. This is done if the woman is older than a certain age, or has other risk factors. The test can catch a condition that will result in a child being mentally disabled. The test costs a little under $1,000.

Now, consider whether a woman ought to get such a test if she is younger than the cut-off age. There is still a very tiny risk that the kid has the condition. What are the cons? There are two: the cost and the fact that the test itself can lead to a miscarriage.

There is no way that anyone other than the end-valuer can make a decision between those options. For instance, one person might say that any risk of miscarriage is worth being sure that one does not have a baby that is mentally disabled; another person may not. One person may be willing to pay a few thousands to be absolutely sure to rule out even the extremely low-probability risks,another may not.

When groups make blanket recommendations, they are not dealing in science (as suspect as their science may be). The more fundamental problem is that they are telling one what values one ought to choose.

Instead, these "recommendation panels" approach the subject like this: the cost of a mammogram is $200, the chance of getting breast-cancer is x%, the cost of fixing it if it is detected is $D and if undetected is $U. So, The additional cost ($U - $D) can be multiplied by the x% to get an "Expected Cost". If this cost is greater than $200, the procedure is not cost-effective! Often, they will add in fictional costs to account for the worry from false-positives and for the non-monetary hassles of ending up with breast-cancer. When they do this, they are assigning values to us, the end-valuer.

Step 2 is not the job of scientists. They're modern day priests / social-engineers masquerading as scientists.

Gus Van Horn said...


Interesting analysis, to which I would add: The state is usurping the role of the value-er, which means that we are treated like state property under such calculations.

That said, I think that in a free society, many of the things the government now does might be done (although from an individual-centered perspective) by panels of scientists working either for private companies (like insurers) or consumer-interest entities (like the Consumers' Union and, I believe, UL). This would be for the simple reason that we can't all be experts at everything. (We would generally know more, but the kinds of boards I am envisioning would have reputations and accountability.)

Finally, all of that said, I don't think that government-issued information is ENTIRELY useless -- at least not yet. I think it will rapidly become such as the state gathers more and more control of the economy and is in the business of ensuring compliance with orders over providing information.


Mo said...

This is just the tip of the iceberg and I believe this was a “trial balloon” to ascertain the public’s reaction to the “less is more” Obamacare ideology embedded within the Senate Healthcare Bill concocted behind closed doors. I know in the first edition of the Baucus Bill there was a section described as a “regional imaging board” that was to oversee imaging studies.

It should be obvious to those who pursue information not forthcoming within MSM that technology that looks inside the body to find abnormalities will be heavily regulated. No need to treat cancers unseen.

Gus Van Horn said...

That's an interesting take, but I could see that change either way (as reflecting the increased cost to "society" or a trial balloon), or both.

Steve D said...

“Finally, all of that said, I don't think that government-issued information is ENTIRELY useless -- at least not yet.”

That’s referring to the opposite trap of course which will also prevent a solid evaluation of the evidence. Not everything from the government is incorrect. While no one can be an ‘expert’ on all subjects we should at least have some knowledge in the various fields which impact our lives (nutrition, economics, biology etc.). At minimum you should be able to recognize when something may require further research.

Here is an example in a quote from a more recent post:

“Though the risks of occasional 'casual' smoking are overblown by just about everyone”

It depends upon what you mean by a casual. Look at the data. The negative health effects of smoking become very minimal below about half a pack a day but you wouldn’t know that by just listening to the more prominent sources of information. They would have you believe that touching a single cigarette is dangerous.

One of my pet peeves is the issue of 2nd hand smoke. So called studies which prove that 2nd hand smoke are dangerous always focus on subjects who have been exposed to large amounts over a long period of time like for example people who have lived with a heavy smoker or worked in a high smoke environment for decades. Then they show a barely significant (or sometimes non significant) result (p=0.1 anyone?) with tiny deltas and use the study as evidence that smoking should be banned in public parks?! The question is if you can barely measure an effect with the heaviest possible exposures what is the effect of casual exposure?

By the way a lot of scientists seem to be susceptible to this trust in propaganda as well. How can a person who is just finished arguing how incorrect the government or media is in his area of expertise, quickly swallow everything they say on other subjects?


Gus Van Horn said...


I never bought second-hand smoke as a serious health hazard, but I was still surprised to learn that half a pack a day isn't terribly dangerous. (But I'll still stick to my once-a-year-or-so cigar: My mouth always feels too nasty the next day for me to want to smoke any time soon again anyway.)

The government's anti-smoking laws are clearly in violation of individual rights, but I have noticed that the problem is compounded by what I can only surmise is a cultural phenomenon: Overblown fear of any and all risk, as if living is not inherently risky.

Make a semi-plausible case that something is risky to most people today and you can gin up a panic out of nothing in no time.


Steve D said...

“Make a semi-plausible case that something is risky to most people today and you can gin up a panic out of nothing in no time.”

Like the H1N1 virus for example? The statistics show it is substantially less dangerous than the seasonal flu yet somehow they were able to make people panic about it. I saw a reporter make a comment that a lot of people were not getting the vaccine because the pandemic (don’t you love that word!) was over. The reporter was derisive of this view but it is correct. Factor in the incubation phase, the time it takes for the vaccine to be effective as well vaccines probable effectiveness (still not really known) and the vaccine is almost useless at this point. In order to figure this out you need to understand a little about how viruses propagate and spread and know some of the actual numbers. Instead of explaining this to his viewers the reporter made the comment that ’We are not out of the woods yet’ with the seasonal flu coming up’ somehow equating the two flues and suggesting by implication that somehow the man was stupid not to get the H1N1 vaccine because of the seasonal flu was about to start?!

“I was still surprised to learn that half a pack a day isn't terribly dangerous”

It’s probably most accurate to state that your chances of having a serious health problem because of a half a pack a day are pretty low (low single digit percentages at most for general health problems and at least an order of magnitude less for lung cancer). The figures always given by the media almost always state the problem in terms of the absolute numbers of people affected (e.g. 20,000 people killed by handgun, flu, automobiles etc.) because it sounds a lot worse that way than the actual percentage. They never mention the millions of people who are NOT harmed by smoking.

“I'll still stick to my once-a-year-or-so cigar”

It would not surprise me if at some point evidence emerges that very light smoking is actually beneficial.

“Overblown fear of any and all risk, as if living is not inherently risky.”

I saw a very interesting figure once which showed the actual risk vs. the perception of risk for a number of different activities and they were strongly inversely correlated. The most risky things (like driving) were perceived to be low risk while the least risky things (like living next to a nuclear power plant) were perceived to have the highest risk. I am not entirely sure all the causes of this phenomenon although a necessary condition must be bad epistemology.


Gus Van Horn said...


This is very interesting and it ties in to a question related to some issues I have been pondering lately. (Off and on, I have been considering risk assessment as it pertains to moral decisions.) Since you appear to know a lot more about this type of information than I do, you might be able so save me some leg work.

I've barely looked for this yet, but as you note with smoking, the frequency and amount of smoking (as well as plenty of other aspects of the individuals concerned) are important in assessing its risk/estimating its danger through statistical methods. I see that, and I know that if I looked around in the smoking literature, I could probably find this information, for smoking anyway.

But people use information like this about other things, too, all the time (e.g., life insurance companies making actuarial tables). It would seem to me that this kind of information is available in a more accessible form somewhere. If I am right, where?

Incidentally, this is one area that a free market in medicine would probably clean things up A LOT. Life insurance companies have a vested interest in making a profit from willing customers, rather than avoiding any and all risk from "misbehaving" dependents. (e.g., as at the the end of this.) A life insurance company will thus not fan the flames of fear when it uncovers a risk, but try to rationally assess it, given that some people will elect to take it willy-nilly.

"Like the H1N1 virus for example?"

From what I've picked up from the news, the risk posed by that illness is WAY overblown for the population at large, but perhaps very high for certain ethnic groups, like Amerindians, with certain genetic variations in their immune systems.

Of course, the sum total of all this reporting (and socialized medicine) is that there is a vaccine shortage. I'm not losing sleep over this, but I have a substantial amount of Amerindian heritage and would have taken the vaccine long ago if I could have. But since everyone's getting it through government controlled channels, I get to wait until December.

"pandemic (don’t you love that word!)"

Yeah. Let's scare everyone by employing a technical term that is correct -- in the technical context of epidemiology -- by using it inappropriately in a general context.


Steve D said...


Most of my information comes from reading the scientific literature or following links from the internet or MSM back to the original studies. I haven’t yet found a single good source of information. Over the years, I’ve looked at a number of reviews/papers but in a rather piecemeal fashion. I am a biochemist so this not my major interest or expertise. I’ve looked into areas of minor interest to myself such as epidemiology, hazards, energy production, nutrition etc. so in these areas I would say I am well informed but not an expert.

The mortality tables which I’ve seen tend to lump all the smokers together or into a small number of easily definable categories. This skews the risk assessment for people who smoke less than or greater than the average. This is perhaps good enough for a life insurance company but not really precise enough for an individual to make a quality life decision.

The bottom line is that I am not sure that the precision of information you want is available anywhere other than the scientific literature. Even with scientific studies at least some effort has to be put into understanding the quality and methods. There are certainly journals which focus on review or metastudies. I would probably start there - I don’t have a list of journals handy but I could probably generate this pretty easily.

A lot of studies I’ve seen focus on the heaviest exposures. I guess its easier to see the effects at these exposures and this also has the additional effect of making the situation seem worse than it actual is. I guess you can panic more people that way.

One thing to point out about the smoking tables though is when you see a statistic such as smokers are 10 times more likely to get lung cancer it is important to keep in mind that the overall risk of lung cancer is very low anyway so that in the final analysis even the risk for smokers is still pretty low.

Also I always try to look at the overall picture when accessing risk. People often look at the risk in isolation and that just is not sufficient to make a good decision. An example which comes to mind is nuclear energy. I am very pro nuclear and I often get into argument with people about this. They will often start by listing all the bad effects of using nuclear energy. My strategy is to try to show them how in fact in each case nuclear energy provides an advantage over the other forms of energy (cost, safety, environment, reserves etc.). My point is that EVERY form of energy will have problems , nothing is or ever will be perfect , so lets evaluate them by comparison and make the best choice. Using this type of comparative analysis smoking, compare the risk of casual smoking with other activities? I am not sure smoking always loses.

Steve D said...


“But since everyone’s getting it through government controlled channels, I get to wait until December.”

The chance of getting a serious case of the flu now are very low since the pandemic has already been on the decline for several weeks (also you have to add in the incubation period). Most of the population has already been exposed and this will be even more so by December. Once the level of people exposed and therefore mostly immune reaches a certain level it becomes very difficult for the virus to spread. Also the vaccine has not been fully safety or efficacy tested (it was super fast tracked) so we have to extrapolate from the data for other similar flu vaccines to get an estimate on how well it will work or on how safe it is. (My best guess is pretty safe and somewhat more effective than the seasonal flu virus since they are targeting only one virus type). My son (8yrs) got the vaccine but only because they were set up at the school when we were there and we only had to wait 5 minutes so there was virtually no cost on our part for a minor benefit. I am not likely going to get the vaccine myself.

If you think you may be more susceptible to this flu or to getting a more serious case of it though for whatever reason your calculation may change and our decision should reflect this. Come to think of it, this is another reason why we need to be careful of public sources of information and make our own decisions. - all these factors vary amongst people.

It’s a good thing this is only the swine flu though. They finally got the vaccine distributed after the pandemic was mostly over! Can you imagine what would have happened if it had been an extremely virulent and deadly pathogen? The response would have been too late and we would have been in serious trouble.

Yes, there are people who get seriously ill from H1N1 and generally these people have other medical issues. Very few (if any) healthy people are dying from this. I don’t know of any specific case where someone with no complications has died although this has probably happened.

A quote from your previous post on H1N1:

“If we have to have influenza, I would clearly chose novel H1N1.”

I agree. I haven’t had a case of influenza since I was a child but if I had to have one I would definitely chose H1N1 over the seasonal flu.


Gus Van Horn said...

"I haven’t yet found a single good source of information."

It could well be that there isn't, too. The ways smoking affects people differs than the ways diet or hygiene do, so it could well be up to the end-user of the information to tabulate it, and may make only relatively crude comparisons possible.

No big deal. This discussion has helped me clarify my thinking about how the kind of issue I have become interested in needs dealing with.

"The bottom line is that I am not sure that the precision of information you want is available anywhere other than the scientific literature."

I don't think precision is necessarily important, so much as being able to make the following types of statements:

"One thing to point out about the smoking tables though is when you see a statistic such as smokers are 10 times more likely to get lung cancer it is important to keep in mind that the overall risk of lung cancer is very low anyway so that in the final analysis even the risk for smokers is still pretty low."

... and ...

"Using this type of comparative analysis smoking, compare the risk of casual smoking with other activities? I am not sure smoking always loses."

Agreed, and that's exactly what I am thinking of.

Smoking seems to be a sort of "gold standard" of unhealthy voluntary activity, and it would be very nice to be able to compare, say, the heightened risk of lung cancer due to smoking to other types of heightened risks, such as that of dying in a car accident or of something else due to chronic exposure to a given level of a given type of radiation.

I have a gut feeling that many such risks are actually both quite low and quite overblown, but gut feelings aren't very persuasive. I'd really like to be eventually be able to say something like, "Lifelong use of umptysquat is generally about twice as dangerous as smoking for the population at large." (As I write this, I now think I see another reason WHY such tables likely don't exist: The big problem with smoking is heightened risk of lung cancer while that of umptysquat might be cardiac arrest.) The comparisons I want to make are certainly possible, but will take work.

(All that said, as I noted earlier in my own personal "example" with the flu, people are individuals and a generalized risk is only a starting point for making one's own quality of life decisions.)

As an interesting aside. I once worked in nuclear power, so I am quite familiar with real, but manageable, yet overblown risks.

In any event, thanks for a very interesting and helpful discussion!

Gus Van Horn said...


Just spotted the continuation in the comment queue.

I agreed that I'd rather have the H1N1 until I saw the data on Canadian aboriginals. Now, I'm not so sure. Unless I start hearing about problems with that vaccine, I'm getting it as soon as I can.


Steve D. said...


"I agreed that I'd rather have the H1N1 until I saw the data on Canadian aboriginals."

The first thing that came to my mind when I read the article on the Canadian aboriginals was the comparative question: I wonder what their statistics are for the seasonal flu and other flu-like illnesses. All sorts of related questions popped up.

There was no real statistics in that article so its hard to assess the real risk. It sounds like there really is not enough information yet to make any definite conclusions. The hypothesis regarding HLA seems reasonable, though.

I do know that percentage of cases which become very serious is much lower than 1% (more like about 0.3% or so). They said lower than 1% in the article so technically that was correct. Also, this is the percentage based on cases where H1N1 is confirmed - in the majority of cases swine flu is so mild most people don’t even know they have it.

Still the article seemed to be designed to emphasize the bad points.

“I have a gut feeling that many such risks are actually both quite low and quite overblown”

When I compare risks I see that minor risks are treated as major and vice versa. Therefore, at least relatively, some risks ARE overblown. By the same token some risks must be understated - the question is why. What is the reason the risks are treated differently? Automobile accidents are one of the most common causes of injury and death yet the major point against them for most people is that they burn carbon.


Gus Van Horn said...


And so we return to one of the points you originally brought up, about high (actual) risks being rated as low and low (actual) risks being rated high.

I think PART of the problem is related to the normal fear of the unknown. We drive cars all the time and we have some grasp of the inherent hazards. (And driving is something of a special case: We have some personal control over its safety, although statistically speaking, maybe not as much as we'd think.) But what about the nuke plant next door? We (i.e., in the general sense of the public) know nothing about radiation, or what is or isn't being done to prevent a catastrophe. There is NO familiarity AND what we do know is very, very ugly sounding. We don't know that our plants are designed to fail safe, but we do know that high doses of radiation can cause cancer. We don't know that the radiation from the nuclear fuel is heavily shielded or that it drops off in intensity with the square of distance, but we do know that people in Japan died horrible, horrible deaths from the atom bomb.

Many technological hazards seem to be a sort of modern-day equivalent of a malevolent "God of the gaps," if you will. A natural, knee-jerk type of reaction is to exercise what control one can: To get rid of the evil, scary technology completely.


Steve D said...


I agree that unfamiliarity is part of the problem.

“But what about the nuke plant next door?”

That’s a good example. I grew up in Deep River, Ontario which is right next door to Canada’s main experimental nuclear reactors. (Chalk River Nuclear Laboratories). My father was a nuclear engineer before he retired. (one reason I am fairly well informed about this subject). Since most people were related to nuclear engineers and physicists people were very familiar with nuclear power and very pro nuclear. When they had a country wide vote to find somewhere to store the spent fuel our county was the only one in the entire country to vote in favor of storing it locally. So familiarity is important.

I don’t think that is the entire problem though. I think another problem is epistemology. When I come across something I don‘t understand and it is interesting or seems important I educate myself. I don’t sit there in a fog. I certainly don’t pretend to know more than I do. How many people today could describe how a vacuum, freezer, internal combustion engine, television or computer works? Probably not that many.

Even with scientists I often see a surprising amount of ignorance in areas just a little bit outside their area of expertise. Further a field, especially topics like economics and history are even less well understood. I could list pages of examples and probably you could as well.

The other side of the issue is not only are people unfamiliar but that most do not put in any effort outside their narrow fields to become familiar - didn’t Ayn Rand talk about blindness vs. the refusal to see?

This is a wider issue than just risk assessment. This may be the one of the crucial issues hampering Objectivism from becoming widely accepted. I am not sure how to combat it though.

As an aside I meant to mention in the last post that I think that nuclear fission is at the present the best solution to our energy problems. I admit I may be somewhat biased but as the technology stands now we could with reasonable safety, efficiency and cost provide all the energy we require. 4th generation reactors will probably be available in 15-20 years and they will eliminate the issue of nuclear waste. Hydro is almost used up so the only other real solution is fossil fuels. They have their own problems regardless of whether AGW is occurring (not yet I think). Yet, I’ve seen entire essays on energy solutions which do not even mention nuclear!


Gus Van Horn said...

"The other side of the issue is not only are people unfamiliar but that most do not put in any effort outside their narrow fields to become familiar - didn’t Ayn Rand talk about blindness vs. the refusal to see?"

This is the crux of the issue. We can't be specialists at everything, but we do have to open to the idea that we may need to look at something more deeply or for the first time from time to time.

How do we know when to go that extra mile versus not wasting a whole bunch of time on something that really isn't that important?

That's a very interesting question, and the beginning of the answer is: How good is the indication that my knowledge about something may be deficient?

Steve D said...

So, we don’t really NEED to know everything. If we can tell when and when our knowledge is deficient and the issue becomes a topic of importance then we can research the topic until we understand it to the level we need make our decision.

We need to know enough about a subject not to understand it completely but at least to know that our knowledge is deficient.

I guess this means we need to have some basic familiarity with all important subjects. At least to the point where we can ask intelligent questions.

Gus Van Horn said...

Can't argue with that.

RE: Chalk River. As in CRUD. Heh!