The uninsured, it’s said, use emergency rooms for primary care. That’s expensive and ineffective. Once they’re insured, they’ll have regular doctors. Care will improve; costs will decline. Everyone wins. Great argument. Unfortunately, it’s untrue.  A study by the Robert Wood Johnson Foundation found that the insured accounted for 83 percent of emergency-room visits, reflecting their share of the population. After Massachusetts adopted universal insurance, emergency-room use remained higher than the national average, an Urban Institute study found. More than two-fifths of visits represented non-emergencies. Of those, a majority of adult respondents to a survey said it was “more convenient” to go to the emergency room or they couldn’t “get [a doctor's] appointment as soon as needed.” … Medicare’s introduction in 1966 produced no reduction in mortality; some studies of extensions of Medicaid for children didn’t find gains.

HT Tim Starr.

]]> http://www.overcomingbias.com/2010/03/uninsured-er-fallacy.html/feed 12 http://www.overcomingbias.com/2010/03/why-wash.html http://www.overcomingbias.com/2010/03/why-wash.html#comments Thu, 11 Mar 2010 18:45:26 +0000 Robin Hanson http://www.overcomingbias.com/?p=22185 In 2005, Boston-based doctors published the very first clinical trial of alcohol-based hand sanitizers in homes and enrolled about 300 families with young children in day care. For five months, half the families got free hand sanitizer and a “vigorous hand-hygiene” curriculum. But the spread of respiratory infections in homes didn’t budge, a result that “somewhat surprised” the researchers. A Columbia University study also found no reduction in common infections among inner-city families given free antibacterial hand soap, detergent, and cleaning supplies. The same year, University of Michigan epidemiologist Allison Aiello summarized data on hand hygiene for the FDA and pointed out that three out of four studies showed that alcohol-based hand sanitizers didn’t prevent respiratory infections. Then, in 2008, the Boston group repeated the study—this time in elementary schools—and threw in free Clorox disinfecting wipes for classrooms. Again, the rate of respiratory infections remained unchanged, though the rate of gastrointestinal infections, which are less common than respiratory infections, did fall slightly. Finally, last October, a report ordered by the Public Health Agency of Canada concluded that there is no good evidence that vigorous hand hygiene practices prevent flu transmission. …

In hospitals, outside of these clinical trials, just half of doctors and nurses regularly clean their hands before patient care, despite widespread publicity. More worrisome: In hospitals where massive educational efforts have increased hand-washing rates from 40 percent up to 70 percent, there has been no overall reduction in infection rates. Even in highly regulated places like hospitals, the promising benefits of hand-washing remain largely unrealized.

More here; HT Tyler.

Added 12Mar: Yvian lists may pro-washing studies here.

Bloodletting was used routinely until 1836 when French physician Pierre Louis conducted one of the first clinical trials in medicine.  Louis compared pneumonia patients whom he treated with aggressive bloodletting and those he treated without it.  Louis found that bloodletting was linked to far more deaths. … George Washington, the first president of the United States, … died two days after a doctor treated his sore throat by draining almost five pints of blood.  … A remarkably high percentage of medical decisions still reflect the often-obsolete practices that a doctor learned in medical school, the ingrained traditions of a hospital or region. (p.13) …

What she thought was a straightforward study of how leader and coworker relationships influence errors in eight nursing units. … [She was] flabbergasted when nurse questionnaires showed that the units with the best leadership and best coworker relationships reported making 10 times more errors than the worst. … Better units reported more errors because people felt psychologically safe to do so. …

Nurses whom doctors and administrators saw as most talented unwittingly caused the same mistakes to happen over and over.  These “ideal” nurses quietly adjust to inadequate materials without complaint, silently correct others’ mistakes without confronting error-makers, create the impression that they never fail, and find waits to quietly do the job without questioning flawed practices.  These nurses get sterling evaluations, but their silence and ability disguise and work around problems undermine orgainzational learning.  (pp105,106)

Clearly most med errors are not reported, and docs reward nurses more for covering doc asses than for improving patient outcomes.

To most folks such regulations seem intolerably intrusive.  But when people are asked to justify our common and extensive regulations and subsidies of medicine and education, they often mention exactly this issue – that such interventions make sense because we all have a stake in the work of others via the income taxes those folks pay.  Why the asymmetry?  Why do folks think these arguments make sense regarding medicine and education, but not regarding choice of career or location?

My guess: humans inherited intuitions that the community should have more say in and contribute more to medicine and education.  This is the way our distant ancestors did things in their small nomadic forager bands, and we intuit we should act similarly today.  The stuff about managing our cut of others’ income is just a rationalization.

With [Major League Baseball]’s adoption of mandatory testing for steroids, many thought that home run rates would drop dramatically. They didn’t, and many felt that the lack of a test for HGH could be part of the explanation. Well, it’s time for the scientists working on such a test to start something else more important.

That is John Bradbury.  He interprets:

The illegality of growth hormone actually promotes its use in sports. … The banning of a drug by anti-doping authorities sends a loud and incorrect signal that it works. … Therefore, I believe that legalizing growth hormone is needed to send the signal that it doesn’t work, largely to undo the widespread common belief that growth hormone does improve performance. … Think of the powerful effect it would have if MLB pulled growth hormone off its banned list. I can’t imagine a more powerful signal of a drug’s lack of potency as a performance enhancer. If we are going to be paternalists, let’s be effective paternalists.

Added 5Mar: See also here, HT Tyler.

Fifteen years ago, [Irving Kirsch] began conducting meta-analyses of antidepressant drug trials. … Kirsch’s conclusion is that antidepressants are just fancy placebos. … Drug trials are double-blind: neither the patients (paid volunteers) nor the doctors (also paid) are told which group is getting the drug and which is getting the placebo. But antidepressants have side effects, and sugar pills don’t. Commonly, side effects of antidepressants are tolerable things like nausea, restlessness, dry mouth, and so on. … This means that a patient who experiences minor side effects can conclude that he is taking the drug, and start to feel better.

But after 6000 words of such skepticism, Menand still concludes: take the meds.  Why?  Because impressive authors have written eloquent testimonials:

The recommendation from people who have written about their own depression is, overwhelmingly, Take the meds! It’s the position of Andrew Solomon, in “The Noonday Demon” (2001), a wise and humane book. It’s the position of many of the contributors to “Unholy Ghost” (2001) and “Poets on Prozac” (2008), anthologies of essays by writers about depression. The ones who took medication say that they write much better than they did when they were depressed. William Styron, in his widely read memoir “Darkness Visible” (1990), says that his experience in talk therapy was a damaging waste of time, and that he wishes he had gone straight to the hospital when his depression became severe.

The only reason Menand can imagine resisting such artists is a perverse religious desire to suffer:

What if there were a pill that relieved you of the physical pain of bereavement—sleeplessness, weeping, loss of appetite—without diluting your love for or memory of the dead? Assuming that bereavement “naturally” remits after six months, would you take a pill today that will allow you to feel the way you will be feeling six months from now anyway? Probably most people would say no. … Gerald Klerman once called “pharmacological Calvinism” … the view, which he thought many Americans hold, that shortcuts to happiness are sinful, that happiness is not worth anything unless you have worked for it.

Numbers schmumbers – only uncivilized animals, or religious nuts, would not let eloquent authors soothe their savage doubts, until they accept being comforted by their culture’s conventional ways to show that folks care.

A few points on the insurance status and mortality debate:

— Normally we require overwhelming empirical data to overturn a principle that has strong theoretical support.

— The empirical data to support the “insurance status doesn’t impact mortality” conclusion is not overwhelming. [Matt lists contrary studies] …

— I don’t believe that the people touting these studies really believe them; if widespread beliefs about the desirability of health insurance are totally wrong, this should have dramatic policy implications that should be explored.

Strong theoretical support?!  Here’s what our theories actually say:

1. If people buy something, they expect positive net value from doing so.  If they do not buy something, they expect negative net value from buying.
2. On average people tend to be right about what gives them net value.  Yes we err, but that isn’t typical.  Theory does not suggest large values gained.
3. Insurance is sometimes a good buy, but not always.  Potential insurance customers must weigh moral hazard, admin costs, state-dependent utility, etc.
4. People are usually but hardly always honest about why they buy what they buy.

So standard theory suggests that those who buy insurance expect a net value, but that those who would not choose to buy expect a net loss if forced to buy.  When people are forced to buy insurance but have the option to to use such subsidized services, theory also suggests that users expect to gain thereby, ignoring subsidy costs.

If we ask people why they buy med insurance, or why they use med services, they usually say they buy to gain health, which suggests that those who buy gain more health.  But this theory doesn’t at all suggest that they get a lot of health.  And the data only suggests the effect is weak, not that it is zero.  Also, this expectation of honesty is pretty weak; data can easily overturn it.

Much of our data on the health effect of med insurance is about forced insurance, which people would not have bought for themselves.  Theory does not at all lead us to expect that this will provide a net gain.  We do expect a net gain from using subsidized services, ignoring the subsidy cost, but not a large gain.  And the only theory that says this would be a health gain, as opposed to some other kind of gain, is the honesty theory, that people are right about why they buy things.

Matt, many of us really do believe that the health gains of subsidized med are on average small, and that health gains may well not be our main gains from med services.  We may instead gain the comfort of showing and seeing that folks care.  The main theory this conflicts with is our common belief that folks are not usually mistaken about why they buy what we buy.  But that is hardly “a principle that has strong theoretical support,” right?

I suspect the problem here is that Matt said “theory” when he meant “common sense.”  Yes our society has a common, almost religious, belief that more med is typically very healthy.  But we do not have good theory supporting that belief.

That is Megan at her blog.  More from her in the Atlantic:

Ezra Klein declared that Senator Joseph Lieberman, by refusing to vote for a bill with a public option, was apparently “willing to cause the deaths of hundreds of thousands” of uninsured people. … In the ensuing blogstorm, … few people addressed the question that mattered most: …  If we lost our insurance, would this gargantuan new entitlement really be the only thing standing between us and an early grave?  Perhaps few people were asking, because the question sounds so stupid. Health insurance buys you health care. Health care is supposed to save your life. So if you don’t have someone buying you health care well, you can complete the syllogism. …

The possibility that no one risks death by going without health insurance may be startling, but some research supports it. Richard Kronick of the University of California at San Diego’s Department of Family and Preventive Medicine, an adviser to the Clinton administration, recently published the results of what may be the largest and most comprehensive analysis yet done of the effect of insurance on mortality. He used a sample of more than 600,000, and controlled not only for the standard factors, but for how long the subjects went without insurance, whether their disease was particularly amenable to early intervention, and even whether they lived in a mobile home. In test after test, he found no significantly elevated risk of death among the uninsured. …

The only truly experimental study on health insurance, a randomized study of almost 4,000 subjects done by Rand and concluded in 1982, found that increasing the generosity of people’s health insurance caused them to use more health care, but made almost no difference in their health status. … Analyses of the effect of Medicare, which becomes available to virtually everyone in America at the age of 65, show little benefit. In a recent review of the literature, Helen Levy of the University of Michigan and David Meltzer of the University of Chicago noted that the latest studies of this question “paint a surprisingly consistent picture: Medicare increases consumption of medical care and may modestly improve self-reported health but has no effect on mortality, at least in the short run.” …

We should have had a better handle on the case for expanded coverage—and, more important, the evidence behind it—before we embarked on a year-long debate that divided our house against itself. Certainly, we should have had it before Congress voted on the largest entitlement expansion in 40 years. Unfortunately, most of us forgot to ask a fundamental question, because we were certain we already knew the answer.

Forgot?!  This is no random memory failure.  For many decades health economists have known that the best available evidence shows little or no relation at the margin between med and health.  The health economists advising all the major sides have long known this.  When the data is this noisy, there will always be exceptional studies, and as Megan says, the left prefers to cite exceptions that find more med tied to more health; the right prefers to avoid the issue.

These tactics are far from random accidents; neither side wants to contradict the US public, with their religious-level faith in the healing powers of medicine.  If we were considering a vast new grocery store or car entitlement, the public would hardly “forget” to wonder if that would really give us more nutrition or a faster commute.  But the US public has little religious-style fervor on grocery stores or cars.

How often do you see theists wonder if God is really as good as folks say, or patriots wonder if their nation really deserves their allegiance?  That’s how often you will see the US public question the value of medicine.

The currently popular version is that Africans have too many concurrent (at the same time) long-term partners.  There is some evidence that this happens in African more than elsewhere, and there are theoretical reasons to expect it to speed sex epidemics.  But a December review says the case is far from closed.  From Lancet in October:

A four-city African study actually found lower rates of concurrency in places with larger HIV epidemics, and a study using nationally representative surveys in 22 countries (all but one of which was in Africa) concluded that ‘‘the prevalence of concurrency does not seem correlated with HIV prevalence at the community level or at the country level, neither among women nor among men.’’ Additionally, Wellings and colleagues reviewed global sexual behaviour and could not find sufficient data to assess whether rates of concurrency differ across the world.

The main reply is:

[Critics] offer no credible alternative explanation … It is simply not plausible that serial monogamy by itself could generate the explosive generalised epidemics.

But Karl Smith and David Friedman suggest bad med instead:

Much of the transmission may be due to sloppy medical procedures, in particular the reuse of needles for injections.

In fact, there is a whole journal devoted to this thesis:

Seven years ago the International Journal of STD & AIDS (IJSA) began actively encouraging reexamination of the prevailing view that penile–vaginal sex was driving African HIV epidemics, … Although the IJSA-published dissenting views have largely been ignored, dismissed or fiercely resisted by established HIV researchers and allied health agencies.

A 2007 Annals of Epidemiology paper found:

In regression analyses, nonuse of disable syringes is associated robustly with greater HIV prevalence in all models. … Greater HIV prevalence also is associated with higher Gini Index, less female economic activity, less urbanization, and less percentage of Muslims.

World-wide, resusable needles are the second biggest binary predictor of HIV (after Sub-Saharan African location and before gender-literacy ratio):

Focusing on Sub-Saharan Africa, resusable needles are a huge predictor, as is a U-shaped dependence on Tetanus coverage:

For balance, see a brief critique of this view:

If injections were a major source of transmission, one would expect high rates during childhood, when children receive most preventive vaccinations and many health care visits. Yet a study of mother-child dyads in public health facilities in South Africa found only 1.4% of HIV-positive children aged 2–9 years had HIV-negative mothers. While this study needs to be replicated in other settings, it does not indicate that injections play a major role in transmission among children.  Further, if transmission were primarily due to medical injections, one would expect similar rates of HIV infection among males and females. However, there was a five-fold difference in our dataset (10.6% of adolescent females vs. 2.1% of adolescent males were HIV-positive). Potterat et al present data suggesting teenaged females were more likely to have received recent injections, but their argument is flawed in several ways. … Although antibiotic injections are often used to treat sexually transmitted infections (STI) such as syphilis, in our survey, more male than female adolescents ever sought treatment for an STI from a health worker (48% vs. 29%, respectively).

I’m not saying I’m sure bad docs are a big cause of African AIDS, but the possibility sure seems to deserve more attention than it has been getting.  More related papers here, here, here, here, here, here, here, and here.

Added 10a: For you skeptics, yes African data on cause of death is suspect, but overall death rates are much less so, and Africa has clearly suffered a huge death increase relative to pre-AIDS trends.

Added 14Feb: We often think of anti-racism as core to our culture, but when our choice is to think of blacks as sex perverts or to think of med as deadly, we clearly choose the former over the later.  This suggests heroic med is far more central to our culture than anti-racism.

Kübler-Ross began to work on a book … It came out in 1969, and, shortly afterward, Life published an article about one of her seminars. … Angered by the article and its focus on death, the hospital administrators did not renew her contract. But it didn’t matter. Her book, “On Death and Dying,” became a best-seller. …

Her argument was that patients often knew that they were dying, and preferred to have others acknowledge their situation: “The patient is in the process of losing everything and everybody he loves. If he is allowed to express his sorrow he will find a final acceptance much easier.” And she posited that the dying underwent five stages: denial, anger, bargaining, depression, and acceptance. … Today, Kübler-Ross’s theory is taken as the definitive account of how we grieve.

More here.  Pause to see things from those old docs’ point of view.  While we usually prefer to be honest and forthcoming, we make exceptions.  Some of our reasons are selfish, but we also say that telling people some truths only makes them feel bad, without actually helping much to make decisions.

So isn’t imminent death a great examples of a truth that makes folks feel very bad without much helping decisions?  Look how fiercely people avoid thinking about death when it is only a slight possibility, and how more anxious they get as death becomes a larger possibility.

Sure, most folks say they want to be told the truth about imminent death.  But most folks also say they want to know if their partner is cheating, if their career is tanking, if their neighbors hate them, etc.  If you asked folks straight out, most would even say they want the truth on “do I look fat in this.” So if you are going to hide some type of truth from people for their own good, you must do so in the face of the fact that most folks say they want to be told.

Yes, we may like the closure of taking their time in saying goodbye to folks, but don’t similar modestly useful actions correspond to most truths we think of hiding from folks?  Does this gain so obviously outweigh the terror of knowing you will die soon?  I want to be told about my death, but I’m weird and want the truth on most everything.  What is a better example than imminent death of a truth we’d consider hiding from folks?