The Becker-Posner Blog

The Overdiagnosis Problem—Posner

A book published recently and entitled Overdiagnosed: Making People Sick in the Pursuit of Health, by three reputable physicians (H. Gilbert Welch, Lisa M. Schwartz, and Steve Woloshin), argues forcefully that the nation is spending too much money on preventive care. This is doubtless regarded as heresy in some circles: the orthodox view is that prevention is the key to economizing on the expenses of health care: “an ounce of prevention is worth a pound of cure.” The recent health care reform act seeks to promote preventive care. 

Preventive care does reduce health costs in some cases, but not in all, and maybe not in most. The costs of prevention have to be weighed along with the benefits. And private and social costs have to be distinguished. Subsidy programs such as Medicare reduce the private costs of medical treatment to patients, but the social costs are not reduced; their incidence is merely shifted. 

Generally, preventive care has two phases: screening and treatment. The former might seem inexpensive, both in monetary cost and in risk to health, but is not, and for two reasons: the number of people who do not have a condition that is screened for invariably greatly exceeds the number of people who have the condition, so that the cumulative costs of screening are high. And screening creates anxiety, both anxiety over the outcome and anxiety over what to do if the test for the disease in question is positive. An example is the blood test for prostate cancer. It turns out that a huge percentage of men have prostate cancer, but that most of the cancers are benign. The treatments have serious side effects, so for many (especially for elderly) men diagnosed with prostate cancer it is uncertain what the best course of action is. Another example of dubious preventive care is the treatment of mildly elevated blood pressure: blood pressure medicine has to be taken daily and of course must be paid for by someone, and has side effects though less serious ones than prostate treatments, while the benefits in reducing the risk of heart attacks or strokes are modest (unlike the case of highly elevated blood pressure). There are many other examples in which the net benefits of screening for medical conditions followed by treatment if the results of the test for the disease are positive are slight or negative. 

The tendency has been to move the goalposts: to screen for lesser and lesser abnormalities, even though the lesser the abnormality the lesser the expected disease cost to the patient and so the less likely the screening and follow-up treatment are to provide net benefits. Moreover, mild abnormalities are far more common than severe ones, so that moving the goalposts greatly increases the number of persons who have to be screened. When the threshold for excessive cholesterol was lowered from 240 to 200, the number of Americans with excessive cholesterol increased by almost 43 million and all of them are recommended to take drugs to reduce their cholesterol, even though the benefits for persons who are not at high risk of heart disease for other reasons are highly uncertain—yet many of these persons are taking the drugs along with persons who can anticipate a significant benefit. The increased prevalence of screening and preventive treatment has increased the health awareness of Americans and by doing so has increased the innate anxiety that people feel about sickness and mortality. 

Ordinarily we don’t question people’s consumption choices; and it might seem to follow that if people want to take, say, blood pressure medicine to prevent mild hypertension they should be assumed to be maximizing their utility and we should let them alone. But there are reasons to think that screening and treatment of persons who flunk screening tests are excessive from the standpoint of overall social welfare—that aggregate utility would be increased by reallocating many of the resources now used for screening and preventive treatments to other activities. 

We can identify these reasons by considering the full range of factors, other than cost-benefit analyses that support particular forms of screening and preventive treatment. These factors are the incentives of medical researchers (many subsidized by government), health care providers (importantly including pharmaceutical manufacturers), medical malpractice lawyers, American cultural attitudes, our democratic political system, and patients who do not pay the full costs of their medical care. Advances in medical research enable more abnormalities to be discovered sooner—the PSA test for possible prostate cancer is an example—and to be treated. Physicians and other health care providers have an incentive to increase the demand for their services by creating new screening procedures and preventive treatments, although to the extent that preventive care does improve health (as much of it does), acute-care health providers face reduced demand for their services. But apart from dentistry, it is hard to think of areas of health in which preventive care has reduced the overall demand for treatment. 

Although preventive care sometimes involves surgery, as in the case of prostate cancer and other cancers that may be benign, usually it involves treatment with drugs, and thus is strongly promoted by the pharmaceutical industry, often by advertising directly to the consumer. 

Fear of medical malpractice drives physicians to order tests for low-probability conditions, lest they be sued for failure to diagnose a treatable condition. 

Distinctive features of American culture include a strong commitment to business models of economic activity, a high correlation between income and prestige, competitive drive, and a rejection of fatalism. The medical profession, like the legal profession, has embraced a business as distinct from a professional model of service. In a business model, success is measured by profit. Physicians embrace opportunities for increasing their incomes by increasing the demand for their services. 

Americans value longevity not only for the utility that additional years of life confer regardless of how long others live, but as a field of competition: prestige attaches to beating one’s contemporaries in the race to live as long as possible. And this turns out to be for many people a very cheap competition because other people are paying for their medical treatments. The subsidization of the old by the young in the Medicare program increases the demand for screening and preventive care by a politically prepotent voting bloc that has been able to shift most of its medical costs to others. Legal restrictions on exclusions in health insurance policies, and the tax subsidy of employer-provided health benefits, create further gulfs between the costs of medical care to particular persons and what they pay for it. 

So not only is there compelling evidence of what Welch and his coauthors call overdiagnosis; there are good reasons to believe the evidence because the incentive structure for screening and preventive care makes overdiagnosis a theoretical prediction as well as an empirical reality.

How Beneficial is Widespread Medical Screening? Becker

 

 

The United States along with many other rich countries encourages widespread so-called preventive screening for various diseases, including breast, prostate, and other cancers, cardiovascular disease, diabetes, kidney functioning, and many other serious, and even not so serious, diseases. It is often claimed that prevention of diseases should be given the highest priority in order to early detect treatable serious medical problems. But these gains have to be balanced against the considerable cost of extensive preventive screening. The US and other nations are very likely over-screening for various medical problems.

The benefits of screening depend on the seriousness of the disease, and the gain from early detection. The latter in turn depend crucially on the treatment success when a disease is detected early compared to when it is detected at a later stage. For example, treatments for both prostate and breast cancers seem to be more effective when these diseases are detected early, although that is disputed by reputable researchers and physicians. However, it is clear that early detection of Alzheimer’s’ disease has little advantage since there are no effective treatments for this highly debilitating disease. Early detection of this disease may have other advantages if that enables individuals to plan for their increasing mental feebleness, such as preparing or modifying wills, and takng other decisions that require some mental acuity. On the other hand, it may reduce quality of life by causing great fear of the impending disease.

Along with the benefits often come major costs of widespread screening, costs that frequently outweigh the advantages. For example, if only a very small fraction of the population were likely to have a disease, perhaps one chance in 500 or less, it may not make much sense to screen everyone. Take, for example, Huntington’s disease, which affects a very small fraction of the population. With such a low general risk of having the genes that cause this disease, it is much more cost efficient to screen only individuals with a much higher probability of having the disease. These are persons who have family members with Huntington’s. Even then, studies show that many potential carriers of the deadly genes prefer not to be tested since no treatments are available, and they fear living with the news that they will eventually get this disease.

More generally, worry about test results is sometimes a sizable cost of widespread screening. This is partly because many medical tests frequently give false positives that may lead to unnecessary, worrisome, and sometimes risky additional tests and treatments before it is determined that nothing is really wrong. The widely used PSA test to screen for prostate cancer is a good example. Elevated or rapidly rising PSA levels are often indications of tumor in the prostate. However, many individuals have relatively high PSA levels without any tumor presence, sometimes because they have enlarged prostates. Others have rising PSA levels also for other reasons. Annual PSA tests may be indicated for men over age 45 or so with family histories of prostate cancer, but they are less frequently needed for men who have little prostate cancer in their family.

Many tests sometimes lead to serious complications, and any complications should be balanced against the gains from early detection of medical problems. Every blood test carries some risk of infection and contamination. X-rays, MRIs, CT scans, ultrasounds and the like have radiation and other risks. Probing for tumors with biopsies, colonoscopies, and other invasive tests carry a variety of risks.

The benefit-cost evaluation of any suggested widespread screening would depend on the net effect of the factors I discussed. First of all, benefits depend on the size of the gain in treatment from early detection of a disease or other medical problems. This gain has to be adjusted for the probability of finding a serious problem since the lower that probability, the less the expected gain from extensive screening. Costs start with the cost of the methods used to screen, including the value of the time and inconvenience of individuals being screened. The likelihood of various side effects from these methods also must be weighed. The fear engendered by false positives can be important, although that has to be balanced against the peace of mind that comes from getting  (true) negative results.

Preventive care is often considered the gold standard of good medical practice. It is easy to get both doctors and patients to agree when either the government (i.e., taxpayers) or insurance companies (i.e., others who are insured) pay the vast majority of the costs. A rethinking of this standard is warranted, and such a rethinking would be encouraged if individuals paid a much larger fraction of their out of pocket medical expenses.

What Kind of Gun Control? Becker

 

 

The serious wounding of U.S. Representative Giffords and killing of 6 others this January in Tucson, Arizona by a young man using a semi-automatic pistol naturally stimulated considerable anguish. As in all other mass shooting rampages there were also immediate calls for greater gun control, so that guns could not get into the hands of individuals who might use them to kill many innocent victims. In this piece I will consider how successful gun control can be, and the best ways to implement any controls.

This would probably be a safer and better world if no civilians had any guns, aside from policemen, and perhaps some sportsmen, since guns are involved in the majority of murders, at least in the United States. I say “probably” because in such a world criminals would turn to knives, baseball bats, tire chains, even grenades on a very small-scale, and still other weapons. Potential victims, such as shopkeepers and residents of crime-ridden neighborhoods, would in self-defense also acquire similar weapons in order to defend themselves. Nevertheless, since guns are far more lethal than most other weapons, the number of deaths from crime and senseless violence would likely significantly decrease if neither criminals nor victims had access to guns. The total amount of crime would also tend to decline.

Unfortunately, there is no feasible way, certainly not in the United States, to go from the present world to a world without guns. It is estimated that some 60 million Americans own about 200 million guns. This implies more than one gun per American household. Naturally, such an average conceals enormous variation across households and communities. Gun ownership is uncommon in suburbs like Scarsdale and Winnetka, while most households have guns in the inner cities of major cities, like Chicago, Detroit, or Los Angeles.

It is not simply the immense number of guns that makes gun control so difficult, but also the fact that the great majority is illegal and not officially registered. Moreover, the supply of illegal guns is flexible and can be readily expanded as demand increases. Well-organized criminal gangs own the greatest number and have the most sophisticated types of guns. Indeed, drug dealer enterprises go far beyond guns to own explosives, rockets, and other weapons capable of large-scale killings. It is said that the drug cartels of Mexico have weapons that are usually more lethal than those available to most of the local police forces that are fighting the cartels.

In trying to reduce the number of guns in circulation, American states and cities can continue to tighten up on the legal ownership of guns by making the permissible reasons more stringent, such as a shopkeeper in a dangerous neighborhood, by increasing the background checks on applicants for guns to discover whether they have been convicted of crimes or have a history of violence, by requiring longer cooling off period before applicants can take possession of guns, and in many other ways. Undoubtedly, this would reduce the legal ownership of guns, and probably also total gun ownership.

However, despite some dramatic exceptions, the great majority of persons who own guns legally do not intend any criminal actions, nor are they likely to gun down innocent victims. So tightening legal gun ownership will do little to reduce the number of guns in the hands of criminals and unstable individuals. Indeed, it could increase the number held by would-be or actual criminals since the supply of guns available in the illegal market would increase, at least initially, as some of the guns that are pushed out of the legal market by more stringent controls would migrate to the illegal sector. While some criminals may decide they no longer need guns since victims would be less likely to have them, others who would not have used guns before might now decide that guns would give them a greater advantage in attempted robberies.

The most effective way to reduce the number of guns in the hands of criminals without reducing the number of guns legally owned is to punish persons who own guns illegally and those who use guns when committing crimes. Many jurisdictions already punish more heavily individuals who use guns while committing crimes, but it may be necessary to make that additional punishment more severe. The expectation that punishments will be severe to apprehended criminals who had used guns in their crimes will induce some criminals either to use less lethal weapons, or to go out of the criminal business entirely.

Punishing illegal possession of guns is also common. Here, however, a distinction should be (and is often) made between possessors who appear likely to either have committed or will commit crimes, and those who are clearly possessing guns illegally because they live in dangerous neighborhoods, or run shops that may be held up. The former deserve serious punishments, while the latter groups should be lightly punished.

So overall I do not believe that making the legal ownership of guns more difficult is likely to do much good, and could be harmful. I do see more promise in punishing illegal gun possession, and especially punishing severely persons who use guns to commit crimes.

 

Gun Control—Posner’s Comment

There are two important lessons that can be drawn from Becker’s discussion of gun control. The first is that a problem that is not dealt with in its early stages may become insoluble. It is not only the sheer infeasibility of removing 200 million guns from the American population, but also the emergence of a gun culture, that has ended hopes of disarming the population. The more people who own guns, the more other people will want to own them as well for self-defense; and the further ownership spreads, the more normal it seems. The ownership of guns has always been common in rural areas (the lower population density of the United States compared to Western Europe is an important reason why private ownership of guns is so much greater here), where there are hunting opportunities and police are spread thin. But now it is common in the rough areas of cities as well. Drug dealers cannot rely on police to enforce their deals and therefore have to arm themselves, and their law-abiding neighbors decide they had better be armed as well. (The news media create an exaggerated fear of violent crime, and this also contributes to the demand for guns by law-abiding people.) If population density continues to grow and the drug trade were legalized, gun ownership might begin to fall. 

Gun purchases soared in the economic crisis from which we are now (it seems) recovering. Partly this may have been due to increased cash hoarding (the sale of safes also soared) and to an increase in property crimes, but it may have been due mainly to a generalized fear that increased the demand for symbols of security. 

The second lesson is the unwisdom of the Supreme Court’s recent decisions that have created—on the basis of a tendentious interpretation of the drafting history of the Second Amendment and an intellectually untenable (as it seems to me) belief in “originalist” interpretations of the Constitution—a constitutional right to possess guns for personal self-defense. The result is to impose a significant degree of nationwide uniformity on a problem that is not uniform throughout the nation. The case for private gun ownership is much stronger in largely rural states, such as Arizona—states in which there is a deeply entrenched and historically understandable gun culture and a rationally greater lawful demand for private gun ownership than in the suburban areas of the densely populated midwestern, northeastern, and mid-Atlantic states—than it is in big cities with high crime rates—cities that have long had very strict gun laws many of which may now be ruled unconstitutional. 

Though gun ownership cannot be forbidden any longer, it can (even under the new constitutional regime) be regulated, as Becker emphasizes. Gun-registration laws aimed at denying gun ownership to lunatics and persons with a history of criminal activity, coupled with heavy punishment of dealers or customers who violate or evade the laws, should survive constitutional challenge. Federal “felon in possession” laws already provide for heavy punishment of persons forbidden to own a gun because they have been convicted of a felony, if they are caught with a gun in their possession. Loopholes in gun-registration laws, such as permitting the sale of guns at gun shows without requiring the screening of purchasers, can be closed. And punishment can be enhanced, even more than at present, for persons who use a gun in committing a crime. A reduction in the criminal use of guns would in turn reduce the demand by law-abiding persons, and as that demand fell so might the demand of guns by criminals, given stiff punishment costs. A virtuous cycle might be initiated that would lead eventually to a significant overall decline in gun ownership.

Why Dictatorial Regimes Are Brittle—Posner

Dictatorships, as we are seeing in the Middle East today, and as we saw in Iran in 1979 and in the communist nations in 1989—not to mention France in 1789—have a way of imploding unexpectedly, the unexpectedness lying in the fact that no external event seems to have precipitated the collapse. These events belong to chaos theory: if you rock a canoe, it will maintain an equilibrium until at some unpredictable point you rock it so hard that it capsizes. So I will not be speaking of regimes that collapse because of a catastrophic military defeat, nor of regimes ended by civil war, nor of sesessions, such as the American revolution, but just of sudden collapses, unforeseen because there was no visible triggering event that might have been foreseen. 

Over a long period of time, democratic and quasi-democratic nations change profoundly, but the change is gradual. Dictatorial regimes change in fits and starts, so that most of the time they seem more stable than nonauthoritarian regimes. They experience punctuated rather than incremental change. 

There are several reasons. The obvious one is lack of information. A government that uses intimidation, surveillance, and control of media to quell dissent deprives itself of good information about the population’s concerns. People keep their concerns to themselves out of fear. Grievances are driven underground, to fester. Not having a good handle on what people want, the government risks being blindsided by a sudden explosion of repressed anger. Repression also fosters conspiracy; fearful of expressing themselves publicly, people learn to form secret cabals; they become experts at dissimulation. 

Second, the leadership of an authoritarian regime has difficulty obtaining information even from its own officials, or more broadly of managing disagreement and absorbing and responding to criticism. Without fixed terms of office and rules of succession, the position of leaders is insecure: they maintain their position by charisma or fear, by projecting an image of infallibility and omniscience, and these sources of power are undermined by criticism, which is often implicit in “bad news” conveyed to leaders by their subordinates. Even without being critical, the subordinate who warns his leader about popular disaffection is implicitly claiming to have knowledge that the leader did not have. 

Third, and again attributable to the absence of set rules for peaceful transition of leaders, authoritarian regimes tend to be conservative in the sense of reluctant to change even in response to known problems. If you do not have a good handle on public opinion, it is very difficult to predict the consequences of change—change may convey weakness, create expectations that cannot be fulfilled, empower the advocates of change, and undermine belief in the infallibility and omniscience of the leadership. 

Fourth, and again related to the absence of regular rules of appointment and succession, the leadership of authoritarian regimes tends to be old and sclerotic. Retirement is dangerous. The leader will have made enemies and when he relinquishes power he is defenseless against them. By clinging to power he grows out of touch, and is ill equipped to respond decisively and effectively to a challenge. 

Although there are exceptions (particularly in East Asia), authoritarian regimes tend to be bad for economic growth, and this is still another source of potential weakness. No person can rule alone, or by fear alone; he has to reward his key officials, and so corruption tends to be common in such regimes. Also, the military tends to be larger and more expensive than actually required for national defense, because the regime depends on force and therefore most cultivate the loyalty of the military. It is not that a large, well-paid army is necessary to maintain internal order, but that if the army is not coddled it may overthrow the regime, or fail to come to its defense in crisis. 

When an authoritarian regime suddenly collapses, this is seen by the world as an occasion for rejoicing—democracy has triumphed. That is the response of most of the media to the current crises in Tunisia and Egypt. And in fact a sudden collapse often is followed by a democratic interlude, as happened during the French, Russian, and Iranian revolutions. But I emphasize “interlude”; there is nothing automatic about a democratic succession to a collapsed dictatorship. 

Indeed it is likely that if, unlike the formerly communist nations of Central and Eastern Europe, a country has never had a democratic government for more than a brief period, the flowering of democracy in the wake of the collapse of the authoritarian regime will also be brief. Admittedly there are numerous exceptions. Russia is one, though it is less democratic today than it was immediately after the collapse of the communist regime. Japan is a partial exception, because it did have a parliamentary government before World War II, though it was never really democratic. India is a real exception, and there are others in Africa and Latin America. As these examples show, nations are capable of transitioning from authoritarian to democratic societies. There are many democratic nations today, but, apart from the ancient Greek city states, there were virtually none before the nineteenth century, and few before the twentieth. Most of these, however, emerged from authoritarian government by a process of evolution, rather than suddenly; there were democratic roots in the American colonies and Great Britain, for example, long before democracy became the regime of either polity. 

In a country without democratic or liberal traditions, the party to win the first election and become the governing party will think it the most natural thing in the world to endeavor to retain power, by whatever means available, once power is achieved. And the party to win the first election might be whatever conspiratorial faction was best organized; it might have no commitment to democracy. So the first election might be the last. The old regime’s techniques and institutions of repression would be at hand to facilitate the takeover by the winning party. That is why the media’s celebration of the emergence of democracy in the Middle East today is premature, and why the Obama Administration is beginning to back away from its public celebration of what is happening on the streets of Cairo and other Egyptian cities.

Abrupt Change of Authoritarian Regimes-Becker

 

 

The Tunisian and Egyptian political eruptions were pretty much totally unexpected by the governments of the United States and of other countries, and by the vast majority of experts on Egypt and the Islamic world. To be sure, experts were aware that the government of say Egypt was not popular among many segments of the population, including The Muslim Brotherhood, most intellectuals, and many members of the growing middle class. However, the timing and speed of the uprising there (and in Tunisia) was rather a complete surprise since Mubarak and Ben Ali were in power for over 20 years, and seemingly in rather complete control.

I was first impressed by the unexpected and speedy nature of the overthrow of authoritarian regimes in 1979 when a combination of religious and leftwing groups forced the Shah of Iran from power. Until very close to the end he looked invulnerable: he seemed to be in full control of a strong and well-equipped army, and had an active and dreaded secret police, the SAVAK, that imprisoned anyone who vocally attacked the government. That the overthrow was unexpected is objectively measured by the stability of the international value of the Iranian currency, the rial, until just a few weeks before the Shah was ousted. Had the overthrow been anticipated, the value of the currency would have plunged as Iranians and others tried to get out of rials into dollars and other hard currencies. The rial did plunge in value shortly after the revolution appeared to be succeeding.

The rapid disintegration of the Soviet Union is another telling example. In 1989 my wife and I took a train from West Berlin through East Germany to go to Warsaw. The customs agents in East Germany were unpleasant, and the East German government headed by Erich Honecker seemed totally in charge. Much to my surprise, less than six months later, close to one million younger men and women were demonstrating in the streets, and the government was soon quickly gone, along with most of the Russian empire.

The unexpected nature and the speed of the overthrow of these and other authoritarian regimes is what is so glaring and challenging to theories of authoritarian rule. Analytically, what happens is that over time such a regime may be shifting in unnoticed ways from stable equilibrium positions, where the government is in rather complete control, to an unstable equilibrium where seemingly small events trigger massive changes, including the ouster of the government. The overthrow of the government may be quick and without much violence, as in the East German and Tunisian cases, or involve considerable violence, as during and especially after, the Iranian revolution.

Such unstable equilibria are sometimes called “tipping points”. This term was first used to describe rapid changes in housing neighborhoods from being mainly white and Christian to “tipping”, and then rapidly becoming mainly black or Jewish. A neighborhood may remain basically say all white until a few black families move in. If more black households move in over time, their fraction may become large enough that many white residents begin to panic, and put their houses up for sale. After that the neighborhood quickly “tips” into becoming a mainly black neighborhood.

The basic underlying reason that authoritarian regimes fall quickly, with or without violence, is that, as Posner emphasizes, they do not have any natural succession process. A strong man like Mubarak would be in power, but as he ages and gets weaker who is to succeed him? His son or confidants? Opposition groups may begin to see opportunities, or the unhappiness and frustration of young people and others may spontaneously erupt into mass demonstrations, as in Egypt, or in Iran after frustration over the outcome of the presidential elections two years ago. Sometimes these demonstrations succeed, as in Tunisia and apparently now in Egypt, and sometimes they fail, as in Iran after those elections, and in the 1989 Tiananmen Square demonstrations in China. 

Will similar demonstrations spread to the rest of the Arab world in North Africa and the Middle East that without exception have non-democratic regimes? Already the Jordanian government and a few others have started to make concessions to the opposition, including giving greater representation to various disaffected groups. I do not know how many of these governments will change radically and speedily. The theory offers little guidance on the timing of major political changes, but I do believe that large changes in this region toward freer elections and greater representation will occur before very long.

The Internet, Facebook and other online social networks, are changing the dynamics of the political landscape in all countries, including Islamic countries. In addition, the middle classes are growing in importance throughout Middle East and North Africa. As a result, these countries will experience the same aspirations for greater freedom of expression and greater representation in the government, as is found in other parts of the world. Eventually, these aspirations will force a conversion of the political institutions of these Islamic countries into something that may not be the same as Western democracies, but will offer more contested elections, greater political and social freedoms, and probably also greater economic freedom.