Richard Cohen of the Washington Post returned to one of his areas of core incompetence last Tuesday with a lunging attack on Republican leaders John Boehner and Mitch McConnell. In a single column, he managed to recycle at least half a dozen myths, fallacies, and half-truths of what sometimes passes for conventional wisdom on the left bank of the Potomac.
Cohen was offended by Boehner's comment that the American health care system is "the best health care system in the world," going so far as to suggest that the next speaker of the House "might be French." (Them's fightin' words!) In his case against the American health care system's superiority, Cohen first appeals to the CIA. That's an unusual choice, given the agency's record of accuracy when it comes to fact-finding and forecasting, even on topics within its primary portfolio (see, for example, "demise of the Soviet Union" and "weapons of mass destruction"). Worse, it turns out that the CIA figures he cites are based on statistically misleading measures of life expectancy and infant mortality.
Life expectancy at birth is a particularly limited measure of health care performance across nations, because it generally fails to account for such important variables as lifestyle, culture, income level, and educational achievement. Life expectancy at older ages, such as at 65, gives a clearer picture--though it does not eliminate the confounding distortion of non-medical factors--and using that measure, the apparent life expectancy gap between the U.S. and other comparable nations narrows. In fact, if one goes further out on the age curve to age 80 and over, one finds that the U.S. probably leads the developed world in life expectancy.
These differences highlight the U.S.'s focus on subsidizing health care for the elderly, for whom medical interventions are more frequent, costly, intensive, and arguably more beneficial, and to whose future health non-medical factors matter less on the margin. (Their likelihood of voting is also higher…) A study published earlier this month in Demography finds that at age 55 and beyond, Americans are sicker by far than the English, yet older Americans don't die earlier than their British counterparts: Death rates were equivalent for 55-to-64-year-olds, and beyond age 65, Americans had a slightly greater probability of survival. Why is this so? Perhaps because the U.S. health care system diagnoses and treats illnesses (particular among the elderly) more aggressively than does the National Health Service--though, of course, all that extra screening and more intensive treatment costs more money.
The next old chestnut of international health care comparisons that Cohen serves up is infant mortality (deaths in the first year of life). Major problems with infant mortality statistics have been pointed out by others in the past and include differences in data definition and common health care practices. For instance, American medical practice more commonly resuscitates very small premature and nonviable-birth babies; these babies later die but are treated as "live births" in U.S. statistics. Countries such as France and Japan are likely to classify such babies as stillbirths, which aren't counted. Infant mortality rates are also affected by outside factors such as the mother's behavior and lifestyle (e.g., obesity, tobacco use, excessive alcohol use, recreational drug use, and marital status).
Somewhat better measures of perinatal mortality (death in the first week, plus fetal deaths that meet or exceed the minimum gestation time or weigh standards) and of birthweight-specific mortality reveal much smaller differences between U.S. rates and those of comparable nations. And, of course, I should note that Richard Cohen, who is pro-choice, mourns for infants who die "before they can get a cupcake with a single candle" but not for aborted fetuses, which have a mortality rate of 100 percent.
Next up on Cohen's checklist is "avoidable mortality," which purportedly estimates deaths from causes that should not occur in the presence of timely and effective health care.
The most notable proponents of this measure are British researchers Ellen Nolte and C. Martin McKee, whose 2008 study (supported by the Commonwealth Fund, which probably never met a person they didn't think deserved more comprehensive levels of health insurance) compared trends in health care-amenable mortality in different nations.
They concluded that the United States started with a relatively high amenable-mortality level in 1997–98 and then saw unusually small reductions over the next five years, relative to comparable nations. But, once again, there is much less here than meets the eye. For one thing, the study failed to adjust the proportionate share of deaths within given populations due to amenable mortality for changes in overall national mortality rates--in other words, the share of all deaths occurring in a given time period that were due to amendable mortality.
Even Nolte and McKee acknowledge that death is typically the result of a complex chain of processes including social and economic factors, lifestyle factors, and preventive and curative health care, and they concede that this renders the underlying concept of amenable mortality somewhat less than definitive as evidence of differences in health care system effectiveness.
Nolte and McKee were even more forthcoming on the limits of avoidable mortality measures in their 2003 British Medical Journal article "Measuring the Health of Nations" and in their 2004 literature review of avoidable mortality, "Does Health Care Save Lives?" As they observed:
Available data reflect only what is measurable and not necessarily what is important.
Declines in avoidable mortality may have been confounded by simultaneous changes in disease incidence.
Conditions amenable to medical intervention form only a small proportion of total mortality, and analyses of trends are thus likely to overemphasize the overall impact of health services.
Differences in avoidable mortality may be due, at least in part, to differences in diagnostic patterns, death certification, or coding of cause of death.
Partitioning deaths among different categories is an inexact science.
The choice of an upper age limit (including deaths only to age 75) is problematic and arbitrary.
There was at that time little evidence for an association between observed geographical variations in amenable mortality and other measures of health care provision.
Other critics of the Nolte/McKee approach point out that their study essentially only demonstrates how many people in each country died from an arbitrary list of particular diseases and conditions. It does not determine if individuals received care or if they could have been saved by care they did not get.
Returning to Cohen--he alludes to another cliché of health care mythology: the free-riding uninsured people who postpone treatment until they land in overcrowded and expensive emergency rooms. The statistical reality is that among the under-65 population, the uninsured are no more likely than the insured to visit the emergency department, nor are their visits more likely to be triaged as non-urgent. On the other hand, persons under 65 with Medicaid coverage are more likely to have multiple visits to the ED than those other two categories. Adults with Medicaid accounted for most of the increase in ED visits from 1997 to 2007.
In the 2006 National Hospital Ambulatory Medical Care Survey (NHAMCS), it was easy for the reader to see in the ED summary that those covered by Medicaid visit the ED more often than the uninsured (82 visits per 100 persons, as opposed to 48 per 100 persons). But the 2007 version has a different format, and it is now much harder to find figures for the visit rates to the ED per 100 persons based on payment source/insurance coverage. (Any connection between a new HHS administration selling expanded Medicaid coverage and this change from longstanding report format must be purely coincidental.) By the way, did you know that the number of visits to emergency departments actually decreased between 2006 and 2007, from 119.2 million to 116.8 million? Yes, I must have missed reading that headline in the mainstream health media, too.
Thoughtful observers might consider the possibility that an increase in ED visits primarily reflects broader delivery problems (e.g., physicians who don't do evening or weekend hours, or answer e-mail, provider resistance to low-cost clinic competition, etc.) rather than increases in the number of uninsured Americans. They might also wonder whether the new health care law's plan to increase coverage primarily through expansion of Medicaid will help or aggravate the problem of emergency-care overuse. But Richard Cohen only had enough space, or attention span, to use the uninsured-in-the-emergency-room image as a throwaway line.
Cohen's paragraph on the use of Bethesda Naval Hospital by some anti-Obamacare members of Congress is another miss. Criticism of the special health care perks that members of Congress can get for a $500-or-so annual fee may be justified to some extent. They essentially receive personalized, all-that-you-want primary care from official congressional physicians on call in the Capitol, and this ties them in to quick, no-fuss referrals to specialists and admissions at Bethesda Naval--and, despite some disavowals on the record, when members check in to Bethesda, they do receive VIP-like treatment. However, this is not dissimilar to the executive health care benefits enjoyed by top officials in private corporations or by government officials in countries with national health care systems. (It's good to be the king!) Moreover, the same deal could not be extended to all, under either a private market or a public program, without either breaking the bank or putting practically everyone first in line (and therefore right back in the middle). In any case, taking advantage of such benefits, arguably in part on the taxpayers' dime, is nothing new in Congress, and sometimes is no guarantee of quality care.
Essentially, Richard Cohen's column is an overwrought, highly politicized reaction to the periodically shallow rhetoric of some Republican officeholders who refer to U.S. health care as the best in the world--which, in some respects, U.S. health care is: for instance, in cancer detection and treatment and in a number of relatively sophisticated procedures for life-threatening illnesses. But that's not the point. The real issues are (1) how to improve it, particularly in terms of more consistent quality and greater affordability; and (2) how to refrain from worsening it, along with the economy, through a harmful prescription (Obamacare). Some Republicans have focused more on #2 than #1, which is equally important but more complex, but correcting their emphasis is hardly the most important mission we face.
As a matter of full disclosure, I have met with Rep. John Boehner on several occasions as a member of his outside economic- and health-policy advisory group. I've also discussed health-policy-reform proposals with his staff. He doesn't need my affirmation to show that he understands the need for substantial reform beyond the repeal of Obamacare. This article was written without any contact with his office.
As for the worldly Richard Cohen, he's on the record as ambivalent about torture in general, but apparently he's all for torturing health care statistics until they support his policy preferences. Then again, he's pretty much a blank canvas when it comes to health care. He wrote this last year about President Obama: "I feared that if I did get an exclusive interview I would be expected to ask him something about health care reform, about which I know next to nothing. What was worse, despite reading six newspapers a day, watching cable news shows, network news shows, the NewsHour and being online all the livelong day, I could not fathom what the president wants to do with health care."
Sometimes, a weak effort at satire delivers a damning admission of fact.
Thomas P. Miller is a resident scholar at AEI.