One of the disappointing things about healthcare policy research is that its volume is inversely proportional to its quality. Each year, sheaves of research papers are produced by academics and think-tankers, thick with tables and charts, purporting to argue that 62% of all bankruptcies are due to medical expenses, or that 45,000 people a year die because they don’t have health insurance. These studies are then broadcast uncritically by the press, and repeated as gospel by soundbite-seeking politicians. Unfortunately, the methodologies used in such research are often poor, and in the two examples above, intentionally misleading. (Megan McArdle is one of the few writers who has tackled this subject well.)
Such is the case with a new study published last week by the Commonwealth Fund that argues that, compared to six other developed countries, “the U.S. health care system ranks last or next-to-last” on measurements of its quality, access, efficiency, equity, and “healthy lives.” Overall, the study ranked the Netherlands first, followed by the United Kingdom, Australia, Germany, New Zealand, and Canada, with the U.S. ranking dead last.
The study is typical of the genre: drawing conclusions that are not warranted by the data; failing to account for alternative (and more plausible) explanations; and using flawed methodologies. The point of view of the authors is clear: in the first paragraph of the report, they write that “newly enacted health reform legislation in the U.S. will start to address these problems by extending coverage.” But they do their cause no favors with such a tendentious report.
First of all, the authors attempt to assess the quality of U.S. and international health care systems not with hard data, but with subjective surveys. The survey assessed the effectiveness of health care by asking patients and physicians questions like: “Did you receive reminders for preventive and/or followup care?” and “Do you believe a medical mistake was made in your treatment or care in the past 2 years?”
While it is interesting to look at such surveys, it is simply not serious to draw hard conclusions from them. U.S. consumer culture is famously more demanding than that of other developed countries; while Americans still defer to their doctors and pharmacists, they do so appreciably less than Europeans do. They are more likely to suspect error, and far more likely to sue if they believe an error has been made. An international poll that does not even attempt to normalize for these cultural differences is not informative.
There is an energetic debate among empiricists as to how best to measure quality in a healthcare system. Blunt instruments like overall life expectancy don’t take social factors, like violent crime and obesity, into account. European authorities manipulate infant mortality statistics in order to make their national health systems look better than those of rivals (much in the way that some U.S. localities pad crime statistics).
One promising approach is to examine survival rates after a serious diagnosis has been made: for example, how long does the average patient live after being diagnosed with a stroke? With a heart attack? With breast cancer? This approach has several qualities in its favor: it is quantitative; it can control for population disparities between countries by focusing on a particular disease; and it focuses on the actual purpose of delivering health care: treating disease.
As with the infant mortality problem, such national statistics can be manipulated, but so long as the level of disclosure is sufficiently detailed, one can attempt basic comparisons. The Lancet Oncology has published two sets of interesting studies on this subject as it relates to cancer: CONCORD, a worldwide analysis of cancer survival rates; and EUROCARE, a European one. CONCORD found that, in all cancers studied, 91.9% of Americans survived for five years after diagnosis, compared to 57.1% for Europeans.
Another important element of health care quality is speed of service: how long did it take to get an appointment to get an MRI scan? How long did it take, once surgery was prescribed, to undergo the procedure? Delays in diagnosis and treatment are important, because a disease can worsen when it hasn’t been discovered or isn’t being treated. The Commonwealth survey asks its participants some relevant questions in this department, but de-emphasizes—or doesn’t ask—the questions that really matter. In addition, the questions it does ask, it asks in ways that blur, rather than reveal, differences. Hence, the survey ranks the U.K. higher than the U.S. on timeliness of care, even though vast experience shows that the U.K. is among the world’s worst, and the U.S. among the world’s best, on this front. (The British waiting time problem got so bad that Tony Blair brought in Donald Berwick, President Obama’s nominee to head the Centers for Medicare and Medicaid Services, to tackle it.)
One important thing to remember when comparing the American health care system to that of other countries is that we actually have three different systems: one for those with private insurance (177 million people in 2010), one for those under Medicare (45 million), and one for those under Medicaid and SCHIP (40 million). A rigorous survey of U.S. health care quality would break out the distribution of those in each program, and the respective performances of each.
One thing we do know is that people with Medicaid fare worst: because Medicaid severely underpays doctors and hospitals for the care of its beneficiaries, many physicians refuse to take Medicaid patients. As a result, wait times and health outcomes for Medicaid patients are far worse than those for people with private insurance or Medicare. The Affordable Care Act expands the Medicaid population by 20 million, which is certain to put additional strains on this already-broken system.
Hence, though the authors of the Commonwealth study are sanguine about the prospects of Obamacare to improve the quality of American health care, the likely outcome is the opposite.