The predictable lead sentence in a recent New York Times story proclaimed not only the obvious — that the economic crisis in the U.S. reduced use of routine medical care – but observed that such cutbacks are much deeper here than in countries with universal health care systems. And the supposed culprit was the usual villain – higher out-of-pocket (OOP) health care costs in the U.S., compared to those in Britain, Canada, France, and Germany. All of this was according to a new study, “The Economic Crisis and Medical Care Usage,” published by the National Bureau of Economic Research (NBER).

As if.”

The competition is intense among many half-baked theories purporting to explain changes in relative levels of health care use and health care spending. In this case, the NBER study assumed facts that were not in evidence – that Americans face significantly higher OOP costs for health care than people in other comparable nations. It confused absolute dollar totals (in a larger health economy) with the more decisive “share” of health spending that is paid OOP. Study authors Annamaria Lusardi, Daniel J. Schneider, and Peter Tufano relied on Organization for Economic Co-Operation and Development (OECD) health data for 2007 to calculate that OOP health payments accounted for 2.0 percent of gross domestic product (GDP) in the United States, but only 1.5 percent in Canada, 1.4 percent in Germany, 1.0 percent in Great Britain, and 0.8 percent in France. However, they made no adjustment for the much higher share of GDP for total health spending in the United States, compared to those other countries. The technical term for this is called detaching the numerator from the relevant denominator, to make the former look larger compared to another less-related one.

Other, more reputable measures by the Centers for Medicare and Medicaid Services and the OECD calculate this differently, by comparing OOP health spending relative to total national health expenditures.

The most recent national health expenditure statistics compiled by the Centers for Medicare and Medicaid Services show that our OOP share of health spending continues to decline (11.5 percent estimated for 2009, 9.7 percent projected for 2014), which reflects a long-term trend. (You may click on the charts in this post to enlarge them for easier viewing.)

U.S. Out-of-Pocket Spending as a Percentage of Total Health Expenditures, 1960-2008

Moreover, the U.S. OOP share of health spending, as of the last comparative figures available from the OECD in 2008 (12.1 percent), was below that of Germany, Canada, and the weighted average of all reporting members, respectively.

Out-of-Pocket Spending as a Percentage of Total Health Expenditures, Select Countries, 2008

Of course, one can quibble about whether other measures might be more appropriate (such as OOP spending relative to personal care expenditures, or PCE, which eliminates some categories of health spending) and how precise and accurate the data sources for various nations’ types of health spending really are. More granular cross-sectional analysis of health spending for the population between ages 18-65 and for “routine” versus higher-cost acute or chronic care would be desirable, but current national health spending accounts do not separate OOP spending into such distinct categories. However, the broader trends over time, regardless of the measures used, remain similar in direction and magnitude. The United States is far from a high cost-sharing outlier among industrialized nations and, in fact, the relative share of its health spending that is paid OOP has been declining over time.

OOP Spending as a Percentage of Total Health Expenditures, Select OECD Countries, 1998 and 2008

Using another cross-national online survey administered in June and July of 2009, Lusardi, Schneider, and Tufano found that, both absolutely and comparatively, Americans reduced their routine medical care more than respondents in the other four countries “since the economic crisis.” This finding might appear to be consistent with other evidence that patient visits, drug prescriptions, and other medical procedures are down over the last year (“Americans Cut Back on Visits to Doctor”) – except for other reports by the same insurers of “surging health-care usage” in the U.S. just a year ago – which was the time of the NBER study’s survey!

Lusardi, Schneider, and Tufano also observe that wealth losses for U.S. households since the start of the economic crisis were much more pervasive than losses in households of the other four countries, and unemployment increased more in the U.S. than in all but Germany. So it should not require too great a leap of faith to assume that spending on health care might not remain immune indefinitely to broader economic forces. (Health benefits packages and the insurance underwriting cycle do tend to lag behind changes in economic trends.) President Calvin Coolidge once said, “When large numbers of men are unable to find work unemployment results…” To update that unremarkable observation, so too does lack of health insurance coverage follow from downsizing or disappearing employers. And, amazingly enough, when overall wage and income growth slows during a severe economic downturn, some people might just trim their spending on more discretionary, “routine” health care.

The Link Between Economic Downturns And Health: Not What You Might Expect  

Another complicating factor behind any recent slowdown in the growth rate of health spending is the counter-intuitive effect of economic downturns on population health. Lusardi, Schneider, and Tufano briefly cite the work of scholars like Christopher Ruhm that finds that health conditions may improve rather than decline in economic downturns. However, they emphasize only the less consequential factor that reduced work hours or unemployment provides additional time for making medical appointments and seeking routine health care. Ruhm concluded that the overall effect of a reduction in wages and desired work hours on seeking more health services is ambiguous (less active workers also have less income to pay for health care). But he also found that health worsens as the economy improves. For example, individuals consume more alcohol and drive greater distances, increasing the risk of involvement in motor vehicle crashes. Moreover, the effects of macroeconomic conditions on health may occur over a period of years, and the effects of transitory versus permanent changes in economic conditions may be quite different. Temporary income growth worsens health, whereas permanently higher levels of wealth improve it.

In a later study focused on the risk of death from acute myocardial infarction (AMI), Ruhm found that heart attack death rates rise in the year the economy expands and grow further if the lower rate of joblessness is maintained. The reasons for this relationship include longer working hours that reduce sleep, elevate stress, decrease alertness, increase physiological or psychological symptoms, and increase the risk of injury. Long working hours also make it more difficult to undertake time-intensive health-producing activities (e.g., exercise and healthy diets) and lead to higher rates of obesity, heavy drinking, and smoking. (As a classic Ronald Reagan quote goes: “It’s true hard work never killed anybody, but I figure, why take the chance”).

Long work hours may also make it more difficult for workers to care for their aged dependents. Spillover effects of increased economic activity also include air pollution, traffic congestion, and increased isolation or loss of community support.

A similar study in 2004 of German states by Eric Neumayer also found that recessions tend to lower overall mortality rates. Although his economic model of utility maximization assumed that less time will be spent on routine medical check-ups in economic upturns, it also assumed that the higher opportunity costs of leisure time during those periods would mean less time for other health-preserving activities, less time for cooking lower-calorie and better-quality meals at home, and more consumption of calorie-rich prepared food. Economic upturns also were assumed to increase job-related stress, work-related accidents, and consumption of alcohol and tobacco. Neumayer’s empirical analysis weighed the effects of this economic model against those of a model focused on the social and psychological aspects of hardships caused by economic downturns, which might point in the opposite direction. It concluded that the positive impacts of recessions on mortality more than compensate for negative health effects for certain sub-groups of the population.

None of this means that the best prescription for better health is to “take a long recession and see me in the morning several years from now.” It appears that although temporary income growth worsens health, permanently higher levels of wealth improve it. Could the apparent slowdown in recent health spending partly be due to short-term improvements in health, and not just the inability to afford purportedly higher OOP costs during an economic downturn? Demonstrating changes in mortality due to the economy isn’t quite the same as showing that they also directly influence morbidity, let alone levels of health spending, right away. But a poor economy and improved health point in the direction of less, not more, health spending.

Conventional Wisdom: Wrong But Self-Reinsforcing

However, the self-reinforcing biases of many health news headlines, opinions of health policy “experts,” and interests of health services vendors remain relatively immune to empirical evidence on this front. Despite occasional throwaway comments that aggregate health spending is too high and/or unaffordable, the default presumption is that the actual consumption of care should remain unburdened by the economics of paying more of its full price out of pocket – at least at the margin. Early dollar deductibles are resisted as discouraging essential preventive care (aka dropping by the doctor’s office whenever the first unclear symptom appears, or to see if one might be discovered along with a billing code for it…). Partial cost sharing for larger medical expenses is seen as too punitive and overtaxing the limited abilities of patients to assess more complex tradeoffs. And leaving all the other mid-range types of health care cost decisions subject to cost sharing apparently either would single out the chronically ill too harshly and leave them prone to even greater health problems in the future, or it would jeopardize the underlying financial health of a health delivery system based on opaque cross-subsidies that detach prices from values. Before you know it, not a single dollar of health spending can be left at risk to the dangers of cost sharing. Hence, the headline-based ‘horror” when modest evidence appears that Americans might actually be consuming somewhat less health care, and seeing slower growth in health premiums, yet still managing to live another day!

The reality is that reasons for slower growth in some types of U.S. health care spending within the last few years are rather complex, though partly related to declines in our economy. Despite some apparent growth in the nominal dollar amount of potential cost sharing in certain segments of the private health insurance market, this first-party exposure to some of the costs of care has not yet translated into increases in the share of U.S. health spending that is paid out of pocket. Moreover, Americans remain at the lower end of the cost-sharing continuum, compared to citizens in other comparable industrialized nations. And, if left unchanged and then fully implemented, the new Patient Protection and Affordable Care Act will push us to spend even higher relative shares of other people’s money. Whether that actually improves overall health and whether we can afford it (compared to other alternatives) is a topic for another day.