Are we on the flat-of-the-curve?

My post yesterday about “flat-of-the-curve” medicine and waste concluded with some references to other papers, some of which I’ve had a chance to look at. They are very thought provoking. Below I list the ones I read with their abstracts. Following that are my general comments.


Chandra, A., and D. Staiger. 2007. Productivity Spillovers in Health Care: Evidence from the Treatment of Heart Attacks. Journal of Political Economy 115(1):103–40.

A large literature in medicine documents variation across areas in the use of surgical treatments that is unrelated to outcomes. Observers of this phenomenon have invoked “flat of the curve medicine” to explain it and have advocated for reductions in spending in high-use areas. In contrast, we develop a simple Roy model of patient treatment choice with productivity spillovers that can generate the empirical facts. Our model predicts that high-use areas will have higher returns to surgery, better outcomes among patients most appropriate for surgery, and worse outcomes among patients least appropriate for surgery, while displaying no relationship between treatment intensity and overall outcomes. Using data on treatments for heart attacks, we find strong empirical support for these and other predictions of our model and reject alternative explanations such as “flat of the curve medicine” or supplier-induced demand for geographic variation in medical care.

Doyle, J. 2005. Health Insurance, Treatment and Outcomes: Using Auto Accidents as Health Shocks. Review of Economics and Statistics 87(2):256–70.

Previous studies find that the uninsured receive less health care than the insured, yet differences in health outcomes have rarely been studied. In addition, selection bias may partly explain the difference in care received. This paper focuses on an unexpected health shock—severe automobile accidents where victims have little choice but to visit a hospital. Another innovation is the use of a comparison group that is similar to the uninsured: those who have private health insurance but do not have automobile insurance. The medically uninsured are found to receive 20% less care and have a substantially higher mortality rate.

Doyle, J. 2007. Returns to Local-Area Health Care Spending: Using Health Shocks to Patients Far from Home. NBER working paper, number 13301. National Bureau of Economic Research: Cambridge, MA.

Health care spending varies widely across markets, yet there is little evidence that higher spending translates into better health outcomes, possibly due to endogeneity bias. The main innovation in this paper compares outcomes of patients who are exposed to different health care systems that were not designed for them: patients who are far from home when a health emergency strikes. The universe of emergencies in Florida from 1996-2003 is considered, and visitors who become ill in high-spending areas have significantly lower mortality rates compared to similar visitors in lower-spending areas. The results are robust across different types of patients and within groups of destinations that appear to be close demand substitutes.

My Comments

All of these papers, and the one reviewed earlier by Kaestner and Silber use strong methods to handle the endogeneity of treatment intensity. One would expect sicker individuals would be both more likely to die and more likely to receive higher intensity care. Unobservable factors could be correlated with both. Hence, in these studies, instrumental variables and appropriate sample inclusion rules are applied. One study, Kaestner and Silber, restricts attention to Medicare inpatients with in-hospital complications. Another, Doyle (2005), above, restricts attention to individuals in auto accidents. Doyle (2007), also above, looks at patients away from home. These sample selection criteria are designed to exploit a random event and select for individuals who experience an intensity of treatment unrelated to unobservable characteristics.

Another common thread is that all studies focus on serious, acute, or emergency health events. These are exactly the type of health care events for which I would be least likely to expect “flat-of-the-curve” care. If there is waste in the health system (in the sense of high-cost, low-return), I’d expect most of it to be in non-emergency situations. In fact, most of the recent growth in health care expenditure has been in the outpatient setting. Consequently, I am not surprised at the findings of these papers, and I don’t think they thoroughly dispense with the idea that there is “flat-of-the-curve” provision of care. It’s just not found where they looked.

I also don’t think it is easy to prove or disprove the “flat-of-the-curve” hypothesis in non-acute, non-emergency situations. One can’t randomize patients to various levels of intensity of outpatient care, say (or not easily anyway). If one could identify measures of intensity, the approach of Chandra and and Staiger might be applicable. The impact on health of such care would be far downstream, however. Thirty day or even multi-year mortality rates likely wouldn’t be sensitive enough. One would have to consider other health outcome measures.

All in all, I don’t think one can draw a broad conclusion about the degree of “waste” or the likelihood of “flat-of-the-curve” medical practice. As good as these studies are, they have limits. This area is worthy of additional thought and research. I’m sure folks are working on it.

One final thing worth noting is that the paper by Doyle (2005) provides evidence of health returns to insurance, and for a population relatively more likely to be uninsured, as compared to the elderly, say. (Doyle excludes observations over age 65.) This is precisely the type of evidence that motivates coverage-expanding health reform. Certainly other evidence exists (I’ve reviewed some of it), but there is not a lot focused on the population of relevance, particularly working-age adults, as most drivers would be. This is one study not easily dismissed as irrelevant or methodologically weak.

Hidden information below


Email Address*