Two new recent research papers remind us that it may be difficult to cut U.S. healthcare spending without harming quality. The first, written by a research team led by University of Chicago economist Tomas Philipson, appears in the latest issue of Health Affairs and has deservedly garnered a fair bit of media attention. The authors examine cancer spending and survival times for patients in the United States and ten European countries during the period 1983-1999 (later data were not available.) Their data confirm what we already know about health spending; the average cost of treating a cancer patient was about $15,000 higher in the United States. But the data also show that the typical U.S. cancer patient lives nearly two years longer; most of the difference is attributable to prostate and breast cancer patients. The gain appears to be due to greater longevity rather than early diagnosis. Using generally accepted measures of the value of a life, they conclude that the benefits of additional health spending outweigh the costs by a factor of 4:1 or higher. The latter calculation does not consider QALYs (quality adjusted life years) and so may be overstated. The authors acknowledge that other nations may do a better job of cancer prevention, so that their overall approach to cancer may be superior to that in the U.S., but they can find no evidence of this one way or another.
Philipson’s study suggests that U.S. healthcare consumers may get a substantial bang for their higher bucks. Maybe the U.S. system is not so inefficient after all. What about efficiency within the U.S. system? Some providers are far more expensive than others. Is the higher cost worth it? A new study by a team led by MIT economist Joseph Doyle, and released as an NBER Working Paper, suggests that you may get what you pay for within the United States. Doyle and his colleagues ask whether higher cost hospitals in the United States achieve better outcomes than lower cost hospitals. It is not easy to answer this question, because higher cost hospitals may admit more severely ill patients. This results in a statistical problem known as selection bias that is difficult to eliminate with available severity measures.
Doyle et al. do something exceedingly clever to overcome this problem. They take advantage of the fact that in New York State (which is an excellent source of hospitalization data), ambulances are dispatched to patients based on ambulance dispatch boundaries. In addition, more than one ambulance company may serve a given location. When this happens, ambulance assignment is based on a rotational assignment. Because different ambulances tend to serve different hospitals, the upshot is that emergency patients who live near each other are effectively randomly assigned to different hospitals. This eliminates any concerns about selection bias. Once we understand their methodology, the results are easy to describe. Quoting from their abstract: “Higher cost hospitals have significantly lower one-year mortality rates compared to lower-cost hospitals.” The effect is dramatic – hospitals in the 75th percentile of costs have 4 percentage point lower mortality rates than hospitals in the 25th percentile. (The authors acknowledge that there seem to be diminishing returns with no further reduction in mortality above the 75th percentile in spending.) Their results seem to reflect greater treatment intensity rather than higher prices. I wonder if this has more to do with the quality of the providers at hospitals that do more expensive procedures, rather than the procedures themselves.
What does it all mean? Philipson’s study provides compelling evidence that there are consequences to wringing costs out of a nation’s healthcare system. The United States spends far more on cancer care than European nations, but also delivers the best outcomes in terms of cancer survival. (Within Europe, the nations that spend more also have better outcomes.) We can agree that there is fat within every healthcare system, and the U.S. may have more fat than most. But it appears that when it comes to putting a nation’s healthcare system on a diet, one cannot cut out the fat without cutting too close to the bone. And in the case of cancer treatment, European nations appear anorexic. Doyle’s study leads us to a similar conclusion about spending by U.S. hospitals.
A final note: You can now follow me on Twitter @daviddranove. I will be tweeting about new Code Red blogs and occasionally when I see something about the fascinating world of health economics.