By Atul Gawande
The more expensive it is, the better it must be. That’s how people used to think about health-care costs. As has become apparent in recent years, however, this is flat wrong. The costs of care vary wildly depending on where you live—by three hundred per cent or more, as I reported in “The Cost Conundrum,” in 2009. And research has found no consistent relationship between cost and quality across the country. Some of the most expensive places are among the most mediocre.
But nearly all this research was based on the analysis of government insurance programs, especially Medicare, the program for the elderly. Private insurers do not have to make information on whom they pay, how much, or what they pay for publicly available. Only government insurance programs do. A fascinating study out this week, however, manages to crack open the black box of private insurance. It analyzes payment data compiled, for the first time, from three of the country’s largest commercial insurers—Aetna, Humana, and UnitedHealthcare—which cover fourteen per cent of the U.S. population.
In some ways, the study finds very similar results. The costs of care for the privately insured vary from town to town just as crazily as they do for the publicly insured. But the patterns are strikingly different. The most expensive places for Medicare are not the most expensive places for private insurers. In fact, there was essentially zero correlation between where a city ranks in Medicare spending and where it ranks in private-insurance spending—even when you only consider people undergoing the exact same procedure.
Are doctors and hospitals treating people with Medicare differently from people with private insurance? This doesn’t appear to be the case. Instead, the new study found two very different stories explaining the difference between public and private health-care costs.
In Medicare, the prime driver of differences in costs between similar communities was differences in the number of tests and treatments given. In “The Cost Conundrum,” for instance, I compared health care in two Texas border towns: McAllen and El Paso. They had the same levels of poverty and poor health, but El Paso had half the per-capita Medicare costs despite the same or better results on quality measures. McAllen’s doctors were ordering more of almost everything—diagnostic testing, hospital admissions, pacemakers, coronary stents, surgical procedures—in many instances by twice as much or more. Spending on home health services alone in McAllen was five times that in El Paso. There appeared to be a huge amount of unnecessary care. That was the story of Medicare. And it made clear: more is not better in health care. It can sometimes be disturbingly worse.
When I wrote about McAllen in 2009, data for people in McAllen with private insurance were much more limited, but they showed substantial differences in the amount of care given to them, as well. The newer, bigger set of data released this week, however, quantifies another large driver of local cost differences for the privately insured: the prices insurers pay for hospital treatments.
Medicare can use its authority to set prices for hospitals. Private insurers can’t. They have to negotiate with individual hospitals. In general, they end up paying higher and far more variable prices than Medicare does. (There’s a big exception: Congress has prevented Medicare from setting drug prices—and as a result it often pays higher prices than other insurers do.) That, the new research finds, can lead to a completely different cost picture for the privately insured.
As the Times points out, for instance, the research shows that Grand Junction, Colorado—a town I’d reported to have some of the lowest costs for Medicare in the country—also has some of the highest hospital costs for private insurers because of higher hospital prices. McAllen is high cost for Medicare but turns out to be an average-cost town for Aetna, Humana, and United, because they’re paying higher prices in other towns. El Paso is average for Medicare and among the lowest-cost places for private insurers.
Differences in the number of tests and treatments given from place to place are still huge for the privately insured. But the cost of health care is like the cost of groceries—the total depends on the price of every item and on how many items you get. Both Medicare and private insurers have adopted policies and reforms that are reducing unnecessary tests and treatments and improving preventive care. In McAllen, as I wrote in May, such changes have saved Medicare an estimated half-billion dollars for that one community alone. But cutting costs for privately insured patients also requires addressing prices. And that’s a different matter entirely.
When your grocery store is the only one in town, it can jack up prices without losing customers. The same goes for hospitals. The study found that hospital prices in monopoly markets are fifteen per cent higher than in those with four or more hospitals.
It’s the Cost Conundrum Squared. The bigger the hospital, the more it can adopt systems that deliver better-organized, higher-quality, less-wasteful care. But the bigger the hospital, the more power it has to raise prices.
We have a few ways out of the conundrum. We can regulate the prices hospitals charge insurers—this is what Maryland does. We can break up big hospitals. We can encourage hospitals to become the insurers. (That’s what Kaiser Permanente in California has done. It provides members with prepaid care at its hospitals and clinics.) Or we can expand Medicare to more and more people until we’re single payer.
But, whichever way we go (and this being America, we’ll no doubt try to do some hodgepodge of it all), we cannot let the complexities blind us to the core concern. The one thing the medical profession is not rewarded for is providing better, higher-value care. We are financially rewarded either for doing more stuff or for securing monopoly power. In a fee-for-service payment system—a system of paying doctors and hospitals by pill and procedure—we are actually penalized for making the effort to organize and deliver care with the best service, quality, and efficiency we can. That’s why both public and private insurers are rolling out reforms like “bundled payments” (paying a package price for certain conditions) and “accountable care” (sharing savings doctors and hospitals produce from more efficient care)—they want to replace our system of paying for stuff with one that pays for outcomes.
This is a difficulty for any area of human service. (We haven’t figured out how to effectively, on a large scale, reward schools, police departments, or fire stations for their results, either.) But these are the experiments that are going to matter—no matter what insurance a person has.