A few weeks ago in central Massachusetts, a previously healthy 15-year-old named Alex Hall awoke in the middle of the night because his chest hurt. Initially Alex thought the feeling would pass, but when it continued, his worried parents drove him to a local emergency room. There a blood test suggested that -- incredibly -- he was having a heart attack. An ambulance quickly took Alex to a tertiary care center staffed by pediatric specialists, and as the on-call pediatric cardiologist, I was paged to help out.

Alex's heart probably had been attacked by a virus, making him the latest victim of a small but worrisome epidemic in the area. Over the next two hours, the inflammation spread like wildfire through his heart, and the orderly spikes on Alex's cardiac monitor soon became disorderly seismic waves. "I'm scared," he whispered, looking at the monitor. Soon after, his lungs began filling with fluid and he started gasping for air.

Critically ill with a rare condition, Alex needed the kind of specialized but pricey care that is frequently blamed for busting health-care budgets.

The same week Alex came to the hospital, a team of researchers led by John E. Wennberg at Dartmouth Medical School reported massive state-by-state variations in Medicare spending, with New York and New Jersey spending almost 50 percent more per person than North Dakota and Iowa without, on average, having people live longer as a result. A lot of the discrepancy was chalked up to more intensive inpatient care in high-cost states. Wennberg argues that because health care is paid for in a piecemeal fashion (the more doctors do to people, the more they earn), states with lots of specialists and hospital beds are rewarded for overusing them. Health care, he concludes, is perversely "supply-sensitive."

To some, the Dartmouth data encourage the notion that if the supply of specialists and hospital beds were suddenly cut, doctors might reserve fancy care for patients who really needed it, and thus costs would fall. But as Alex's case suggests, these cost controls will require hard choices -- and, inevitably, haphazard rationing of health care.

As the boy's heart continued to fail, our team realized that the only hope for survival was a machine that would bypass his heart and lungs entirely and deliver oxygen to Alex's brain and body. Called extra-corporeal membrane oxygenation, or ECMO, this technology required further escalation: We'd have to send him to another specialized center, at Children's Hospital in Boston.

The resources mobilized to save Alex were staggering. A specialized, fully staffed intensive-care truck from Boston arrived to transport him to the ECMO center, where a team of cardiac experts worked to surgically connect Alex to the device. For eight days, Alex lay connected to the artificial heart and lungs under general anesthesia to buy time for his real organs to recover.

ECMO is very expensive, available only in certain centers and not standard therapy, making it just the kind of care that causes geographic disparity in medical costs. But as reported last year in the journal Circulation, ECMO has been used worldwide in about 700 cases in which even the most advanced resuscitation had failed because of a child's severe heart failure, infection or lung disease. The lead author (who, incidentally, helped care for Alex) wrote that ECMO "rescued one-third of patients in whom death was otherwise certain." In short, the treatment snatched hundreds of children like Alex back from the dead. But no one could tell ahead of time exactly which ones would live.

Critics of American health care sometimes point out that high-cost regions have higher mortality rates and conclude that too much medicine is harmful. But that's like noting that obese people are on diets and deciding that low-calorie foods cause fatness. Centers that offer ECMO, for example, will have the highest costs and highest mortality since they attract the desperately ill. And while life expectancy in the United States supposedly places 42nd among developed countries, it's hard to imagine how a boy like Alex could have survived in such places as Guam or Jordan, which rank ahead.

Of course, America lags behind many countries in providing a basic level of service for all citizens, since we lack universal health care. But for many, the system offers almost unlimited resources, which are often taken for granted. (In the March/April issue of Health Affairs, a former editor at the British Medical Journal whose daughter failed to get proper care in England writes how American health care "can replace a sense of resignation and futility with action and hope.")

Still, the estimated $2 trillion that Americans spend on health can't continue to grow. Indeed, health-care costs may need to come down. But these cuts will never be painless, since they can almost never be targeted only at useless expenditures. We've known this since the 1970s, when the Rand health insurance experiment found that patients cut back equally on both superfluous and necessary visits when asked for small co-payments.

It's unclear whether doctors given limited resources would be any better at rationing care. For now, they can't always tell when ECMO is necessary or futile. The same is true for many other expensive therapies, such as new cancer drugs, organ transplants and innovative psychiatric care. That's why cuts will always be somewhat arbitrary, no matter who's making the decisions: In the end, some people won't get the treatment that may save them.

Alex survived and continues to recover. Because of a minor paperwork error, however, his name remained on a heart transplant wait list, and a transplant coordinator called his home to check on him. You can take him off the list, his dad told the coordinator: Alex was one of the lucky ones.

Darshak Sanghavi is an assistant professor of pediatrics at the University of Massachusetts Medical School and the author of "A Map of the Child: A Pediatrician's Tour of the Body."