THERE IS GOOD NEWS on the health-care spending front. According to the federal Centers for Medicare and Medicaid Services, the nation’s health-care bill grew only 3.9 percent in 2011, the third straight year at that pace — which is about half the average annual rate between 2003 and 2007. Consequently, health care is no longer consuming a rising share of the overall economy; it’s held steady at 17.9 percent of gross domestic product since 2009.

Alas, no one knows exactly what caused this welcome trend, which almost none of the experts predicted. Therefore no one can be sure it’s permanent. Declining spending may reflect the pinch of recession on consumer pocketbooks. That’s good, to the extent that it proves cost pressure can make consumers forgo health care they don’t really need; but it’s not so good to the extent that cost pressure forced consumers to forgo care they did need. Either way, spending would rise as the economy continues to recover.

It’s easier to identify factors that did not contribute to the downward bending of the cost curve. Health information technology is a case in point. At one time, its cost-cutting promise seemed immense. A 2005 Rand Corp. analysis estimated that the nation could save up to $81 billion a year by digitizing patient records and other data; best of all, it was said, more accurate, accessible information would also help improve patient care. All that stood in the way was the fact that doctors and hospitals had no incentive to invest in equipment that would save other people money. So President Obama’s 2009 economic stimulus bill allotted $27 billion in incentives to health-care providers who adopted electronic health records and showed they’re being used to improve patient care.

But as an article released Monday by Rand researchers argues, the subsequent investment has bought very little in the way of either improved care or lower costs. It’s possible that computerization may actually have increased spending. Between 2006 and 2010, Medicare payment to hospitals receiving federal dollars for electronic records grew faster than payments to hospitals that did not take advantage of the incentives, according to a Sept. 21New York Times report cited in the new Rand article.

What IT’s boosters did not anticipate was how hard it would be to develop a truly “inter-operable” system — that is, one that enables various providers to access a patient’s data no matter who created the record. A lot of money went toward automating billing offices, which made it easier for some providers to “upcode” — charge more for the same services.

Health care still runs overwhelmingly on the fee-for-service principle, which means that the more doctors and hospitals do, the more they get paid. Digitizing that system may just reinforce it. Or, as the Rand article, published by Health Affairs, puts it: “[R]evamping of health care delivery is unlikely to happen before payment models are realigned to favor value over volume.” America’s health-care cost problem is a question of institutions and incentives, which are not necessarily susceptible to technological fixes.