In a sobering keynote address at the 2010 World of Health IT Conference and Exhibition, eHealth researcher Enrico Coiera of University of New South Wales, Australia said industry enthusiasts who back, uncritically, national-scale HIT systems need to prepare themselves for some very bad news.

"We've yet to experience our first health IT plane crash, a health IT failure that claims many lives," Coiera said Wednesday. "But I think that will happen... I think it's unavoidable given what we're doing. We need to do our best to mitigate that."

Coiera titled his talk "The Dangerous Decade," because while he predicts unprecedented growth in healthcare information technology in the immediate future, that proliferation will come with some unwanted side effects -- especially when deployed on a national scale.

"I think over the next 10 years we will build more health IT than we have ever built before," he said. "These systems will be bigger and more complex. The costs and benefits are so large that they will significantly impact national GDP -- people are going to notice it.

"We have no choice but to do this," Coiera continued. "The danger is that health IT is still in its infancy. We are doing things we have never done before."

To date, Coiera said, national HIT projects typically have come in two flavors: a top-down, single system implementation exemplified by England's NHS National Programme for IT; and a decentralized, bottom-up federated system best characterized the United States approach. Neither had been an unqualified success, and in fact both approaches have faced significant criticism.

But experience with both of these models has suggested a "middle out" approach. "The middle out way is that we agree in totality what the common goals are, what the standards are... but allow people to innovate in their own interest," Coiera said.

Governments have policy expertise and should describe the high-level goals and build some common public goods which can't be built by private industry; namely workforce training, standards, infrastructure like broadband, and patient safety legislation.

"The thing that government should not do, at any cost, is design, buy or operate IT," he said. That should be left to private industry and local or regional healthcare providers.

Given the complexity of the healthcare IT needed to address patient safety and rising costs, Coiera said there were three risks to watch: the safety of IT systems, unrealistic expectations about those systems, and addressing the wrong problem.

"We're still focused on the technology," he said. "I worry that a focus on the electronic health record dominates the thinking of most people. We are missing the easy wins," such as decision support and e-prescribing.