Why Halamka's health IT predictions might overestimate the future

John Halamka, CIO of Beth Israel Deaconess Medical Center in Boston and a professor of medicine at Harvard Medical School, is one of the most respected opinion leaders in health IT. He's also one of the smartest people I know. Yet his new piece in the MIT Technology Review, predicting where health IT will take us in the next five years, is too optimistic by half.

I don't dispute Halamka's contention that the pace of electronic health record adoption will accelerate dramatically, partly because of the federal government's incentive program. Nor do I disagree with his argument that health IT will be essential to transforming the provider payment system in ways that can control cost growth.

But I'm not convinced that cloud-based computing is going to result in an explosion of EHR implementation. It's true that much faster Internet speeds have made remote hosting feasible for physicians who will not tolerate a page load-up time of more than two seconds. But much still depends on the kind of Internet connection they have, how much local traffic is sharing the link they're using (in the case of cable), and what kinds of files they're downloading.

Moreover, usability can be affected by whether or not a particular EHR is a web-native application. Some physicians still don't like the idea of their patient data being stored in cyberspace. And, despite the advantages of someone else providing and maintaining a server, doctors still have to deal with other hardware problems.

Tablet computers are great, and new consumer-oriented ones like the iPad and Samsung Galaxy Tab have more flexibility and apps available to them than old Tablet PCs ever did. They're also lighter, cheaper and have a longer battery life than their predecessors. But it remains to be seen whether the new EHRs that are native to the iPad will provide the same experience as the equivalent desktop or tablet PC application, and whether doctors will find it just as easy to document on them.

Now we come to interoperability. The "new standards for secure email of data between providers" already are being integrated into some EHRs. But the Direct Project protocol for these secure email exchanges only allows physicians to "push" messages and attachments to other trusted providers who have Direct addresses--assuming that both parties have a health Internet service provider (HISP) available. It doesn't let doctors locate and retrieve data that they need at the point of care.

Consequently, the Direct protocol will not lead to the creation of a national health information network that is a "federation of subnetworks," as Halamka predicts. In fact, John Kachelski, CEO of the Wisconsin Statewide Health Information Network (WISHIN), recently told me that his organization plans to use Direct to jump-start his health information exchange. This will help overcome the absence of a statewide infrastructure and the lack of EHRs in many doctors' offices. But it's a big leap from there to a "robust bidirectional exchange," he pointed out.

Halamka is exactly right when he says that "only a few hospitals and cities in the U.S. are able to securely exchange health records, and even fewer have economic reasons to do so." In fact, the rise of enterprise-wide private HIEs that connect a health system's hospitals and their staffs--along with the distrust of competing systems toward one another--is the biggest obstacle to the development of regional HIEs. But without them, universal connectivity will remain a dream.

Halamka also is right when he says that more shared decision-making between providers and patients will be required to improve care. But the greater availability of healthcare information on the Internet does not necessarily produce more educated patients. In many cases, it results in confusion and misinformation.

As for the use of remote patient monitoring, it's true that a different reimbursement system would prompt some providers to use a variety of tools to keep patients from getting sicker and ending up in the hospital. But few health plans--which have been taking financial responsibility for care all along--are sufficiently convinced of remote monitoring's effectiveness to pay for it. Even if they did, healthcare organizations would have to figure out how to use low-level clinicians to track the data and only notify physicians when necessary.

I also don't think home monitoring will ever return us to the days when doctors made house calls for even trivial problems. More likely, providers will deploy the technology to the sickest patients.

My crystal ball is too foggy to tell whether Halamka is right when he says genomic information will be used for everyday medicine in five years. Suffice it to say that we're a long way from being able to predict whether a given individual will develop a particular disease, even if they have a family history. Being able to use genomic information to tailor treatments to individual patients is still science fiction.

You might think my critique of Halamka's vision is too pessimistic by half. But the effectiveness of information technology has always depended on how it's used--and I don't see that changing in the foreseeable future. - Ken