Is the genie all powerful?

George Anstadt, in a comment on Blake's post on the etiology of crisis, draws attention to the potential benefits of new forms of information technology in health care, which he calls the "digital Genie." This seemingly very wonkish field is one that we should all try to learn something about, as health care consumers and as citizens.

Health Affairs recently published a theme issue on this subject. (Unfortunately, non-subscribers have access only to abstracts of most articles.) The more conventional term is Health Information Technology, which incorporates Electronic Medical Records (EMR) and Computerized Physician Order Entry (CPOE). These are linked, in that the physician's orders automatically enter the medical record, and information in the medical record can theoretically alert the physician if it affects the appropriateness of an order (e.g., a drug allergy or other counterindication).

People hold out great promise for HIT. If all, or at least most provider institutions used the same data format, medical records could be easily exchanged between doctors and hospitals. Consumers could, in principle, truly own their own records because all the records from all providers could be aggregated on a single server. (Did you know that legally, you already own your own records? You can ask to see them, and get copies. But hardly anyone does.) The record system could be programmed with "process of care standards," so that if someone was due a vaccination, or a screening test, the physician would be automatically alerted. They would also provide comprehensive databases of patient medical and treatment histories so that statistical analyses could be readily done establishing the compliance with quality standards of individual practitioners, hospitals, and health networks. CPOEs could be programmed to reduce prescribing errors and adverse events. The systems could automatically send reminders to people with chronic conditions about self-care. And so on.

This all sounds good although as Dr. Anstadt points out, it won't do anything directly about financial and other barriers to access. But as excited as many people are by the promise of such systems, there is a long road to travel before reality can be even close to the vision, and many potential pitfalls. Famous curmudgeonly critics David Himmelstein and Steffi Woolhandler, in the Health Affairs issue, run down the down side.

First, there's been talk of this digital revolution in HIT for decades, but progress has never come close to the hype. Many hospitals have installed computerized systems that they have ultimately abandoned; and while HIT has indeed gradually been installed in more hospitals, the much-touted safety and quality improvement capabilities are often lacking, and don't always work.

A major problem is that electronic systems don't do anything by themselves. They have to be integrated with human work processes and individual human behaviors. This is extremely difficult to achieve. It is also far from straightforward to determine what really constitutes quality of care. While numerous so-called "process indicators" have been developed -- e.g., what tests should be performed under what circumstances, what drugs should be prescribed, etc. -- there is little evidence linking any of these to better outcomes, and many physicians resist them on the grounds that all cases are individual and their judgment may override the one-size-fits-all recommendation. There is also no particular reason to think that patients will generally comply with the reminders and recommendations the systems generate for them.

Economics are also a problem. These may be good investments for large hospitals, but probably won't repay the investment for small and mid-size physician practices and community clinics. (They might repay the investment from a social perspective, but that won't motivate, or even enable, the individual practices to invest.) There would have to be a substantial government subsidy. They also require technical expertise to install, operate and maintain which doesn't currently exist. And inter-operability -- the single standard so that records will really be portable and exchangeable -- is a huge task that will require a major government investment and mandate, which will be accompanied by massive controversy. There is little discussion in Health Affairs of the privacy issues but they are obviously huge.

So HIT is a tool. Like any tool, it has to be wielded by a skilled user, or in this case, millions of skilled users, all working in concert, within systems of work that function effectively with the technology. A hammer and saw won't build a house, that takes a carpenter. We may be able to use HIT to make health care better, but the basic issues diagnosed by Blake will only be solved by people.