The Reality-Based Community

 This may cause a little discomfort

I don't suppose a single vote on March 4 or November 4 will depend on the issue of electronic medical records. But I'll write about them anyway:

• It's an important issue for health care;
• There's a clear difference between HRC and Obama on the matter;
  and Yanks have the bonus of
• A costly British fiasco to jeer at.

The idea of a comprehensive national system of electronic medical records (EMR) is that clinical notes, and the associated data files from scans and tests, would be kept by all medical providers in the country in electronic form, in standard formats, and exchanged securely by standard protocols. This article by Robert Charette has more.

If you can get EMR right, as the Veterans Administration, the Mayo Clinic, the Regenstrief Institute in Indianapolis, and Finland seem to have done, there are great benefits in continuity and quality of care. Suppose you need treatment from unfamiliar doctors or nurses, for example in ER, if you have a medical problem while travelling, or when shifts change in a hospital. In these unsettling environments an accessible digital health record can improve and speed up diagnosis, avoid duplicate tests, and reduce medical errors. But what the USA has today is a vast patchwork of incompatible systems. There is a nice little federal initiative called the NHIN puttering away to develop standards for a national system, but obviously you don't get implementation for a trivial $61m a year.

The three main Democratic candidates based their health care proposals on Jacob Hacker's plan but Hacker was silent on EMR. HRC and Obama do mention it in their programmes, so briefly that I'll just quote them:

HRC:

Ensure That All Providers and Plans Use Privacy-Protected Information Technology: The proposal will give doctors financial incentives to adopt health information technology and facilitate adoption of a system where high quality care and better patient outcomes can be rewarded.

Obama:

Obama will invest $10 billion a year over the next five years to move the U.S. health care system to broad adoption of standards-based electronic health information systems, including electronic health records, and will phase in requirements for full implementation of health IT. Obama will ensure that patients' privacy is protected.

There is clear blue water between them. Both want to protect privacy. HRC mentions neither a specific sum of money, nor standards. Obama will spend a lot of money and secure common standards.

Should the federal government make this a spending priority?

Yes. Obama is right.

EMR is difficult. The well-managed Kaiser Permanente HMO has run into severe problems - and so has the huge NHS project for England and Wales (government, critics). There are risks in all large IT projects. But medical EMR has two special features that make it the one from hell.

Sunk costs
Hospitals and GPs have made very large existing investments in incompatible systems, including training; and the data to be covered are very wide-ranging as well as technically specific. This means that any shift to common standards will be costly, time-consuming, and resented. I remember the mess when my old employer merely shifted from WordPerfect to Word : changing EMR is more like changing an air traffic control system on the fly - with similar risks to life and limb.

The economies that nationwide EMR offers are those of networking. The efficiency gains are largely external to a given provider, the costs internal. When switching costs are fairly low, as with Betamax and VHS, the market will eventually impose a standard that enables the benefits. But in medicine, the costs are very large, and incurred by complex and change-resistant bureaucracies. For this reason alone, gentle suasion and small subsidies won't do it. There's an impeccable case in principle for government subsidy and/or regulation to secure standardisation; and in either case the pressure must be large.

Privacy and confidentiality
Medical privacy is very difficult and EMR makes the problem more acute.
Underestimating the problem was one main reason why the British NHS project spiralled from £6.2bn to £12.4bn (2006 estimate, so it's probably more now). It is pardonable in campaign documents just to say that "privacy will be guaranteed", but the whole point of electronic medical records is to share them. So do you allow privacy freaks to opt out entirely? Do you block off certain categories of data, like psychiatric records? The British solution was to create a short spine record accessible by any GP or hospital nationwide - i.e. by about 250,000 individuals - with details kept locally and released when needed (you can SFIK opt for release by permission). But what about children? Unconscious or mentally handicapped patients? In a nationwide system, can you guarantee security as least as good as with paper?

The USA is more diverse culturally than England, and you would expect an even wider range of views on privacy. But if you accommodate everybody's preferences, there is no national system. To get one, even after maximum consultation, in the end you need political leadership for the least bad solution.

After endless rows the English NHS finally allowed total opt-outs on the national spine record, and a release-with-permission option. (I don't see any plans for local opt-outs; if you tell your GP, you tell the local hospital). But when the new system went online in the pilot area, Bolton, it turned out that only a few availed themselves of the right. The quite wide opposition to the spine record has been driven by GPs - who will rarely benefit, see above, but will bow to the inevitable - rather than by patients. It's better to placate the few hardliners among the latter, at the cost of losing universal coverage, than to force them into a system they detest.

I've not heard of similar opposition in Wales and in Scotland, which now run their own parts of the health service. The objection appears to be to sharing information with strangers beyond a certain degree of familiarity. So the USA shouldn't even consider a clinical spine record, just an administrative one. The aim should not be to replace the current patchwork by a single, centralised scheme as in England, but a rationalised quilt of fewer but larger, up-to-date, and intercommunicating ones: a Baby Bell approach.

Postscript
Rather as an afterthought, and at comparatively trivial cost, the NHS set up a system that allows everybody to create their own personal health page. In future any patient (and of course anybody who can hack the password) will be able to see their summary care record as well via this page.. There's no movement of opposition to HealthSpace, though it lowers the effective security level drastically. But it gives patients a sense of control.