This is an issue I have been working on for 30 years. As a VA employee, I was one of the initial designers of the VA's VistA system, then as a VP and chief scientist at SAIC, one of the designers of the DoD's CHCS system. In the 1990's, I worked on a (rejected) proposal to install a variation of the VistA system at Kaiser.

The basic problem with this approach is that it attempts to integrate around the enterprise, not the patient. Each organization places themselves at the center of the information system, and places other organizations at the periphery. N organizations require N-squared integration efforts and agreements.

Imagine that Kaiser had an intercom system connecting all of their vital offices. Each critical office had fixed button position, so that you could instantly talk to the desired office. Now, imagine that VA had the same, and the two announced an effort to "integrate" their two systems. Even if they succeeded in figuring out how to do this, it would be a complex, brittle, non-scalable solution. And if they wanted to integrate with Scripps Health, for example, the complexity would start over again, only more complex as there would be a three-way negotiation involved.

Trying to "integrate" VA and Kaiser information in this way is like trying to integrate their intercoms. What about DoD information from Balboa Naval Hospital? Do we initiate a DOD-VA-Kaiser sharing agreement? (DoD and VA have been at this for 30 years now, and are still bickering about it. I've demonstrated three different VA-DoD interfaces; all of which have been shot down for political, not technical reasons). So now we are layering the political complexity of "integrating" two federal agencies and a large private organization. What about "integrating" Scripps Health? Do we start the whole process again?

And why are we doing this on a regional basis? Even if the two-way integration effort in San Diego works, what happens if the patient moves to Long Beach? What happens to folks in smaller organizations, or who are uninsured and unemployed? Do they simply not have a medical record?

To complicate this already complicated situation, we are asking all of Kaiser to trust all of VA with the information exchanged. As each new tendril in the tangled web of enterprises is added, all have to trust everyone else. And the security of the overall system is only as good as the weakest participant. One bad apple in the system can spoil the entire system; creating an ever-larger barrel through enterprise integration creates an ever-greater incentive for the bad apples to appear.

Then we have to look the evolution of the system over time. This effort is essentially a "stop the world, I want to integrate it" approach. If one enterprise moves in an innovative direction, the others will not be able to participate until they all move in lockstep in the new direction. If advances in genomics or proteomics take place, for example, how will this be communicated. And what if this innovation happens outside of the enterprises who have lock-stepped integrated their complexity?

This represents a brittle, complex, unsustainable model of integration that is a good way to channel billions of dollars of federal funding to vendors and the management teams dedicated to dealing with this.

As a taxpayer and someone who has played a lead role in this process for 30 years, I have to say that it appears to me that we are trying to get out of a hole by digging it deeper.

The way out of this complexity trap is to reframe health information around the individual patient. The patient is at the center, and providers are at the periphery. If someone wants to access that information, it happens according to the policies defined by the access "shell" around the information - and this is appropriately visible.

Health care enterprises (whose half-life is probably one third that of the patients they serve) would integrate with these personal health records. If they want to add their own enterprise approach on top of this, fine. But we shouldn't be paying federal monies to support the internal IT of health care providers any more that tax payers should be paying to support IT within airline companies.

Another way out of the situation is to frame health care IT as a "space" rather than an "integrated system." The web, for example, was designed as a "space for information to exist" rather than an integrated system for Google to search. This invites a different model of patient-centric activities and peer-to-peer linkages, rather than enterprise-centric, hierarchical and regional activities.

The situation we are in is a little like asking Ma Bell to invent the Internet. They had a perfectly good system (in their perspective) of circuit-switched phones and a great billing system to track long-distance calls by the minute. Why invent all of this dynamic routing, packet switching, and make users support their own routers when Ma Bell could do it centrally? Just give everyone a modem, and let them dial in and pay the monthly charges.

The health care industry is acting like the Ma Bell. "We are the smart center, and you are the dumb edges of the network" Inverting this model to a "smart edges" approach to health information technology is a very necessary approach. But I'm afraid we're not going to get there as long as we pour money and efforts into systems such as this.

Tom Munnecke
Encinitas, Ca
http://munnecke.com
(papers on patient centered records: http://munnecke.com/blog/?page_id=248 )