January 15, 2003

OpenEMed helps detect epidemics -- and bioterrorism

- By Robin 'Roblimo' Miller -

"OpenEmed is a distributed healthcare information system built around the OMG distributed object specifications and the HL7 (and other) data standards and is written in Java for platform portability," says the first sentence on the project's home page. Its objective is to aggregate data from clinics, emergency rooms, pharmacies, and individual physicians so that epidemics or bioterrorism attacks can be rapidly spotted and, hopefully, stopped before they affect large numbers of people.

"The Defense Threat Reduction Agency, a DoD agency, funded this work as part of their biodefense network," says project leader David Forslund.

To demonstrate the kind of situation where OpenEmed can be of value, Forslund points out a well-documented 1993 Cryptosporidium outbreak in Milwaukee, Wisconsin that reportedly made 403,000 people ill and killed 100. Forslund says, "The occurrence was detected by a pharmacist who noticed people buying a lot of Imodium at his store. Before that, there were a lot of calls about diarrhea [to local doctors and hospitals]. If people had connected these, the disease could have been detected earlier."

If the disease had been detected earlier, a simple "Don't drink tap water until further notice" announcement could have been broadcast by local media, and hundreds of thousands of people might not have gotten sick, and a few lives might have been saved.

"[OpenEmed's] main goal is to detect bioterrorism," Forslund says, "but we also want to make it able to react to natural disease outbreaks."

A lack of standards is the big problem

Doctor A has a record system from SuperDoc Programming, Doctor B has one from UltraMed Software, and three different hospitals in town each have their own records systems, all incompatible. This is why the Object Management Group has a sub-group devoted specifically to healthcare matters.

CORBA is important here, and as this diagram shows, a number of leading software vendors and systems suppliers in the health care industry are implementing CORBA in their products.

"The important thing is that these implementations are completely standards-based," says Forslund.

He also says the Internet has made the task of sharing medical information both easier and harder, that "you can now publish medical information on the Web, and this knocks prices down." It also means that where there once only a few hundred medical information systems -- mostly mainframes -- piping information back and forth, "you can now have a few hundred thousand pipes talking to each other."

And this huge number of potential information-sharing systems, Forslund says, is why you need "more than data standards, you need functional standards, a la LDAP, with a standard API you can call to, that you can plug and play and make a system out of."

Practice, not theory

OpenEmed grew, in part, out of the TeleMed project sponsored by the Los Alamos National Laboratory. TeleMed is still active and doing its job today.

There was a "live" test of the OpenEMed system in 2002 in Albuquerque under the acronym "B-SAFER" (Bio-Surveillance, Analysis, Feedback, Evaluation, and Response). A basic HTML slide show describing the systems components and how they interact can be found here. "We had this stuff running for a while," Forslund says. "By running in Albuquerque we dotted i's and crossed t's. This is no longer an academic exercise. We had to make everything run, and we had an epidimologist sitting there, looking at our data."

Soon to be part of Homeland Security?

Forslund says CDC (Centers for Disease Control and Prevention) is interested, but that primary funding in the future will probably come from the new Department of Homeland Security. He says commercial software companies dealing with medical records have decided disease outbreak monitoring "was not where the money was since about 1999 or 2000. It has been quiet since then.

"After 9/11 it went especially quiet, just when it was needed. [The software] companies didn't see a market."

Now, of course, OpenEmed can be considered a vital part of our national security infrastructure, therefore well worth federal sponsorship. And if it happens to help detect a few natural disease outbreaks and alert medical people to them so that they can save a few lives, that's nice too.

Why Open Source?

The answer should be obvious, but here it is in Forslund's words: "By making it open source and free, we hope we can get it deployed to many sites, because the broader the base, the more sensitivity we have to a problem."

Not only that, now that the code is released there is no chance of the effort completely dying because the few proprietary software vendors that currently dominate the medical records market decide information sharing between different health care providers isn't profitable. Now, with or without them, OpenEMed can continue its basic mission: to prevent illness and save lives.


  • Open Source
Click Here!