Kaiser, diabetes group team up for online project

The proposed Internet site will let diabetics forecast their health by answering a set of questions.

By Tyler Chin, amednews staff. July 14, 2003.

The American Diabetes Assn. and Kaiser Permanente are developing a Web site that will let diabetics virtually predict their diabetic future, arming patients, the developers hope, with information that can be used to make better health care decisions.

The site, which will be free and is expected to become available within two years, will ask those with and at risk of developing diabetes to answer a comprehensive set of questions relating to age, sex, height, race, body mass index, glucose count, cholesterol level, family history, and medication types and dosage.

See related content
Earlier about Kaiser

Based on the answers, Archimedes, a computer modeling software, will simulate what will likely happen to individual users within whatever time frame the users select, said Richard Kahn, PhD, chief scientific and medical officer of the Alexandria, Va.-based American Diabetes Assn.

The idea is that after patients see what the future holds for them, they can change variables and use the computer software to learn what steps they can take to live healthier and longer lives, Dr. Kahn said.

That's because Archimedes models the outcomes physicians and patients can expect from different treatment options, said David Eddy, MD, an adviser to Kaiser Permanente and co-developer of the predictive software program.

The way Archimedes works is akin to how Sim City, a popular computer game from Electronic Arts Inc., works, Dr. Kahn said. Sim City players create a fully functioning virtual city by building roads, hospitals, industry, etc. What sort of city they create will depend on the things they build.

"Say that Archimedes tells [patients] that in five years they are likely to be dead," Dr. Kahn said. "Then Archimedes says, 'Let's look at what happens if they change something about themselves -- reduce their blood pressure, their lipids, glucose; start taking aspirin; move closer to the hospital or take drugs more often -- this is what will happen.' Then the person and their doctor can actually be more intelligent about planning treatment."

Project participants said they are confident Archimedes' predictive power is highly accurate because they have validated the software against more than a dozen clinical trials.

Still, Dr. Kahn acknowledged that he expects some doctors and patients will resist using Archimedes because they question its accuracy or think that they know more than computers.

In any event, Dr. Kahn believes the question of user resistance is a moot point. "People will use the Web site if they feel there's value to it. If they don't, they won't. It's not like we're trying to sell things to someone. It's free."

Oakland, Calif.-based Kaiser, the American Diabetes Assn. and the New York-based pharmaceutical company Bristol-Myers Squibb Co. are helping fund development of Archimedes. Gateway Inc. Potway, Calif., will donate the computer processing power the simulations will require.