Futuristic Clinical Decision Support Tool Catches On
IndiGO healthcare software, which predicts which patients will develop chronic conditions and the best interventions to head off disease, signs users.
5 Key Elements For Clinical Decision Support Systems
(click image for larger view and for slideshow)
After being tested successfully at Kaiser Permanente, a futuristic decision support tool from San Francisco-based Archimedes is beginning to build traction elsewhere. Archimedes' application, known as IndiGO, applies algorithms to clinical databases to predict the chances of patients developing a serious chronic condition. It also shows patients and doctors which interventions have the best chance of heading off a heart attack, a stroke, or another adverse health event.
The MyAccess Health Network in Tulsa, Okla., one of the 17 Beacon Communities that have received health IT grants from the Department of Health and Human Services, last week agreed to use IndiGO. Physicians could ultimately use IndiGO "to inform the health decisions of as many as 810,000 Oklahomans," according to the announcement. The MyAccess Network is a coalition of 50 provider organizations and Blue Cross and Blue Shield of Oklahoma.
The Colorado Beacon Consortium, another Beacon Community, also recently announced it would use IndiGO; Fairview Health Services, a big provider organization in Minneapolis, has already adopted it. Meanwhile, Kaiser Permanente plans to expand the use of IndiGO to its Southern California division.
Archimedes, which was cofounded by David Eddy, MD, an adviser to Kaiser Permanente and a leading expert on evidence-based medicine, developed IndiGO in late 2008. The next year, Kaiser began piloting and studying the decision support tool in its Hawaii organization.
The results of that soon-to-be-published study were positive, said Peter Alperin, MD, VP of medicine for Archimedes, in an interview with InformationWeek Healthcare. Doctors liked IndiGO and were able to incorporate it in their workflow, he noted. The patients who used IndiGO in one Kaiser clinic found it helped them make better healthcare decisions and were more compliant with their doctors' recommendations as a result.
Even more significant, according to Alperin, was the fact that the risk of developing coronary artery disease (CAD) in these patients dropped 22%, on average, during the nine-month trial of IndiGO. By comparison, patients' risk of CAD declined only 9% in another clinic that did not use the tool.
IndiGO derives its predictive power from algorithms developed through mathematical modeling of large public databases. Included were clinical trial data from the National Institutes of Health and longitudinal data from long-running studies such as the Framingham Heart Study. Kaiser Permanente provided the only private data that Archimedes used, said Alperin.
When a patient visits a doctor who uses IndiGO, the tool is applied to the data in his or her electronic health record. While it helps to have some longitudinal data, all that is required to activate the algorithms is information from a single point in time, Alperin said. In addition, the program may consider claims data that shows a patient's past medical history.
Based on that data, IndiGO will assess the risk of a patient having a heart attack or a stroke, developing diabetes, or acquiring some other chronic disease over the next five years. In addition, it will rank the interventions that could be used to ameliorate that risk in order of their potential benefit to the patient.
Many patients, of course, have more than one health condition, and these comorbidities can influence the outcome. Alperin noted that IndiGO's "risk equations" are based on datasets that include patients with these comorbidities, which are factored into the predictions and the benefits of interventions.
IndiGO doesn't adhere to evidence-based guidelines, because not all patients follow the expected disease paths, Alperin said. For example, hypertension is a factor in heart disease and stroke, but not all patients who will develop those conditions have high blood pressure. "You want to see which factors in any given patient cause these events. Often, conventional markers are missing in particular patients at risk for a particular disease."
Alperin added, "We don't advocate that people abandon traditional guidelines. Our algorithms analyze the same data that evidence-based guidelines are based on. We're just analyzing it in a different way."
When are emerging technologies ready for clinical use? In the new issue of InformationWeek Healthcare, find out how three promising innovations--personalized medicine, clinical analytics, and natural language processing--show the trade-offs. Download the issue now. (Free registration required.)