VHA to test prediction systems for care improvement

The Veterans Health Administration wants to determine the "value and fitness" of new computerized clinical reasoning and prediction systems, according to a statement of work issued to potential vendors this week.

According to VHA, such systems must be able to process both structured and unstructured data in a timely manner to "obtain optimal answers to clinical questions." VHA plans to test the ability of such systems to break down natural language questions from physicians for the purpose of detecting previously unknown trends, identifying patterns and developing and using predictive models.

"For several years, the VHA Office of Informatics and Analytics has been monitoring the maturation and breakthrough successes of reasoning and prediction class systems, in the context of an overwhelming growth of the medical literature corpus," the statement says. "Concurrently, VHA is accumulating significant amounts of clinical data. The technology advances, relentless increases in medical literature, and the staggering volume of clinical data we collect have motivated VHA to perform an in-depth, first-hand assessment within our environment of the value clinical reasoning and prediction systems may offer."

For the assessment, VHA will install such a system at the U.S. Department of Veterans Affairs' Austin (Texas) Information Technology Center; the work is expected to last no more than 24 months.

The system will be expected to support up to 100 authorized VA users, as well as 100 complex "patient case" results per hour, with a three minute response time for each case.

In related news, VA Chief Information Officer Roger Baker told reporters this week that the secretaries of Defense and Veterans Affairs will meet next week to approve a software development approach that will enable an accelerated rollout of an integrated electronic health record, FierceGovernmentIT reports.

To learn more:
- read the VHA's statement of work (.pdf)
- here's the FierceGovernmentIT article