Many reports on health information technology in the past few years tell a tale of a dream deferred — standard-setting deadlines blown, adoption targets undershot, groundswells of demand dematerializing. But in the alternate world of large, integrated health plans, where the better angels of our troubled system’s nature may be said to reside, some of the future promise of health IT has already arrived. The possibilities of a “rapid-learning health system” are laid out in detail in a collection of papers published online today by Health Affairs.

A few of these integrated health plans, which have fully deployed electronic health records up and running, have begun to use their systems to answer practical research questions on the effectiveness of new treatments that would take years to answer using such traditional methods as randomized controlled trials. At a time when the pace of scientific innovation has outstripped the capacity of the existing clinical research infrastructure, and while oceans of patient data sit in vast, disconnected repositories “untouched by human minds,” a golden opportunity to accelerate clinical learning presents itself. And some of these organizations — including the Veterans Health Administration, Kaiser Permanente, and the Geisinger Health System — have seized it.

For example, since its transformation into a “wired” system in the 1990s, the VHA has developed a diabetes registry that holds detailed clinical data on 600,000 patients, which allows the system to track and manage higher-risk patients well enough to improve clinical outcomes by large orders of magnitude. Similar public program potential exists to enhance postmarketing surveillance by the FDA and help Medicare use patient data to make coverage decisions.

Like health IT, evidence-based medicine has been touted by the health services research community for decades as the answer to the health system’s troubling quality failures and unexplained variability. The history of evidence-based guidelines has been a checkered one, in large part because of medicine’s resistance to standardized practice. But Paul Wallace’s account of the use of rapid-learning approaches to standardize cancer care at Kaiser Permanente shows that standards and variations are two sides of the same coin in real evidence-based practice, which can be used together to enhance both best practice norms and clinical discretion for individual patients — laying to rest the canard of “cookbook medicine.”

The possibilities for expansion and enhancement of rapid learning opportunities, and the policy implications of these new vistas, are outlined in an overview paper by Lynn Etheredge.

Click here to view a webcast of a briefing on the rapid learning collection, which will be available on Tuesday, January 30, at 12 p.m. EST.

Tags: Health IT