Today’s First-Generation Decision Support Systems: Not Yet Able to Turn Doctors Into Sheep

Inside Healthcare Computing has graciously agreed to make previous Mr. HIStalk editorials available from its newsletter as a weekly "Best Of" series for HIStalk. This editorial originally appeared in the newsletter in July 2006. Inside Healthcare Computing subscribers receive a new editorial every week in their Electronic Update.

My colleague Ross Koppel, a sociologist and Penn professor, wrote an editorial in The American Journal of Managed Care (released today) titled “Defending Computerized Physician Order Entry From Its Supporters.” In it, he stresses that CPOE and clinical decision support systems (DSS) are separate systems, despite popular perception. Their implementation is often divergent and their benefits and shortcomings confused (or intentionally misrepresented).

Ross is right, and his sociologist’s view is important to our little world of geeks and IT-friendly doctors. We’re expecting a lot from immature CPOE and DSS systems that most hospital executives can’t define, even when they’re plunking down hard-earned capital dollars for them.

(I should mention that Ross wrote another article awhile back that riled up vendors, consultants, and HIMSS, in which he described one hospital’s increased error rate with CPOE implementation, finding that his one, small discouraging word was met with choruses of indignation from the “CPOE is Nirvana” crowd.)

CPOE is a smart typewriter that, standing alone, has little ability to improve patient outcomes. It prevents transcription errors, although those seldom harm patients because they’re caught anyway. CPOE makes it easy to choose common order defaults instead of “winging it.” Beyond that, the benefits (both clinical and financial) come from DSS, not CPOE, even though the hospital executives signing a multi-million CPOE deal as their cornerstone of patient safety automation probably missed that point completely.

DSS systems are, unfortunately, mostly frightfully immature, even more so than CPOE. Early adopters share war stories of sky-is-falling alerting, inflexible third-party rules, the inability to customize and personalize, and performance-sapping rules engines incapable of delivering alerts of any more sophistication than the old hard-coded screen edits.

Still, the real problem is right down Ross’s alley. Hospitals usually buy CPOE and DSS because they’ve failed to control physician behavior otherwise, often euphemised as “reducing practice variation” or “practicing evidence-based medicine.” They want software to do the dirty work that they can’t or won’t: telling physicians that they’re wrong and forcing them to change. When docs don’t follow the new cookbook medicine rules any better than the old ones, CPOE and DSS get the blame and everyone involved in the project pretends to have been somewhere else when the vote was taken to buy it.

I’ve been involved in two CPOE/DSS implementations, both involving large IDNs and well-known vendors. In both cases, hospital administration ill-advisedly shot their patient safety technology wad on CPOE, confident that it would improve patient care better than any other investment. Physician adoption was universal in one, minimal in the other, but one element was common to both: 90% of the expected DSS benefit never materialized. The carefully but naively drawn up list of post-implementation metrics was hidden away once everyone realized that we hadn’t really changed anything of importance for our multi-million dollar investment. We had bought ourselves a smart typewriter.

No software contains a switch that turns resistant physicians into docile, rule-following sheep who make better decisions under the watchful eye of Big Brother’s can’t-miss medical guidelines. But if your hospital has already spent a few million on CPOE and DSS thinking that was the case, you’ve learned that already.

Maybe the next generation of systems will offer value that physicians recognize. After all, they want the best outcomes for their patients, too. Where they disagree is that we have the answer right now with these first-generation CPOE and DSS applications.

This editorial is copyright-protected by Algonquin Professional Publishing, LLC., publishers of Inside Healthcare Computing. Please do not copy, forward, or reproduce this material without prior permission.  To obtain permission or for more information about Inside Healthcare Computing’s reprint policy, please contact the Customer Service Department at 877-690-1871 or go to http://insidehealth.com/ihcwebsite/reprints.html.

Mr. HIStalk’s editorials appear each Thursday morning in the subscribers-only version of Inside Healthcare Computing’s E-News Update.  To subscribe, please go to:  https://insidehealth.com/ihcwebsite/subscribe.html or call 877-690-1871.