Bar codes on medication have been hailed as a remedy for drug errors in hospitals. Medication errors are not a small problem; the Institute of Medicine in 2006 estimated that at least 400,000 preventable medication mistakes occurred annually in U.S. hospitals, leading to about 9,000 deaths and $3.5 billion in extra cost. About one-third of U.S. hospitals now use bar codes to track medications and match up patients and their medicine. Most other hospitals are expected to introduce bar coding in the next few years.

But an interesting story by Josh Goldstein in today's Philadelphia Inquirer, however, reports that what works for your local grocery store doesn't always work so well on the hospital floor. A system that seems great on paper can be flawed in practice if, say, the computer can't be taken into an isolation room (meaning it's down the hall where the nurse can't hear the beep-beep alert of an error) or nurses who find ways of working around the bar code system, for instance, when they have to retrieve insulin for four different patients from a refrigerator on a different floor, scan in the code, and then carry all the medicine on one single tray back to four patients all awaiting different doses.

The piece summarizing findings by researchers at the Universities of Pennsylvania and Wisconsin is quite interesting. It hadn't ever occurred to us, for instance, that the simple technology could be so mismatched with nurses' actual workload and habits that they'd find ways of basically outsmarting it. But we wish Goldstein didn't bury the good news. If you read to the end of his story, you find out that most of the problems are fixable. Not just in theory but in practice. By following eight key recommendations, he wrote, four of the five hospitals studied were able to greatly improve the system and reduce the errors.

The most important thing that hospital executives can do to make the bar code systems work as intended, the article continues, is to "spend time actually watching how the systems are used and to demand that the vendors design systems for the real world of care."

Simple fixes. Solved problems. Reduced costs. Saved lives. Sounds good to us. Wouldn't it be nice if all the rest of the health care cost-coverage-quality challenges in our system were so simple?