Health Information Technology

Computerized drug-laboratory interaction alerts change doctors' prescriptions

A new study shows that when a computerized order entry system uses a clinical-decision support system to provide alerts about laboratory-drug interactions to outpatient care providers, they respond to the alerts and use the information to improve patient care. After receiving an alert, doctors either changed the dosage or stopped the order for a medication if it would worsen the problem indicated by the lab results, or they ordered laboratory tests to monitor potential problems for a drug.

The study, supported by the Agency for Healthcare Research and Quality (Contract No. 290-00-0014), focused on drug-laboratory interactions for 18 high-volume and high-risk medications that can lead to abnormally high or low blood potassium levels, kidney toxicity, low blood platelet count, or liver inflammation. Andrew W. Steele, M.D., M.P.H., and Sheri Eisert, Ph.D., of Denver Health, and colleagues compared outpatient clinic provider medication and laboratory orders at baseline and after implementation of the drug-laboratory interaction alerts.

During the post-alert period, an alert was displayed 11.8 percent of the times a potential drug-laboratory interaction was processed, with 6 percent for only abnormal laboratory values, 5.6 percent for only missing laboratory values, and 0.2 percent for both types of alerts. Alerts for an abnormal laboratory result significantly increased the percentage of time that clinicians did not complete the medication order (5.6 percent at baseline vs. 10.9 percent post-intervention) or the ordering of a recommended laboratory test (39 vs. 51 percent). Overall, ordering of the appropriate laboratory test increased by 33 percent with the presentation of the alert.

See "The effect of automated alerts on provider ordering behavior in an outpatient setting," by Dr. Steele, Dr. Eisert, Joel Witter, M.D., and others, in the September 2005 PloS Medicine 2(9), pp. 864-870.

Return to Contents
Proceed to Next Article