CHICAGO -- Errors in drug ordering, dosage and monitoring that may have serious consequences for patients persist in hospitals even after the adoption of computerized medication systems, according to a study in the May 23 issue of Archives of Internal Medicine, one of the JAMA/Archives journals.
Several broad-based studies during the past 15 years have demonstrated that injuries resulting from the use of a drug, called adverse drug events (ADEs), account for up to 41 percent of all hospital admissions and more than two billion dollars annually in inpatient costs, according to background information in the article. Many of these studies have also suggested that as much as a quarter of inpatient adverse drug events might be prevented through the use of computerized physician order entry (CPOE) and related computerized medication ordering and administrative systems.
Jonathan R. Nebecker, M.S., M.D., of the Veterans Administration Salt Lake City Health Care System, Salt Lake City, and colleagues conducted a daily review of the electronic medical records from a random sample of patients admitted to a VA hospital during a 20-week period in 2000. Because the Veterans Administration Healthcare System is a leader in implementing multiple computerized interventions, including computerized physician order entry, bar code-controlled medication delivery, a complete electronic medical record, automated drug-drug interaction checking and computerized allergy tracking and alerting, the authors suggest that a VA hospital offers an appropriate setting for testing the impact of computer systems on adverse drug events. Pharmacists used standardized criteria to classify adverse drug events.
Among 937 hospital admissions, 483 clinically significant inpatient adverse drug events were identified. An adverse drug event was considered clinically significant when a change in the patient's treatment plan was required. There were 52 adverse drug events per 100 admissions. One quarter of the hospitalizations had at least one adverse drug event. Nine percent of the adverse drug events resulted in serious harm, the other 91 percent were classified as moderate, requiring monitoring, interventions or discontinuation or adjustment of the dose of the problematic drug. The most common error types were failure to provide for expected adverse drug reactions, for example, prescribing potassium with diuretics to avoid a low potassium level (36 percent); failure to start or complete adequate monitoring for common adverse drug reactions (33 percent); and prescription of improper doses (33 percent) or inappropriate medications (seven percent).
The researchers found errors occurred at the following stages of care, 61 percent ordering, 25 percent monitoring, 13 percent administration, one percent dispensing and 0 percent transcription. The authors note that the computer system was successful in eliminating problems reading physicians' orders (transcription), but did not resolve the other problems associated with administering medication, drug selection, dosage and monitoring.
"In conclusion, our study found high rates of clinically important ADEs related to problems in drug selection, dosage, and monitoring in a VA medical center after the adoption of computerized systems that offered minimal decision support for these specific aspects of the medication process," the authors write. "...our findings do imply that purchasers of CPOE systems should not rely on generic CPOE and bar code medication administration systems alone to dramatically reduce ADE rates. Rather, health care organization desirous of preventing ADEs should consider whether candidate computerized medication systems offer decision support functions that address the most troublesome aspects of the medication administration process."
(Arch Intern Med. 2005; 165:1111-1116. Available post-embargo at www.archinternmed.com.)
Editor's Note: This study was supported by grants from the Veterans Administration Health Service Research and Development Service, Washington, D.C., the Geriatric Research, Education, and Clinical Center and the Salt Lake Informatics, Decision Enhancement, and Surveillance Center, Salt Lake City, and the VA Cooperative Studies Program, Albuquerque, N.M.
Cite This Page: