A study at the VA Medical Center in Salt Lake City underscores the need to develop physician order-entry systems that better model the clinical experience and provide doctors with more useful information, including patient history, said Jonathan Nebeker, lead author of the study and an attending physician at the hospital.
Researchers recorded high rates of adverse drug events related to medication selection, dosage and monitoring at the VA hospital during the final 20 weeks of 2000, even though the facility used a computerized order-entry system that automatically checked prescriptions for drug allergies and interactions.
The study identified 483 clinically significant adverse drug events-requiring a change in treatment plan-in 937 hospital admissions. Of those, 45 adverse events, or 9%, were classified as serious, leading to prolonged hospitalization, permanent harm, near death or death.
About one in four patients, or 241, experienced at least one adverse drug event, and many experienced multiple events.
Some 27% of the events resulted from errors, with the most frequent error being failure to provide prophylaxis for expected adverse drug reactions.
The study was published in the May 23 issue of the Archives of Internal Medicine.
"Many studies have shown that human beings can only simultaneously consider four independent factors in decisions," Nebeker said in an e-mail interview. "It is common to have to consider 10 to 20 simultaneous factors in patient medication decisions," he said.
Those factors include medication interactions, disease processes, patient history and patient preferences.
Nebecker praised the "excellent work" being done in developing computer systems to reduce medication errors but noted that most errors don't result in patient harm.
In the VA study, the most common outcomes of adverse drug events were constipation, hypo-kalemia and hypotension.
The drugs most often involved were narcotic analgesics, diu- retics, and cardiovascular and renal agents.
Computerized decision support should target the most troublesome types of events, Nebecker said. Among other elements, order-entry systems should automatically generate recommendations for prophylaxis and monitoring, suggesting potassium supplements, for example, when loop diuretics are prescribed.