High rates of adverse drug events in a highly computerized hospital

From Clinfowiki
Jump to: navigation, search

Multiple studies have documented the prevalence of medication errors leading to a significant number of adverse drug events (ADEs). These ADEs have a significant impact on the length of hospital stay, total cost of health care and loss of life. Since these findings became public, it has been suggested that computerized patient care systems could prevent most medication related errors leading to improved health care and cost reductions. Recently, the value of these computerized systems has been questioned but the needed data is mostly lacking. Researchers at Salt Lake City’s VA Medical Center decided to investigate this issue. Their medical center has already implemented computerized systems for physician order entry, laboratory results, electronic health records and medication administration. Although their system did check for drug-allergy, drug-drug and some limited drug-disease interactions, it did not provide more advanced decision support features for drug selection, dosing, monitoring and warnings or advice.


During a study period of 20 weeks, 937 patients (41% of admissions) were randomly selected for prospective review for ADEs. Adverse drug events were defined as injury resulting from the use of a drug excluding self-limited harm and other ADEs that were not considered severe according to the FDA. They identified 483 clinically significant ADEs corresponding to 52 events per 100 admissions. Adverse drug reactions accounted for (93%) of the ADEs and the remainder (7%) was due to overdosing or underdosing of medication.

The adverse reactions were further broken down into (90%) dose dependent and (10%) idiosyncratic in nature. Also, the ADEs were due to additive drug-drug interaction in (39%) of the cases and due to drug-condition interaction in (29%) of the cases. Twenty seven percent of ADEs were due to errors, and these errors were categorized as execution errors in (27%) of the cases and planning errors in the remainder 83%. The errors included failure to provide prophylaxis for expected adverse reactions (36%), failure to adequately monitor for common adverse reactions (33%), prescribing improper dose (33%) and inappropriate medications (7%).

The incidence of ADEs in this study is significantly higher than previous reports. This was attributed to thorough review of legible electronic medical records by clinical pharmacists. Also, it was noted that the error at the transcription stage (0%) and the administration stage (11%) were lower than previous reports.


This study showed that electronic medical systems that do not provide for adequate decision support are not likely to reduce the incidence of ADEs during the initial planning and ordering stages. This highlights the importance of considering these features before purchasing a computerized physician order entry system. An ideal electronic system should adequately address all the stages involved in medication ordering and administration, and provide the necessary tools to minimize ADEs.