| Abstract | BACKGROUND: Despite the growing use of error reporting tools, the healthcare industry is inexperienced in receiving, understanding, and analyzing these reports. OBJECTIVE: To assess the accuracy and define the epidemiology of medication error reports. Design, setting, and patients: A retrospective cohort study of 581 error reports containing 1010 medication errors reported between July 2001 and January 2003 at a large academic children's institution. MAIN OUTCOME MEASURES: Correct classification and types of medication errors. RESULTS: Of the 1010 medication errors reviewed, 298 (30%) were prescribing errors, 245 (24%) were dispensing errors, 410 (41%) were administration errors, and 57 (6%) involved medication administration records (MAR). Following expert review, 208 errors (21%) were deleted because they had been inappropriately coded as errors and 97 (10%) were added as they were not initially coded despite having occurred. In addition, 352 medication error reports needed to have the subtype of error reclassified; 207 (59%) of these involved the reporter choosing the non-descript "other" category on the reporting tool (such as "Prescribing other") which was able to be reclassified by expert review. The overall distribution of error type categories did not change significantly with expert review, although only MAR errors were underreported by the reporters. The most common medications were anti-infectives (17%), pain/sedative agents (15%), nutritional agents (11%), gastrointestinal agents (8%), and cardiovascular agents (7%). CONCLUSIONS: Despite clear imperfections in the data captured, medication error reporting tools are effective as a means of collecting reliable information on errors rapidly and in real time. Our data suggest that administration errors are at least as common as prescribing errors in children. Further research is needed, not only in the area of computerized physician order entry (CPOE) for children, but also on ways to make the dispensing and administration of medications safer. |