Aims: To evaluate the frequency and characteristics of preventable medication-related events in hospitalised children, to determine the yield of several methods for identifying them and to recommend priorities for prevention.
Methods: A prospective observational cohort study was conducted over a 12-week period on the paediatric wards at a university-affiliated urban general hospital in New Zealand. For all admissions of greater than 24 hours, medication-related events were identified using a multifaceted approach and subsequently classified independently by three reviewers (using a standardised reviewer form) by event type, type of error, stage of the medication process, and preventability.
Results: There were 495 eligible study patients, who had 520 admissions and 3037 patient days of admission, during which 3160 medication orders were written. Of 761 medication-related events reported during the study period, 630 (83.3%) were identified by chart review; 111 (14.6%) by a voluntary staff quality improvement reporting system; 16 (2.1%) by interview of parents; and 4 (0.53%) events via the concurrent routine hospital-incident reporting system. Excluding duplicate reports and practice-related issues, a total of 696 study patient-specific events were included in the analysis. Excluding the inconsequential events (trivial rule violation and 'other' categories), the majority [368/399 (92.2%)] of events were found to be preventable; comprising 38/67 (56.7%) ADEs, 75/77 (97.4%) potential ADEs, and all 255 (100%) harmless medication errors. Most commonly implicated in preventable ADEs and potential ADEs were, event rate (95%CI): improper dose and the prescribing stage-35 (29 to 42) and 74 (64 to 84) respectively per 1000 patient days; and antibacterial agents and the intravenous route of administration 21 (17 to 25) and 11 (10 to 13) respectively per 100 medication orders.
Conclusions: Preventable medication-related events occur commonly in the paediatric inpatient setting, and importantly over half of the events that caused patient harm were deemed preventable. Voluntary staff reporting in a quality improvement environment was found to be inferior to chart review for identifying events, but a vast improvement on the conventional incident reporting system. Most commonly implicated in the harmful or potentially harmful preventable events, and hence the best targets for prevention are dosing errors, particularly during the prescribing stage of the medication use process, and use of antibacterial agents, particularly when administered by the intravenous route.