SP5 Tenfold medication errors in children – Welsh paediatric surveillance unit study 2017–9

Abstract

AimsTo establish the incidence and characteristics of tenfold or greater and a tenth or less medication errors in children <16 years in Wales to help inform patient safety on a population level.MethodPopulation-based incidence study in Wales, UK, from June 2017 - May 2019 (24 months). Cases were reported from paediatricians and hospital pharmacists using the monthly Welsh Paediatric Surveillance Unit (WPSU).Results46 confirmed incidents in 44 children from 63 notifications were identified. Cases came from 8 hospitals in Wales with 29 (63%) from the sole tertiary hospital. Median age was 1.7 (range 1 week to 15) years and weight 10 kg (0.6 to 59).39 (85%) were overdosing (up to 1000x fold error) and 7 underdosing. 40 different medications were involved, 16 (37%) intravenous. Of 29 cases involving enteral medication, 26 (90%) were liquid formulations. Three cases were discharge medication prescribed or dispensed incorrectly and administrated at home. Stage of errors were primarily in prescribing 37 (80%), administration 7 (16%) and dispensing 2 (4%).18 (42%) cases reached the patient, 10 from prescribing. Seven cases were spotted after multiple doses were given. Six errors resulted in harm, three which required intensive care treatment. No deaths or permanent disabilities were reported. Half (23/46) of all errors reported and two-thirds (12/18) of cases that reached the child occurred in <10 kg children.Several human factor themes were identified: Prescribing confusion between gram milligram and microgram (none reached patient, n=7), confusing between mg and mg/kg (n=6 including 3 underdosing errors), leading zero errors (e.g. 0.1 vs 0.001 mg, n=6) and prescribing reconciliation errors where admitting doctor attempted to prescribe chronic medication in mg by reversing calculating liquid dosage expressed in mL (n=4).During this study period 164,000 hospital admissions occurred in children <16 years in Wales. Our data estimates a tenfold error incidence of 1:3600 paediatric admissions, with drug reaching the child in 1:9000 admissions.ConclusionIn this unique first ever population surveillance study, tenfold errors in children occurred at every stage of medication process and in the full range of care settings. Errors found were very different from those obtained from tertiary hospital single centre study and UK National Reporting and Learning System (NRLS). Strategies for error reduction will be more productive if designed across a whole national healthcare system.