On-site pharmacists in the ED improve medical errors

Objective The objective of the study was to compare errors in the emergency department (ED) with pharmacists present (PPs) for resuscitations and traumas vs with pharmacists absent (PAs). Our hypothesis was that errors would be significantly fewer during PP than PA times. We also hypothesized that times with PP would affect patients greater when disposition was to more critical areas (intensive care unit, or ICUs). Methods The study was conducted during a 3-month period in 2009 in a level 1 trauma center with an emergency medicine residency. This was a cross-sectional cohort study comparing a prospective analysis of patients during the time (10 hour/day) with PP and a retrospective review of the time on the same days (14 hours/day) with PA. Demographics of age, race, and sex were recorded. Patient disposition was either ICU, operating room, non-ICU wards, observation unit, or discharge. Main outcome was errors recorded including medications given but not ordered, medication ordered but not given, and time delays for medications. For demographics and prevalence, descriptive statistics and percentages were used. Percent differences and 95% confidence intervals (CIs) and χ 2 were derived. Logistic regression used predictor variables of age, race, sex, disposition, and presence or absence of pharmacists. An a priori power analysis was performed. The study was powered at 80% with 186 subjects per group (PP vs PA), to find a difference of 20% between the 2 groups in percent of medical errors. Results There were 694 patients included in the 3-month period. A total of 242 presented during PP times and 452 during PA times. There were 383 (55%) male, 301 (43%) female, and 10 (2%) unknown sex. Mean age was 45 ± 18 years in PP group and 48 ± 20 years in PA group ( P , nonsignificant). There was no difference in ethnicity between groups. There were 6 (3%) patients with errors recorded during PP times and 137 (30%) with errors recorded during PA times (difference, 27%; 95% CI, 23-32). Controlling for age, race, sex, and disposition, medical errors were 13.5 times more likely during PA than during PP times (adjusted odds ratio, 13.5; 95% CI, 5.7-31.9). Conclusion With pharmacists absent, over 13 times more errors are recorded in our ED than with pharmacists present. An on-site pharmacist in the ED may be helpful in reducing medical errors.