Long-term reduction in adverse drug events: an evidence-based improvement model.

OBJECTIVES: To develop and test an evidence-based model for reducing medication errors and harm in hospitalized children.
METHODS: Prospective interrupted time series study evaluating the effectiveness of a multifaceted, staged intervention over 4 years in a major urban pediatric referral hospital. Guidelines for safe pediatric prescribing were implemented by using an evidence-based model. Key components included early clinician engagement and improved multidisciplinary communication, consensus development, interactive education, and timely data feedback by using iterative Plan-Do-Study-Act cycles. Impact on medication error and harm (adverse drug events, [ADEs]) was measured by using standard definitions and a multimethod approach. Prospective data from voluntary reports by nursing, medical, and pharmacy staff and intensive chart review were combined. All data were reviewed by a multidisciplinary panel, including causality assessments for ADEs.
RESULTS: Reviewed over 3 time periods were 1011 patients with 6651 medication orders. Total ADEs decreased by > 50% in the first year and this was maintained at 4 years. Greatest improvements were in potential ADEs, which decreased from 12.26 per 100 patients at baseline to 4.60 per 100 patients at 4 years (P < .05). Total medication errors decreased from 4.51 per 100 orders at baseline to 2.78 per 100 orders at 4 years (P < .05). Prescribing errors decreased by 65%, from 4.07 per 100 orders at baseline to 2.05 orders at 4 years (P < .05).
CONCLUSIONS: A multifaceted, evidence-based model for safe prescribing guideline implementation, engaging multidisciplinary clinicians, was effective in reducing medication error and harm in hospitalized children, resulting in sustained long-term improvement.