An Exploratory Study of Quality-related Events Reported by Community Pharmacies in the United Kingdom: Applying Association Rule Mining

Abstract

Errors in healthcare systems can cost resources and sometimes lives. Studying errors in healthcare can help to improve service quality. This study aims to explore the prevalent patterns in medication errors in UK’s community pharmacies by using error data voluntarily reported from 859 community pharmacies to incident management system managed by Pharmapod Inc. from 2015 to 2018. The primary output of the study is three sets of association rules that can characterize the relationship of error-factors, error-categories, and event-types. The sample data of 72733 events of medication errors were analyzed using association rule mining. Three sets of strong rules are identified, including six strong rules from error-factors to event-type, nine strong rules from error-categories to event-type, and twenty-two strong rules from the combination of error-factor and error-category to event-type. The implications of the findings are discussed with the literature.