We thought we would be perfect: medication errors before and after the initiation of Computerized Physician Order Entry

Abstract

Background: Because the Institute of Medicine demanded health care improvement, electronic medical records have been implemented with the hopes of eliminating iatrogenic injury caused by avoidable mistakes. Electronic orders and electronic medical records survived its initial slow adoption and have since had a myriad of identifiable flaws as it becomes incorporated nationally.

Materials and methods: This retrospective study at a university teaching hospital analyzed all medication order errors (OEs) for the 26 wk of paper-order entries before computer physician order entry (CPOE) and 26 weeks after CPOE was initiated. All OEs were included and documented by month as well as severity using standard taxonomy.

Results: Results indicated that CPOE yielded a significant increase in overall medication OE with five of six severity categories remaining the same or increasing in OE. Severity categories A and E saw a significant increase once CPOE began (P < 0.01). Pre-CPOE OEs were 1741, whereas Post-CPOE OEs were 2226, showing an increase in overall medication errors (P < 0.01). After CPOE began, the cumulative successive errors recorded were 112, 290, 267, 307, 412, 399, and 439 with an R(2) value of 0.849 and a P value of 0.003 in the analysis of variance to test regression relation.

Conclusions: As CPOE adapts for its real-world applications, it may eventually prove useful in reducing errors; however, perfection and error free order entry will not be achieved unless objective data analysis guides its evolution.

Keywords: Computer physician order entry; Medication error; University hospital.