The effect of Computerised Physician Order Entry on prescribing errors: An interrupted time-series study at a secondary referral hospital in Australia.

BACKGROUND:Computerised Physician Order Entry (CPOE) software is increasingly used across the world to improve medication safety. However, few high-quality studies have reviewed the impact of CPOE on prescribing errors and patient harm. OBJECTIVE:To investigate the effect of a hybrid CPOE-paper prescribing system on prescribing errors at a secondary hospital site. DESIGN:An interrupted time-series study was conducted by identifying prescribing errors via prospective medical chart review before and after the implementation of CPOE across three medical wards. PARTICIPANTS:The medication orders of all patients admitted to the medical wards during the study period were reviewed. INTERVENTION:Implementation of a CPOE across three medical wards. MEASURES:A blinded expert panel risk stratified the errors according to level of severity, preventability and potential for harm. Pearson's chi square and segmented regressions were used to determine if there were differences in prescribing errors pre- and post-CPOE implementation. KEY RESULTS:A total of 10,535 medication orders were reviewed pre-CPOE and 13,841 medication orders reviewed post-CPOE. Analysis demonstrated that after implementation of CPOE there were reductions in the proportion of orders with one or more of any error (-30.1%, 95 %CI: -36.5%, -23.7%, p < 0.001). Reductions in the proportion of orders with one or more errors were seen across the error categories of dosing errors (-20.1%, 95 %CI: -25.1%, -15%, p < 0.001), procedural/administrative errors (-18.9%, 95 %CI: -22.8%, -15%, p < 0.001), and therapeutic errors (-2.6%, 95 %CI: -4.1%, -1%, p = 0.002). Post-CPOE there were reductions in the proportion of orders with at least one non-intercepted serious error (-12.6%, 95 %CI: -16.4%, -8.8%, p < 0.001). CONCLUSION:The introduction of CPOE was associated with reductions in prescribing errors. There is also evidence that this translated into a reduced risk of harm to patients post-CPOE implementation through the reduction in actual adverse drug events.