Immediate and long-term effects of an electronic medication management system on paediatric prescribing errors: A stepped-wedge cluster randomised controlled trial with one-year follow-up

Abstract

Abstract Objectives: To assess the immediate (first 70 days of use) and long-term (one-year) effectiveness of an electronic medication management (eMM) system to reduce prescribing errors, and their potential and actual harm in paediatrics. Design: A stepped-wedge cluster randomised controlled trial (SWCRCT) over 11 weeks (April–July 2016): 8 clusters randomised for eMM implementation. All medication orders during the trial and for a random sample of medication orders one-year post-eMM implementation (June–September 2017) were reviewed. Clinical prescribing errors were rated for the severity of potential harm on a 5-point scale. Errors (score ≥3 ie potential adverse drug events [ADEs]) identified in the SWCRCT were assessed for actual harm (actual ADEs). Setting: A 310-bed paediatric referral hospital in Sydney, Australia. Participants: Inpatients. Control and Intervention: Paper medication charts (control); eMM (intervention). Primary outcomes: clinical prescribing errors, potential ADEs. Secondary outcomes: procedural errors, clinical prescribing errors for high-risk medications and by error type, clinical errors associated with actual harm (ADEs) during the SWCRCT. Results: 35,260 medication orders for 4,821 patients were reviewed. In the first 70 days of eMM use there was no significant change in overall clinical prescribing error rates (Incident Rate Ratio [IRR] 1.05[95%CI 0.92-1.21], p=0.45). However, potential ADEs significantly increased by 62% (IRR 1.62[95%CI 1.28-2.04], p<0.001). One-year post-eMM, clinical prescribing errors declined by 36% (IRR 0.64 [95%CI 0.56-0.72], p<0.001) and errors associated with high-risk medications decreased by 33% (IRR 0.67 [95%CI 0.51-0.88], p=0.004) compared with rates pre-eMM. Dose, route, frequency and duplicate therapy errors accounted for over 90% of errors in each study period. Dose error rates were more than double that for any other error type. Few errors in the SWCRCT were associated with actual harm (paper 10.01/1000 clinical prescribing errors[95% CI 5.97-16.73]; immediately post-eMM 8.84/1000 errors[95%CI 6.04-12.92]). 71%[95%CI 50-86] of patients with actual harm experienced a dose error. Discussion: This is the first randomised controlled trial of the effects of eMM on paediatric prescribing errors. The first 70 days of use showed no overall improvement and an increase in some categories of errors. The results suggest immediate risks to medication safety during the early periods of system use. A year after implementation error rates significantly declined from baseline suggesting long-term benefits to safety. Optimisation of eMM should focus on features to reduce dose errors due to their high frequency and greater capacity to cause harm.