Health information technology (HIT) is increasingly being promoted as a medication error reduction strategy. Electronic prescribing and smart-pump technology are examples of HIT widely advocated in the hospital setting. In critical care, the risks associated with paediatric infusions have been specifically addressed with calls for the use of standard concentration infusions (SCIs) in conjunction with smart-pump technology. Evidence on the benefits of HIT in the paediatric setting remains limited. This study aims to assess the impact of both electronic prescribing and a smart-pump drug library of SCIs on medication errors in paediatric critical care.Methods
A retrospective, observational study based on an interrupted time series design was conducted in the 23-bed paediatric intensive care unit (PICU) of a tertiary children’s hospital. 3400 randomly selected medication orders were reviewed over 4 epochs: pre-implementation of either technology (Epoch 1); post-implementation of SCIs (Epoch 2); immediate post-implementation of electronic prescribing (Epoch 3); and 1 year post-implementation of both (Epoch 4). Orders prescribed during the study period were included provided they had undergone clinical pharmacy review. Intravenous fluids, epidural/regional blocks, total parental nutrition, chemotherapy and patient/nurse controlled analgesia were excluded. Medication error rates were calculated applying pre-specified definitions and inclusion criteria.1 Novel technology-generated errors were identified and defined using a modified Delphi process. Errors were graded for severity using a combination of two validated grading tools.2,3Results
Overall medication error rate based on all orders were similar in Epoch 1 and 4 (10.2% vs 9.7%; p=0.66). Altered error distribution was however evident. Incomplete and wrong unit errors were eradicated, but duplicate orders increased. Dosing errors remained the most common. 77% of pre-implementation errors were considered likely to be removed by the new technology. 24% of post-implementation errors were considered to be novel technology-generated errors. Examples included incorrect formulation selection and errors on altered electronic orders. In Epoch 2, the implementation of SCIs prior to electronic prescribing significantly reduced infusion-related prescribing errors (31.4% to 12.6%; p<0.01). An infusion error rate of 7.9% was reported post-implementation of electronically-generated standard infusion orders in Epoch 4.Conclusion
The overall medication error rate in PICU was largely unchanged by the introduction of electronic prescribing. Some errors disappeared but new errors directly attributable to the implemented technologies emerged. In the complex PICU environment, dosing errors remain common. A significant reduction in infusion-related errors was found as a consequence of the introduction of SCIs and smart-pump technology. The introduction of electronically-generated standard infusion orders brought further benefits. The results of this study show that the benefits of HIT in the paediatric setting cannot be assumed and highlight the need for further studies with increasing use of HIT in paediatric settings.