Investigating Medication Prescribing Accuracy for Critical Error Types (iMPACT)

Background: The definition and classification of prescribing error varies greatly as the reporting of prescribing errors is subjective and dependent on the experience and knowledge of the clinician auditing the prescribing. There is also no standard data collection technique for these errors and so it is difficult to benchmark healthcare settings and recommend strategies for quality improvement. A list of 80 high risk and/or frequently occurring prescribing errors in secondary care were produced using a two stage consensus technique1. All of the errors had the potential to be prevented by electronic prescribing (eP) and clinical decision support (CDS). The prescribing indicators spanned a number of categories, which were allergy, clinical contraindications, dosing, drug-drug interactions, frequency and route.

Objective: Our aim was to create a tool that would standardise the collection of high severity and/or high frequency prescribing errors in general medicine in secondary care. The prescribing errors captured would be amenable to CDS to ensure that the rate of occurrence of the prescribing errors could be determined before and after the implementation of eP with CDS.

Method: The iMPACT data collection tool was developed as a standalone Windows application that can be accessed on any computer terminal at a hospital site. Auditors are asked to review inpatient medication chart(s) and complete the questions posed within the software. To comply with information governance, minimal patient information is collected (patient initials, age and admission date) and the data stored locally. The documentation of demographic data such as weight, allergy status and completion of venous thromboembolism (VTE) risk is captured. The auditor is then asked to review the medications prescribed and if it appears in the software, select the drug or drug class. A pop up box then appears for each medication selected, which asks one or more closed questions about the prescription to determine whether one of the 80 prescribing indicators has occurred.

Results: Hospital Trusts participating in the study are asked to review a minimum of 4000 inpatient orders pre and post-implementation of eP. The data is exported via NHS email to the University Hospitals Birmingham (UHB) for analysis, with the patient’s initials removed to comply with data protection. An analysis of the indicators is conducted to provide a summary of: 1) The completeness of the documentation process; 2) The number of indicators that occurred during the study period; and 3) The type of indicators that occurred. Each site is provided with a summary of the data to inform quality improvement and system development.

Conclusion: There are currently ten English NHS Trusts auditing prescribing practice prior to the implementation of eP. The iMPACT tool can not only be used to quantify any change in the rate and type of prescribing error as a result of quality improvement initiatives such as ePrescribing, but can also be used to monitor change in prescribing habits over time.

1. Thomas SK, McDowell SE, Hodson J, Nwulu U, Howard RL, Avery AJ, et al. Developing consensus on hospital prescribing indicators of potential harms amenable to decision support. British Journal of Clinical Pharmacology. 2013;76(5):797-809.