Abstract Introduction Pharmacy automation using dispensing robots for the supply of medicines is not new. Improved dispensing efficiencies, maximised storage capacity and reduced dispensing errors are cited as key benefits.1 Published articles describe the impact of automation on dispensing error rates and types for original-pack dispensing2 with a paucity of literature on split-packs. Split-pack dispensing cannot be entirely eradicated: the ability to barcode and return to the robot facilitates the storage of medicines locally where dispensary footprint is limited; a manual process and one that is subject to human error. With anecdotal reports of split-pack robotic dispensing errors, there is an opportunity to explore near-miss errors as part of routine monitoring undertaken. Aim To establish the incidence and type of near-miss dispensing errors associated with split-pack robotic dispensing. Methods Approval was obtained from the Trust Pharmacy Research Committee. The need for ethical submission was waived. Near-miss dispensing errors were recorded in real-time via direct observations using a standardised data collection tool adapted from the Royal Pharmaceutical Society Near Miss Error Log3; with additional data collected on whether prescription requests were for original or split-pack quantities and storage location. All drugs dispensed/ checked in the dispensary between 09:00-17:00 were included (4 days in January 2023). All prescription types for individual patient medication orders were included. Stock drugs and outpatient prescriptions were excluded. A manual count of all items dispensed/ checked was undertaken each day to provide the denominator. Data was entered into Microsoft excel and analysed. Results Overall near-miss dispensing error incidence of 3.5% (72 near-misses/ 2035 items dispensed). Fourteen of fifty-eight (24.1%) erroneously dispensed items had two near-misses identified. Near-miss incidence associated with original-pack dispensing of 3.1% (50 near-misses/ 1590 items dispensed) compared with split-pack dispensing of 4.9% (22 near-misses/ 445 items dispensed). Chi square (with Yate’s correction) did not identify a statistical significant difference between the overall near-miss error rate between the two (p> 0.06). Most common original-pack dispensing error types were ‘wrong instructions on label’ (n=29/50), ‘wrong quantity on label and dispensed’ (n= 6/50), and ‘missing label’ on item (n=5/50). For split-pack dispensing, ‘wrong instructions on label’ (n=9/22), ‘wrong quantity dispensed’ (n=8/22), and ‘wrong quantity labelled’ (n=5/22) were most common. Discussion/Conclusion Local prevalence of near-misses is higher than in reported published literature1. Whilst direct comparisons are difficult to make due to differences in research methods, definitions, and operating systems; the incidence of near-miss errors associated with split-packs was notably higher than original-pack dispensing despite no statistical difference. Similar near-miss error types were identified for original/ split-pack dispensing with ‘wrong instructions on the label’ being the most common. This is consistent with published literature1 on errors with pharmacy automation. With split-pack robotic dispensing, it was noted that the quantity were not always as identified on the barcode label due to the manual returns process. Direct observations have indicated that missing patient information leaflets are a common problem for split-pack dispensing, although not recorded in this study. Further work should focus on causes/ contributory factors to split-pack robotic dispensing errors.
Read full abstract