PSNC publishes data challenging CPO’s claims about dispensing errors and automation
May 18 2016
PSNC has published an analysis of figures on dispensing errors and automated dispensing, which the Chief Pharmaceutical Officer had presented to MPs in March.
Dr Keith Ridge, CPO for England, had told a session of the parliamentary All-Party Pharmacy Group that “dispensing errors in community pharmacy occur in around 3% of all items dispensed, whilst large-scale automated dispensing would reduce the error rate to 0.00001%.”
However, he apologised last week for “inadvertently misleading” the MPs and clarified the figures which had been sourced from a single Swedish model of automated dispensing. PSNC has welcomed Dr Ridge’s clarification, and has now published a paper examining research into the subject, highlighting “statistics produced from the research and comments on their validity in assessing the impact of automation on reducing dispensing errors.”
Prior to the APPG meeting, the CPO had already advocated the use of automated dispensing. PSNC has quoted from a speech he gave on March 8, where is reported as saying:
“Deployed well and surrounded by a robust quality management system, probably approaching the rigour of GMP, then the level of error is reduced to a tiny fraction of current manual dispensing error rates. It then becomes in my view a professional obligation to ensure as much dispensing as possible is done through such equipment.”
PSNC also notes that the reference of a 3% dispensing error rate in the UK is thought to date from a paper published in 2007, based on research led by Bryony Dean Franklin.
“We believe quoting that figure in isolation, without a further explanation of the types of error identified in the research, could overstate the likely beneficial impact of automated dispensing systems on dispensing accuracy,” said PSNC. “It could also lead people to believe that the community pharmacy service is providing a less safe service than is actually the case, as the majority of the errors identified in the research were described as of minor clinical significance.”
Among the points challenged by PSNC is that the research by Professor Franklin is getting on for 10 years old and community pharmacy systems and practice have changed significantly in that time. The introduction of the Electronic Prescription Service will now prevent the majority of labelling errors that had been identified in the 2007 study. other labelling errors which EPS may not be able to prevent will arise due to human error either inputting or checking information in the dispensing process.
As the Franklin paper had identified 1.7% of the errors were content errors, and 1.6% were labelling errors, “we therefore conclude that the research provides no evidence to suggest that there would be a difference in labelling error rates between community pharmacies and automated dispensing systems. The only severe labelling error identified in the research would still have occurred in an automated dispensing system,” said PSNC.
As for content errors, most of these were incorrect quantities, generally unlikely to have any clinical significance. This was also in part due to a greater extent of dispensing from bulk packs in 2007 (so that an additional tablet may have counted as a dispensing error) compared to the much higher level now of original pack dispensing. When product bar codes are introduced on packs for authentication as part of a European-Union-wide Falsified Medicines Directive, pharmacy staff will be required to scan items before handing the item to a patient, offering a further dispensing check.
“The level of labelling errors (1.6%) identified in the research are now unlikely to occur at the same rate due to the introduction of EPS. Automated dispensing systems are unlikely to further reduce this type of error. Wrong quantity content errors are now less likely to occur due to the predominance of dispensing from original packs rather than from bulk containers,” concluded PSNC.