Statistics from Altmetric.com
Editor,—We welcome the coverage given to the major, and potentially fatal, problem of medication errors within managed health care.1 We disagree, however, with the key message that medication errors are uncommon. They are endemic, extremely common, overlooked and often ignored.
Observational studies of medicine administration within hospitals in the UK report an error rate of 3% to 8%.2 In contrast, Ross et al report 195 errors, collated from a mandatory error reporting policy, in 65 months.1 While mandatory reporting is a commendable principle, the reality remains that the majority of healthcare professionals will not report errors, and the majority of medication errors, will not be reported.
Reasons for lack of reporting by nursing staff include confusion regarding the definition of drug errors and the appropriate action to take when they occurred, fear of disciplinary action, loss of clinical confidence and variation in managerial response.3
Voluntary, non-punitive error reporting programmes have been advocated as the most effective way to promote candid disclosure of medical error.4 Unless we are aware of what errors occur, we cannot expect to implement an appropriate system fix.
We suggest that the occurrence of three errors/month, represents a tremendous under reporting of the extent of medication error.1 If patient through put totalled 335 835 patient bed days, and we assume that each day the average patient received 6 doses of medicine, an error rate of 5%, suggests that a more realistic interpretation of the extent of the error iceberg is an incidence of greater than 100 000.
The conclusion therefore that medication errors are uncommon is unfortunately not true. The reality is thatreported medication errors are uncommon.
Dr Ross responds
Editor,—We welcome the debate stimulated by our paper. Indeed, this was our aim in publishing it. We agree with Mr Caldwell that a degree of under reporting is likely. Our system provides a clear definition to all staff of what constitutes a reportable medication error (listed in the appendix). It does not include errors that are averted—such as, misprescribed errors corrected by pharmacists before dispensing. We also noted that error reporting rates vary widely in the literature. We discussed some of the reasons advanced to explain such variations—for example, whether the reporting system is mandatory or voluntary, and the intensity of the search for errors. However, the published evidence about medication error rates in paediatric settings is very limited especially in the context of a nationally funded, universal, health care system like the NHS. There is, therefore, little firm paediatric evidence to support Mr Caldwell's opposite view that errors are “ . . .extremely common, overlooked and often ignored.”
In our experience, most reported errors were minor. Serious events with adverse outcomes were uncommon and, we think, are unlikely not to be reported. If anything, we would suspect it is minor errors that are most likely to go unrecorded. This may be of considerable importance if analysis of minor events highlights system problems whose correction may help avoid future serious incidents.
Mr Caldwell suggests that voluntary systems may increase error reporting. It needs to be recognised that voluntary systems are not a panacea but may also detect only a fraction of overall errors. Again, we would suspect that minor errors might be those most likely to be missed. The thrust of the editorial by Cohen seems to relate to errors with serious adverse outcomes.1-1 There are also some potential difficulties with voluntary systems. For example, how we do we ensure that parents are notified about error occurrence if reporting is voluntary? What happens about errors of such seriousness that issues of criminal negligence arise?
Whether a reporting system is mandatory or voluntary is probably less important than that the system is non-punitive. This is borne out by the findings of Vincer and colleagues1-2 who found an approximately four to six fold increase in error reporting when the punitive aspects of the form were reduced by making it an “incident” rather than an “error” form. We have no doubt that the critical challenge for us all is to make the shift from assuming “errors” arise from individual negligence to recognising that “incidents” more usually arise because of systemic organisational failures. We urgently need to move away from a culture of fixing the blame towards one of recognising and fixing the problem.