Much press attention was aroused by the publication of a follow up to a 2009 study by the General Medical Council (GMC) that found one in 20 prescriptions contained either a prescribing or a monitoring error, affecting one in eight patients. A ‘severe’ error was found in one in 550 prescriptions. The results are interesting, and well summarised here.
The PRACtICe Study sampled records from 15 GP practices and found that nearly one-third of errors involved “incomplete information on the prescription” and that children and people of 75 or over were nearly twice as likely to have an error on their prescriptions as those aged 15-64 years. Underlying causes included factors related to the prescriber (including undergraduate therapeutic training), to the patient, to the team (poor communication was one), to the working environment (time and workload pressures), to the task, to computer systems (generally seen as positive in preventing error but in some cases causing problems) and the interface between primary and secondary care (variation in quality of hospital communication, willingness of GPs to prescribe medicines recommended by specialists). Interestingly, GPs did express concern about what one called the ‘quasi-autonomous role’ of nurses, with GPs being asked to sign prescriptions for patients they had not assessed. The study authors say that this was not one of the highest risks, but they did pick up a small number of errors associated with this practice, and they comment that, “if nurses are to issue prescriptions based on their own clinical assessment practices it would seem most sensible (and in line with current legislation) for them to be qualified as non-medical prescribers”. Training more non-medical prescribers, particularly pharmacists who might be in a good position to manage complicated regimens, is also suggested as a possible way of adjusting GP workload levels.
And were these all ‘GP prescriptions’? Well, of the 6048 prescriptions included in the study, most were issued by GP partners, with a mere 60 being attributed to non-medical prescribers.