Features: Mistakes happen
Of the total orders checked, 50,016 were written by FY1 doctors with an error rate of 8.4%.
27 November 2009
(But how can we avoid them when prescribing?) Research shows how prescribing errors can be avoided.
‘…you’re bleeped to a ward, you’re asked to do it, you don't know how to, so you bleep someone to ask them [but] they’re stressed out and busy as well, so they try to tell you over the phone, they’ve got no knowledge of the patient, you’ve barely got any knowledge of the patient…’ Interviewee 6 (medical school J)
Speculation over the rates of errors made by doctors when prescribing drugs and the subsequent impact on patient safety has led to much research and media attention in recent years, yet little firm evidence has been available about the role of basic medical education. To explore how education specifically can help reduce the risk of prescription errors, the GMC commissioned a project to investigate the cause and prevalence of errors made by FY1 doctors (first year foundation trainees) although the project also looked at doctors at other stages of their careers.
The project, completed this November, provides reassurance of the approach taken in the new 2009 edition of Tomorrow’s Doctors; its findings and recommendations are vital reading for all healthcare professionals.
Scale of the problem
To gain a representative picture of the prevalence and nature of prescribing errors, the team visited 19 hospital trusts.
In total, 124,260 medication orders were checked from all prescribers. 11,077 contained errors; an error rate of 8.9%. Of the total orders checked, 50,016 were written by FY1 doctors with an error rate of 8.4%.
Fifty three per cent of errors were deemed potentially significant and 40% were described as minor. Potentially serious errors were less common at five per cent and potentially lethal errors were found in fewer than two per cent. Significantly, prescribing errors were 70% more likely at the point of admission to hospital. For definitions and details of the methodology please refer to the full report.
What it revealed
The project found that perhaps the single most important influencing factor was the complexity of the system within which prescriptions are made, and that prescription errors are not solely, or even primarily, a problem of the most junior trainees. All levels of prescriber were found to make some errors with the highest error rate being found in FY2 doctors, but even consultants made a significant number of errors.
However, project manager and researcher, Dr Penny Lewis, explains that as FY1 trainees were the primary focus of interest of the study the report pertains primarily to them, but that they likely typify the wider culture of clinical care.
To look at some of the causes of the errors, the project team conducted a separate qualitative study with FY1 doctors. ‘Rule-based mistakes’, applying the wrong rule or failing to apply the right one, accounted for nearly half the errors. Such mistakes were found to result from lack of expertise in defining a problem and applying the correct solution to it. Professor Tim Dornan, project lead, explains: ‘Mistakes of this sort were not primarily due to a lack of declarative knowledge, but to the difficulty of applying knowledge learned during the undergraduate years to practice.’ He adds that such rule-based mistakes could often have been prevented by using readily available sources of help such as written information, senior colleagues, or members of other health professions.
Dr Lewis comments that a lack of knowledge, unsurprisingly, contributed to knowledge-based mistakes: ‘but it was very contextualised knowledge rather than broad principles that was the primary cause of such mistakes’. Lack of a day-to-day, working knowledge of individual patients was an important predisposing factor. Also of particular concern was when FY1 trainees’ adequate knowledge was overridden by the system, notably more senior doctors’ instructions. Being busy, having a high caseload, having to rush, feeling tired, having difficulty concentrating, multi-tasking, and feeling flustered all made it hard for respondents to apply knowledge they already had. Dr Lewis adds: ‘When knowledge was lacking, it could have been remedied (and on some occasions was) by providing better support and/or trainees being readier to take advantage of it’.
Certain environments, such as units with a fast turnover of patients, led to ‘skill-based’ errors, as did short-staffing on wards and tiredness. Unfamiliar drug charts from hospital to hospital were also cited as a major concern and inspired one of the report’s recommendations: to standardise prescription forms.
There were also times when the authority, expectation and behaviour of more senior members of communities of practice led FY1 trainees to change correct decisions to incorrect ones and times where FY1 trainees were put into situations that led to errors for which they were deemed responsible.
Safety nets
The report stresses though that almost all errors were intercepted before reaching patients and causing any harm, which reflected the existence of well-developed safety nets. While nurses and senior doctors played important parts in preventing errors impacting on patients, the contribution of pharmacists was found to be pivotal.
Recommendations
The research has identified five main targets for interventions to improve patient safety by minimising prescribing errors:
• clinical working environments
• undergraduate medical education programmes
• FY1 education
• other parts of the medical education continuum
• interprofessional education.
From trainees' perspectives, under-graduate programmes were said to be variable in quality. Instances of good, practical education included a blend of good science with good clinical teaching, appropriate assessments, stringent teaching and interprofessional education. Less satisfactory experiences were: when prescribing education was left to opportunistic learning or was not practical in nature; when theory dominated over practice; when a programme did not incrementally prepare students for practice; and when shadowing placements were of poor quality. Some specific suggestions for improvement included:
• more training in practical prescribing, including filling in drug charts
• teaching about common prescribing errors
• clinically focused teaching, including common scenarios and interactive cases
• training about treating patients on multiple drugs.
Suggestions included: narrowing the gap between theoretical instruction in pharmacology and practical prescribing; and making practical prescribing a mandatory component of final examinations and included in portfolio assessment. Professor Dornan summarises: ‘The challenge that emerged from respondents' narratives was for medical schools to provide a balanced curriculum, which helped medical students learn basic pharmacology, and link it to practical prescribing’.
Another key finding for Professor Dornan is the need for a national prescription chart (as has been in place in Wales since 2004), resolving unnecessary variation between hospitals. He also highlights the recommendation that senior doctors should not devolve prescribing responsibility to junior, inexperienced doctors without adequate supervision.
Conclusion
Many of the recommendations support the developments in the revised version of Tomorrow’s Doctors published in September 2009 but will also be of great interest to other organisations.
The report offers a number of recommendations but stresses that, if education is to reduce errors, it must also include higher specialist training and continuing professional development for consultants as well as undergraduate and foundation training.
The report concludes: ‘Since it is likely practitioners in the UK healthcare system will have heavy workloads for the foreseeable future, an intervention that standardises prescription charts, makes support systems more readily available in the heat of busy practice, inculcates a safety culture, and encourages trainees to challenge instructions from more senior doctors that they disagree with will likely be needed to reduce the prevalence of errors.’
The full report will be available on the GMC website shortly.