From the Editor

Rocket scientists need not apply

At the time of writing this, The Times newspaper front page lead concerns the necessity to top up the NHS Litigation Authorities legal compensation fund to the tune of £185 million, with the astounding statistic that clinical negligence payments by the NHS have risen by 500% over 10 years.¹ The article goes on to discuss the causes of this escalating cost, concentrating on changes in compensation practice (notably to allow life-long support of brain-damaged babies when negligence is an established cause) and legal practices such as the ‘no-win, no-fee’ basis for many cases. It also discusses imminent changes such as the restriction of legal aid aimed at curbing this financial drain on the NHS. Also discussed on many front pages there is also the issue of who will or won't pay, for what, in which part of the UK, for faulty breast implant material. Medical and manufacturer negligence is big news and big bucks.

Its true cost is of course incalculable, because there is no system for quantifying distress, no effective system for documenting every manifest occurrence and in particular no system for discovering negligence which is not manifest. Physicians may even feel that – evidenced by the sliding scale of medical defence organisation subscriptions – they are only minor contributors to the problem. An article by Barton et al in this issue of Clinical Medicine is a quiet reminder that for physicianly errors there is a huge invisible portion of the iceberg.² In a very simple observational study, they show that the chances of having an incorrectly written prescription chart in a survey of adult medical in-patients in the 24 hours after admission in four Australasian hospitals was well over 90% – not a misprint.

They used a stringent description of prescribing errors. Many doctors might think that with those definitions some of these errors were ‘unimportant’, but no-one could condone undocumented allergies, illegible drug names and unsigned prescriptions. They did not document complex errors (such as dose, route of administration or interactions). Others, closer to home, documented prescribing errors which reduced the probability of timely and effective treatment, or increased the risk of harm, in three UK hospitals, teaching and non-teaching, and reported a 15% incidence of errors.³ Among a random sample of these erroneous prescriptions 19% were reported to be serious. Others have reported that 1–2% of inpatients are harmed by medication errors, predominantly those reflecting poor prescribing.^4,5

Reducing, or even eliminating, prescribing errors requires more than a single strategy. This journal and others recently highlighted the invaluable role that pharmacist review has to play,⁶ while acknowledging that with current staffing structures this is always likely to be subject to delay. Disappointingly, the Health Foundation's Safer Patients Initiative – which aims to induce an institutional culture of safety and in respect of reducing medication errors approached the issue by involving staff in failure mode analysis, identifying weak links, and providing tools to measure adverse events – failed to impact on the quality of prescribing.⁷ Approaches such as the use of electronic prescribing can certainly eliminate errors of legibility and clarity, and have been successfully introduced in critical care wards, for example; though intriguingly in at least one comparative report a computerised physician order system was associated with a number of major errors whereas the handwritten comparison group was not.^8,9 It is clear, however, that well-designed electronic prescribing systems can improve compliance with recognised prescribing standards.

Intriguingly – returning to Barton et al – the pattern of some simple prescribing errors varied between the four different Australasian hospitals and this could clearly be traced in part to the different designs of the inpatient prescription chart. Some designs are better than others and consensus on an optimum approach should not be impossible. As junior trainee doctors move between institutions in the UK, the absurdity of the NHS relying on a variety of different chart designs must be reinforced to them every six or twelve months. With different charts in different hospitals, not only will some forms be more fit-for-purpose than others, but the unfamiliarity of the different charts to doctors moving between institutions will itself impose a risk. Over a decade, workers in Brisbane introduced a revised medication chart, audited its before-and-after use in five hospital sites in Queensland (demonstrating a significant fall in prescribing errors) and last year published their analysis of the implementation of their National Inpatient Medication Chart in 22 hospitals nationwide; the prescribing error rate fell by a third.^10,11 While acknowledging a Hawthorn effect – whereby the process of education and implementation may have played a role in reducing error rates in this before-and-after study – the authors comment that the standard format facilitates the education of prescribers, and it does seem intuitively obvious that a uniform prescription chart must diminish the risks involved as physicians set about their most dangerous task: prescribing to their patients. Come to think of it, a single prescription chart for the NHS would also be strikingly cheaper than, for example, a national IT system. Is it so difficult? If not, why is it taking so long?