Towards standardisation of drug

‘First, do no harm’ (Primum non nocere) is an important maxim of medical practice1 and yet unintended harm occurs within various health systems, including critical care, with alarming regularity.2 The National Patient Safety Agency (NPSA) was set up and tasked with monitoring and reducing adverse event rates in the UK.3 In total, 72,481 medication-related incidents were reported to the NPSA in 2007 from acute Trusts, although under-reporting means that the true figure is likely to be higher.4 The Institute of Medicine, which is part of the National Academy of Sciences and advises the US government, published the report ‘To Err is Human’ in 2000, which concluded that 98,000 deaths occur annually due to preventable mistakes. This is more than deaths from breast cancer, road traffic accidents or AIDS and yet medical error does not have the same level of public attention.5 There are also likely to be cases where errors have occurred but have remained undetected. Medications that require manipulation before administration are prone to these types of errors and studies have demonstrated that a wide variation in administered dose can occur as a result. One such study analysed samples taken from acetylcysteine infusions prepared in the emergency department for the treatment of paracetamol poisoning. It found that only 61% of doses that had been administered were within 20% of the prescribed dose.6 Alarmingly, one dose contained no drug, and a further 4% of samples deviated from the expected concentration by 50% or more. The preparation of controlled drugs in a paediatric critical care population showed that 6% of prepared infusions exhibited a two-fold concentration error or greater.7 This work has recently been extended, with 118 volunteers from a variety of health professional backgrounds who undertook a series of calculations and prepared morphine syringes which were later analysed. Even here under scrutiny, 8% of 464 morphine infusions prepared were found to have two-fold concentration errors.8 In a recent study, discarded syringes containing midazolam, insulin, noradrenaline, dopamine, potassium or magnesium were collected from a UK critical care unit and were analysed. Only midazolam and insulin were within the industry standard for variance in concentration (less than 10%). The mean discrepancies from the expected concentrations in the remaining medications were 144% for magnesium, 20% for potassium, and -12% for dopamine. Box and whisker plots show wide variation in concentrations, which are particularly marked for noradrenaline and potassium. Magnesium and potassium were singled out for further study. The introduction of specific electrolyte charts did reduce the variation in concentration but did not eliminate it.9 The US Institute for Safe Medication Practices has long advocated strict functions and constraints to produce a system that should make errors virtually impossible. This is the most powerful tool for preventing errors.10 Standardisation of practice facilitates this and has been advocated by the UK’s Department of Health and the Audit Commission.11,12 We have previously sought to establish the feasibility of standardising infusion concentrations by ascertaining the extent of variation in practice for a set of medications commonly used in critical care.13 The aim of this follow-up survey is to assess the acceptability to UK critical care units of a draft list of standard medication concentrations and to identify possible factors that would impede the introduction of such a standard.