Coloured drug labels have long been used to improve the recognition of syringe labels in an attempt to minimize drug administration error. However, their impact on reducing drug administration error has not been proved . Medication errors can increase costs and have devastating consequences for patients . Multiple incompatible label systems have been in use in the UK for several years  and this has been highlighted as a risk factor for drug administration error (Webster CS, Merry AF, personal communication, 2000). The Councils of the Royal College of Anaesthetists, the Association of Anaesthetists of Great Britain and Ireland, the Faculty of Accident and Emergency Medicine and the Intensive Care Society agreed to implement a single standard for syringe labelling. The International Colour Coding System (developed by the International Organization of Standardization (ISO)) for syringe labelling replaces previous individual local labelling systems. The need to move to a colour coding standard has generally been acknowledged  although there have also been concerns raised regarding the potential for an increase in drug administration errors during the period of transition to the new system (Souter A, personal communication, 2003). Drug administration error during anaesthesia, of which ‘syringe swap’ (the selection of the wrong syringe and the erroneous administration of its contents) is the most frequent , is known to occur in 0.11-0.75% of cases [1,6]. The incidence of latent errors, i.e. errors which are detected prior to drug administration, has been reported to be 0.4% . Prior to changing to the new system, our department was using the Medilabel® system for syringe labelling. We decided to study the effect of the use of the ISO labelling system on drug error rates prior to its introduction and attempt to identify key areas of increased risk.
Following local research Ethics Committee approval, 31 anaesthetists were invited to induce general anaesthesia for a simulated patient in a designated emergency theatre. Each participant was made aware that they were taking part in a research study and in no aspect were being personally assessed. None declined to take part. A high fidelity Laerdal SimMan (Laerdal Medical Ltd, Orpington, UK) was used in all cases. The presented scenario was designed to suggest the need for a rapid sequence induction. The following information was supplied on a typewritten sheet:
‘You have been asked to anaesthetize the following patient for a laparotomy. The surgeons suspect that he has a bleeding duodenal ulcer. The patient is already in the anaesthetic room and the senior house officer (SHO) has drawn up and labelled any drugs that you might need. The SHO has had to go elsewhere in the hospital, leaving you to anaesthetize the patient. Aside from the current acute abdominal problem, he is otherwise fit and well. He is a non-smoker, drinks occasional alcohol and has had general anaesthesia for minor orthopaedic surgery in the past without problems. He takes regular lansoprazole for indigestion/ reflux symptoms but no other regular medication. He has no known allergies to any medication. He last ate 3 h ago.’
On entering the anaesthetic room, each subject was presented with a simulated 40-yr-old adult patient of approximately 70 kg weight. A trained anaesthetic assistant was present who confirmed that all anaesthetic equipment has been fully checked. Pulse oximetry, electro cardiogram, invasive blood pressure, capnography and inspired oxygen analysis were displayed on a standard monitor screen. A 16-G intravenous (i.v.) access had already been established in the right antecubital fossa. Cardiovascular monitoring and the observed respiratory rate suggested adequate fluid resuscitation had been achieved. The drugs shown in Table 1 had been drawn up in appropriate syringes in advance and labelled with the ISO system of colour-coded labels. The syringes were presented in a single tray in a random arrangement.
Each anaesthetist was asked to induce general anaesthesia in their preferred manner using the drugs available. No information was given regarding the aims of the study.
Following induction of anaesthesia, monitored patient parameters were adjusted in a standardized fashion. An unexpected bradycardia was introduced once the patient's airway had been secured. The heart rate was maintained at 35 bpm despite any intervention for a period of 2 min before returning to the baseline level of 80 bpm. No explanation of the cause was suggested. At this point the scenario was halted.
In order to reduce inter-observer error and distraction while running the simulator programme a single dedicated observer recorded all drug administrations throughout the study. Active error was defined as the injection of the wrong drug. If the anaesthetist selected a syringe by mistake and made purposeful movement towards the i.v. cannula but stopped short of administering the drug, this was also recorded but defined as a latent error. Rearrangement of the drug syringes were permitted. Strategies that potentially reduced the risk of drug administration error were noted. Participants' previous experience with the ISO labelling system was recorded following completion of the scenario. None of the participants reported being colour blind.
Categorical data were analysed using SPSS Release 6.14, SPSS Inc, Chicago, USA.
Of the 31 anaesthetists that took part in the study, 13 were consultants, 14 were specialist registrars (SpRs) and 4 were SHOs of at least 12 months anaesthetic experience. In the group of anaesthetists with no prior ISO labelling system experience (Group 1) a total of 107 drug injections were recorded of which 1 (0.9%) was an active error and 16 (15%) involved latent errors; 11 anaesthetists (58%) performed at least one latent error.
The one active error that occurred involved the administration of fentanyl instead of suxamethonium during a rapid sequence induction. As a result no muscle relaxant was given prior to intubation. The total observed latent errors are detailed in Table 2.
The effects of anaesthetic experience on drug administration error rates in Group 1 are shown in Table 3.
Two error reduction strategies were recorded during the duration of the study: verbalizing drug selection as the drug was administered and the practice of separating essential or required syringes from others prior to commencing induction of anaesthesia.
The effects of verbalizing drug selection on drug administration error in Group 1 are shown in Table 4. Sixteen anaesthetists verbalized each drug injection. One active error occurred and was verbalized as the incorrect drug. The drug administrations associated with latent errors were all correctly verbalized. Three anaesthetists did not verbalize each drug injection.
Twelve anaesthetists in Group 1 spontaneously chose to separate the drug syringes in some manner prior to commencing induction of anaesthesia, whilst seven anaesthetists left all the drug syringes in a single tray throughout the anaesthetic. The effects of this strategy are shown in Table 5.
The second Group of 12 anaesthetists had at some time in their career been exposed to the ISO labelling system (Group 2). Eight of them were consultants and four were SpRs. Within this group a total of 65 injections were made. No active errors occurred but two latent errors were recorded (3% error rate). This represented a 6.9 (95% CI 1.3-26.7) fold reduction in the error rate (P = 0.023) compared with the anaesthetists with no previous exposure to the system.
It should be noted that for 1 month prior to the study, the critical care directorate had embarked on an intensive orientation exercise to increase the awareness of all anaesthetic staff to the potential risks of drug errors occurring during the transition period. Each anaesthetist had been sent a colour A4 sheet displaying the new labels along side the pre-existing system. A copy of this information sheet was also posted in every anaesthetic room (including that used for the study). The assumed resultant level of increased vigilance did not prevent errors from occurring during this study.
Although only one drug was actually given in active error, the latent error rate of 15% involving 58% of the anaesthetists studied was dramatically higher than would be expected in normal practice  even though self reporting probably underestimates the recognition of latent errors. The seniority of the anaesthetist and their prior exposure to the labelling system appeared to reduce the rate of error, although only prior exposure was found to be statistically significant. The practice of separating out the syringes, creating a reduced choice of selection, surprisingly appeared to confer no protection against error. It appears that colour/pattern recognition was the dominant factor in these cases. The effect of verbalizing the drug administered is difficult to interpret given the small sample in the non-verbalizing group although interestingly the error rate appeared to be higher in the verbalizing group.
The temporal distribution of the latent errors was evenly spread through the simulated case. Six occurred during the rapid sequence induction, six during the stabilization phase and six during the simulated critical incident of an unexplained persistent bradycardia.
Certain patterns of error did emerge from the study. Fourteen of the 18 latent errors involved confusion of a muscle relaxant with another drug. Four occurred when fentanyl was picked up instead of suxamethonium. Three occurred when morphine was chosen instead of vecuronium. This is of particular interest given that the new system allocates the previous Medilabel® colour of muscle relaxants, pale blue, to opioids. This colour change issue has already lead to several critical incidents in the South-west region (Souter A, personal communication, 2003). During the critical incident, the reverse situation seemed to occur with four anaesthetists picking up a ‘red’ syringe (muscle relaxant) when they desired either ephedrine or atropine. The previous Medilabel® system used a red colour to signify drugs that had cardiovascular effects. The colour of the label for atropine remains green.
Critical incidents involving ‘syringe swap’ have previously been shown to be most likely to occur in otherwise healthy ASA I and II patients who otherwise do not require complex anaesthetic intervention .
Apart from raised vigilance, strategies have been suggested to reduce error during the changeover period. One would normally expect the anaesthetist administering the drugs to draw up and label his/her own syringes, rather than use those prepared by another anaesthetist as required in this simulation scenario. One department changed to uncoloured labels for a period to ‘unlearn’ the colour associations prior to the new label introduction (Frerk C, Webster R, personal communication, 2003). This has had its critics in departments where this strategy has been introduced (Fouque CA, personal communication, 2003).
Pre-packed syringes for muscle relaxants would provide additional visual cues for correctly identifying and administering drugs, and would eliminate errors in the drawing up and correct labelling of a drug syringe. Purchasing pre-filled syringes however would add an additional burden of cost to anaesthetic department drug budgets. Changing the appearance of the syringe containing muscle relaxant, e.g. by having a coloured barrel, has been advocated although the strength of the supporting evidence is weak . Since the completion of this study the Councils of the Royal College of Anaesthetists, the Association of Anaesthetists of Great Britain and Ireland, the Faculty of Accident and Emergency Medicine and the Intensive Care Society have issued further clarification and guidance on the standardization of syringe labelling .
Although this study does not provide any clear solutions to the problem, it does associate the period of transition between labelling systems with increased anaesthetic risk for patients, particularly with regard to the erroneous administration of muscle relaxants.
1. Fasting S, Gisvold SE. Adverse drug errors in anaesthesia, and the impact of coloured syringe labels. Can J Anaesth
2. Jensen LS, Merry AF, Webster CS, Weller J, Larsson L. Evidence-based strategies for preventing drug administration errors during anaesthesia. Anaesthesia
3. Radhakrishna S. Syringe labels in anaesthetic induction rooms. Anaesthesia
4. Christie IW, Hill MR. Standardized colour coding for syringe drug labels: a national survey. Anaesthesia
5. Currie M, Mackay P, Morgan C et al
. The ‘wrong drug’ problem in anaesthesia: an analysis of 2000 incident reports. Anaesth Intens Care
6. Webster CS, Merry AF, Larsson L, McGrath KA, Weller J. The frequency and nature of drug administration error during anaesthesia. Anaesth Intens Care
7. Irita K, Tsuzaki K, SawaT et al
. Critical incidents due to drug administration error in the operating room: an analysis of 4 291 925 anaesthetics over a 4 year period. Masui
1Presented in part at the Anaesthetic Research Society, Aberdeen, 2nd April 2004.