Research article Introduction Intravenous (i.v.) therapy usually needs to be pre- pared immediately before administration. This may involve dissolving of powder, dilution or transfer of injection fluid from the original vial or ampoule into a container (a syringe or an infusion bag). These processes present multiple opportunities for errors. Thirty years ago Breckenridge investigated prepara- tion and administration of i.v. medication on hospital wards in the United Kingdom (UK) 1 . In his report it was summarised that there was a lack of information and guidelines, as well as inadequate prescribing, which resulted into poor quality of care. Following An observational study of intravenous medication errors in the United Kingdom and in Germany • Veronika Wirtz, Katja Taxis and Nick D. Barber Pharm World Sci 2003; 25(3): 104–111. © 2003 Kluwer Academic Publishers. Printed in the Netherlands. V. Wirtz, N.D. Barber: Department of Practice and Policy, The School of Pharmacy, University of London, UK (e-mail: veronika.wirtz@ams1.ulsop.ac.uk) K. Taxis: Pharmazeutisches Institut, Pharmazeutische Biologie, Universität Tübingen, Germany Key words Administration of i.v. drugs Aseptic technique Germany Intravenous administration Intravenous medication error Medication errors Observational study Severity United Kingdom Abstract Objectives: To investigate the incidence and the severity of intravenous (i.v.) drug preparation and administration errors in two countries and three pharmacy services. Method: A disguised observational method was used to record details of the preparation and administration of prescribed i.v. drugs on two wards in each of three teaching hospitals: one with a traditional British ward pharmacy service (TBP) and two hospitals in Germany, one with a traditional ward stock supply (TGP) and one with a satellite pharmacy service (GSP) with unit dose system. Main outcome measures: Errors in i.v. drug preparation and administration and their potential significance. Results: The number of observed preparations/administrations were: TBP 77/63, TGP 126/109 and GSP 134/106. The preparation error rates were: TBP 22% (95% confidence interval: 13-31%), TGP 23% (16-30%) and GSP 31% (23- 39%). The administration error rates were TBP 27% (16- 38%), TGP 49% (39-58%) and GSP 22% (14-30%). The percentage of administration errors on the wards with TGP was statistically significantly higher than in the other two services. Common errors at the study sites with TBP and GSP were omissions. Wrong rate of administration occurred most frequently on the wards with TGP. The majority of errors were likely to be of ‘moderate’ to ‘severe’ outcome. Careful drug chart reading could possibly reduce omission errors on the wards with TBP. A change of the German nursing law (‘Krankenpflegegesetz’) to legally entitle nurses to administer i.v. drugs could probably result in better training, national guidelines and standards. Conclusion: This study found a high rate of i.v. medication errors of moderate to severe significance. Changes in practice should be considered to make i.v. therapy safer for patients. Accepted February 2003 this report few studies have been performed on the use of i.v. drugs and related medication errors (ME). Three studies investigated only i.v. medication errors: one of them found 151 (84.4%) errors in 179 observed drug administrations 2 , the other reported an error rate of 24.7% for 320 observed preparations and administrations 3 . A recent study found an error rate of 49% in 430 i.v. drug preparations and admin- istrations 4 . These findings are higher than the error rates, between 3% and 8%, found for oral ME 5,6 . Some studies investigated preparation and adminis- tration of oral and i.v. medication in intensive care units (ICU). A study in a paediatric ICU in Switzerland found 18% errors in 231 observed events 7 . Another study in France reported an error rate of 6% observ- ing 2009 events 8 . This error rate was calculated by dividing the number of errors by the number of observed events. Each of the administrations consist- ed of several events. A study done in United States (US) in the early seventies identified a total of 21% errors in 100 preparations observed 9 . Observations regarding aseptic technique were also included in two studies 9,10 showing that the majority of nurses did not follow it. However, error rates of the studies are difficult to compare due to differences in settings, methods and definitions used. Only a few studies gathered data in different phar- macy services and different countries, using the same methods and definitions to allow a comparison of error rates across different systems. However, none of these studies included i.v. ME. Dean et al. 5 found an error rate of 3.0% for the traditional British ward pharmacy system and an error rate of 6.9% for the unit dose system of a US hospital. Characteristic types of errors were identified in each system: omission errors occurred frequently in the traditional ward pharmacy system, whereas unordered or incorrect doses were often administered in the unit dose sys- tem. Another study compared the British ward phar- macy system with the traditional ward stock system and the unit dose system in two German teaching hospitals 6 . In British hospitals, ward pharmacists initi- ate drug supply and review prescriptions on their daily ward visits 11 . In contrast, such patient-oriented pharmaceutical services are rarely provided in German hospitals. The wards have a large stock of commonly prescribed medicines. Few hospitals oper- ate a unit dose system or satellite pharmacy services 12 . The study found a higher error rate (8.0%) for the British pharmacy system than for the tradition- al German ward stock system (5.1%) or the German unit dose system (2.4%) 6 . Again characteristic types of errors were identified in each system. Therefore, the aim of the present study was to investigate the incidence and the severity of i.v. ME during preparation and administration in two coun- tries and three pharmacy services. 104