Reported error rates are likely to be underestimationBMJ 2009; 338 doi: https://doi.org/10.1136/bmj.b1814 (Published 06 May 2009) Cite this as: BMJ 2009;338:b1814
- Bryony Dean Franklin, director1,
- Katja Taxis, assistant professor of pharmacotherapy and clinical pharmacy2,
- Nick Barber, professor of the practice of pharmacy3
- 1Centre for Medication and Service Quality, Imperial College Healthcare NHS Trust/School of Pharmacy, University of London, London
- 2Department of Pharmacy, Section of Pharmacotherapy and Pharmaceutical Care, University of Groningen, 9713BV Groningen, Netherlands
- 3School of Pharmacy, University of London, London WC1N 1AX
The reported error rates in Valentin and colleagues’ study on errors in administration of parenteral drugs in intensive care units are likely to underestimate the true incidence of errors.1
Firstly, errors were identified using self reporting by hospital staff. Only about 1 in 100 prescribing errors and 1 in 1000 administration errors are reported using established incident reporting systems.2 3 The error rate was 7% of all doses given, which is substantially lower than the error rates of 18-173% found in other studies in intensive care and other wards using the “gold standard” of independent observers.4 5
Secondly, the authors assessed only five specific types of error. Other potentially common administration errors associated with parenteral treatment, including use of the wrong diluent, incompatibility errors, and wrong administration rates, seem to have been excluded. Conducting regression analyses on the basis of a potentially small subset of reported errors is therefore likely to be flawed.
Valentin and colleagues gave no information on the role of pharmacists, although pharmacists are routinely part of the critical care team in many of the countries included. Pharmacists reduce adverse events due to prescribing errors in critical care, and having a pharmacist in a multidisciplinary team is likely also to be associated with reduced administration errors.
We advocate the use of independent observers collecting data on actual practice to understand the true incidence and causes of administration errors in intensive care,5 as well as controlled studies to identify the true impact of interventions.
Cite this as: BMJ 2009;338:b1814
Competing interests: None declared.