Errors during the preparation of drug infusions: a randomized controlled trial.

BACKGROUND We investigated the extent and frequency of dose errors and treatment delays made as a consequence of preparing drug infusions at the bedside, rather than using pre-filled syringes. METHODS Forty-eight nurses with critical care experience volunteered to take part in this randomized, blinded, controlled study conducted in the simulation centre of an urban hospital. They assisted in the management of a simulated patient with septic shock. Vasopressor infusions were prepared either by diluting concentrated drugs from ampoules or were provided in syringes pre-filled beforehand by an intensive care unit resident. RESULTS The time taken for the infusion to be started and the final concentration of the drugs were measured. We also measured the concentration of infusions prepared by a pharmacist and a pharmaceutical company. Nurses took 156 s to start infusions when using pre-filled syringes compared with 276 s when preparing them de novo, a mean delay of 106 s [95% confidence interval (CI) 73-140 s, P<0.0001]. One infusion prepared from ampoules contained one-fifth of the expected concentration of epinephrine; another contained none at all. Medication errors were 17.0 times less likely when pre-filled syringes were used (95% CI 5.2-55.5), and infusions prepared by pharmacy and industry were significantly more likely to contain the expected concentration (P<0.001 for norepinephrine and P=0.001 for epinephrine). CONCLUSIONS Providing drug infusions in syringes pre-filled by pharmacists or pharmaceutical companies would reduce medication errors and treatment delays, and improve patient safety. However, this approach would have substantial financial implications for healthcare providers, especially in less developed countries.