Factors causing dose variability in drug administration.

The variability of the drug dose actually given to cancer patients was analyzed. Three variability factors were quantitatively examined (body surface calculation, personalized dose calculation, and drug residuum in commercially available vials) and their variability was experimentally measured. A systematic reduction (mean, 7%; range, 2%-15%) and a random variability (4%-5%) of the dose given were demonstrated. These results draw attention to the role of some of the procedures of routine clinical activity in determining the amount of drug actually delivered. The analysis suggests that personalization of doses must be very accurate in both measurement and calculation and that the staff giving the drug needs to be carefully informed about the importance of drug residuum. The variability of the delivered dose can lead to the misclassification of patients in investigations on the dose-response relationship. This factor may be added to pitfalls previously reported to affect this type of retrospective analysis.