An outbreak investigation was conducted to determine if an increase in bloodstream infections (BSIs) in patients with central venous catheters (CVC) had occurred. Because other methods of obtaining CVC days were not feasible, we used an estimation method based on a random 5% sample of medical records to determine the proportion of days that a CVC was present for each of three patient units. This calculated ratio was used to estimate the total CVC days for each unit. A cohort study was conducted in which the BSI rates before and during needleless device use were compared. This article describes the methods used to calculate this estimated denominator and discusses the need for such a denominator to be used by infection control practitioners when prospective collection of CVC days is not possible.