Evaluation of five systems to determine technician staffing in a unit dose cart-filling area

The development and evaluation of predictive systems to determine staffing needs in a centralized unit dose cart-filling area were studied. Data concerning actual cart-filling time and the hospital's daily census, by total beds and by bed type, were collected over 55 days. Four predictive systems were then developed, as follows: simple average, range average, simple regression, and multiple regression. In addition to these mathematical systems, a pharmacist "best-guess" system was devised, whereby the pharmacist directing the cart-filling area estimated the staffing needs on a daily basis during the trial period. The five systems were then used to predict cart-filling time daily over 14 days. During this time, the actual filling time was recorded and compared with the times predicted by the five systems. The differences among the actual or predicted mean cart-filling times for the five systems were not significant. The pharmacist best-guess system was on average the most accurate in detecting different staffing needs; the advantage of this system is that the pharmacist can evaluate differences in work habits among the scheduled technicians, which the mathematical models would be unable to do. The simple average system correlated well with changes in filling time and most precisely predicted variability in census. Although none of the systems was superior in all respects, a combination of the pharmacist best-guess and simple-average systems appeared to be the best method for predicting daily technician staffing needs in the central cart-filling area.