A88 Eur J Hosp Pharm 2013;20(Suppl 1):A1-A238 robot. Errors were classified in three categories: wrong drug, missing drug/quantity or additional quantity. 2. Workload efficiency: time to prepare a sequence of orders manually or with the robot was measured. Results 1. Manual dispensing error rate was 0.93% (n = 5805 ordered lines; wrong drug: 0.36%, missing drug/quantity: 0.31%, additional quantity: 0.26%). By decreasing this error rate to 0.27% (n = 5840; only conveyor errors leading to missing

more »

... drug/quantity and additional quantity), the automation avoided more than 4500 errors each year. 2. With the distribution of 880 boxes of drugs/hour (reduced to 630 when the automated 'Pro-log' filling system was working simultaneously), the robot significantly increased the distribution speed in comparison with the manual picking (303 boxes/hour). Conclusions This reorganisation contributed to safer and more efficient distribution of drugs. No more incorrect picking of medicines occurred thanks to the high reliability of the robot. Remaining errors could still be reduced by improving the conveyor software. With one single person operating the robot, 2 full-time equivalents were saved, leading to an estimated return on investment in 4.5 years. For medicines remaining outside the robot (i.e. controlled drugs, cold chain drugs or those with an unusual size, shape or weight), a scanning system will be introduced and evaluated by the same protocol. No conflict of interest.