RESUMO
This article puts forward a new proposal to calculate count, turnaround, response, and service time of work orders in a clinical engineering (CE) department. These are calculated by means of a queuing model as a measurement tool. This proposal was tested in a 600-bed hospital with an inventory of 1094 medical devices and with 6 full-time clinical engineers. In April 1999, a simulation (with ARENA 3.01 developed by System Modeling Corporation) of the working of this proposal was performed with desired values being applied to the queuing model. At the end of 2002, real work order data from the database was recorded. As predicted, the results showed that all the indicators of nonscheduled work orders decreased. Response and turnaround time were reduced from 27 to 0.56 hours and 27.48 to 1.13 hours, respectively. From a backlog of 22 outstanding repair orders per month between April 1999 and January 2000, the number was reduced to 4 in December 2002. The queuing model also helped to measure the positive effects on arrival and service rates when users and CE were trained. The difference between simulated and real values was under 5%.