Online Medical Device Use Prediction: Assessment of Accuracy

Cost-intensive units in the hospital such as the operating room require effective resource management to improve surgical workflow and patient care. To maximize efficiency, online management systems should accurately forecast the use of technical resources (medical instruments and devices). We compare several surgical activities like using the coagulator based on spectral analysis and application of a linear time variant system to obtain future technical resource usage. In our study we examine the influence of the duration of usage and total usage rate of the technical equipment to the prediction performance in several time intervals. A cross validation was conducted with sixty-two neck dissections to evaluate the prediction performance. The performance of a use-state-forecast does not change whether duration is considered or not, but decreases with lower total usage rates of the observed instruments. A minimum number of surgical workflow recordings (here: 62) and >5 minute time intervals for use-state forecast are required for applying our described method to surgical practice. The work presented here might support the reduction of resource conflicts when resources are shared among different operating rooms.