Portal de Periódicos da CAPES

Integrating Data Mining and Optimization Techniques on Surgery Scheduling

2012; Springer Science+Business Media; Linguagem: Inglês

10.1007/978-3-642-35527-1_49

1611-3349

Carlos F. Gomes, Bernardo Almada‐Lobo, José Borges, Carlos Soares,

Hemodynamic Monitoring and Therapy

This paper presents a combination of optimization and data mining techniques to address the surgery scheduling problem. In this approach, we first develop a model to predict the duration of the surgeries using a data mining algorithm. The prediction model outcomes are then used by a mathematical optimization model to schedule surgeries in an optimal way. In this paper, we present the results of using three different data mining algorithms to predict the duration of surgeries and compare them with the estimates made by surgeons. The results obtained by the data mining models show an improvement in estimation accuracy of 36%. We also compare the schedules generated by the optimization model based on the estimates made by the prediction models against reality. Our approach enables an increase in the number of surgeries performed in the operating theater, thus allowing a reduction on the average waiting time for surgery and a reduction in the overtime and undertime per surgery performed. These results indicate that the proposed approach can help the hospital improve significantly the efficiency of resource usage and increase the service levels.