Abstract

Introduction: We analyze the assembly of surgical trays in a hospital's sterile services department. The department assembles 520 different tray setups. However, tray assembly times are unknown, imposing a challenge to production planners. To respond to demand, workers from other departments are often called, leading to higher operational costs and more frequent quality problems due to workers' poor training and inconsistency. Methods: Conducting traditional time-motion studies is infeasible in such a high variety production setting. Thus, we used design of experiments to optimize the data acquisition. Assembly times of 36 trays were sampled using a 2-factor nested factorial design. Through regression analysis, we built a model to estimate completion times of trays not sampled in the experiment. Results: A prediction model with 90.8% accuracy was obtained from the experimental data. The model was validated with assembly times from several trays not included in the experiment. Predicted assembly times had an absolute error of 7.83% on average compared with observed assembly times. Conclusions: Design of experiments and regression analysis combined were able to optimize time data acquisition using a small sample of trays, resulting in a model that predicted assembly times within an acceptable margin of error.