The workload of the outpatient departments in Chinese large hospitals is extremely high. Patients often have to wait for a long time before getting their treatments. It is economically expensive to increase medical staffs including nurses and doctors. Therefore, it is critical to optimize staff planning in the outpatient departments to reduce excessive patient waiting time. A high-fidelity simulation model can accurately capture the features of the outpatient service system. But it is very time-consuming to obtain the optimal staff planning decision only based on the simulation model. A simplified queueing model might lead to an analytical solution for the optimal staff planning problem, but it can not fully capture the feature of the real outpatient service system. We propose to use the outputs of the high-fidelity simulation model to drive the output of the low-fidelity queueing model closer to that of the outpatient service system, and then use the data-driven queueing model to make the staff planning decision. Empirical studies on a major hospital are carried out, which demonstrate the effectiveness and efficiency of our method.