Activity-based costing (ABC) is widely used to precisely allocate indirect costs to medical services. In the ABC method, the indirect cost is divided among the medical services in proportion to the volume of "cost drivers", for example, labor hours and the number of hours of surgery. However, the workload of data collection of cost drivers can be time-consuming and a considerable burden if there are many cost drivers. The authors aim to develop a method for objectively reducing the cost drivers used in the ABC method. In the ABC method, the cost driver is assigned for each activity. We assume that these activities and cost drivers are the best combination. Our method, that is cost driver reduction (CDR), can objectively select surrogates of the cost drivers for each activity in ABC from candidate cost drivers. Concretely, the total indirect cost of an activity is temporarily allocated to the medical services using each candidate of cost drivers. The difference between the costs calculated by each candidate and the proper cost driver used in ABC is calculated to evaluate the similarity by the evaluation function. We estimated the cost of laboratory tests using our method and revealed that the number of cost drivers could be reduced from seven in the ABC to four. Similarly, the results of cost estimation obtained by our method were as accurate as those calculated using the ABC. Our method provides two advantages compared to the ABC method: 1) it provides results that are as accurate as those of the ABC method, and 2) it is simpler to perform complicated estimation of hospital costs.