Klinik Laboratuvar Süreçlerinde İstatistiksel Kalite Kontrol

Abstract

Statistical quality control methods are commonly used in production environments to determine whether a process is achieving targeted outputs and to monitor the stability of a process. Similar methods for different purposes are also applied in clinical laboratories, which give medical decision support with vital consequences.

In clinical laboratories, statistical quality control is implemented with two mechanisms: internal quality control and external quality control. Internal quality control is also handled in two main categories: analytical assessment and patient-based process control. Although the analytical assessment using quality control material is widely used in almost all clinical laboratories, patient-based methodologies are not commonly accepted in practical applications. In this study, alternative quality control methods and the underlying reasons behind the negligence towards patient-based statistical quality control are investigated.

It is seen that, at the current state, patient-based quality control studies focus on the average of normals (AON) methodology. An investigation of alternative methodologies shows that the cumulative sum (CUSUM) methodology has not been utilized to track patient data in the literature. In order to test whether this methodology will bring any performance improvement, a new method is proposed by combining AON and CUSUM methods to minimize the number of patients in between the occurrence time of an error and its detection. Comparing the performance of AON method with the proposed method, it is shown that the proposed method is performing significantly better in detecting especially small process shifts.

Moreover, it is observed that the main reason behind patient-based statistical quality control mechanisms for not being used commonly is the lack of a computer program for implementations and tests. In order to provide such a platform, a simulation based computer program is developed for experimentations with the parameters of the methods and to evaluate the performance outcomes of quality control processes.