Referans aralıklarının hasta kayıtlarından elde edilmesi sistemi

In the reference interval studies of biochemical laboratory tests, samples are provided by at least 120 healthy volunteers for the subgroups of the population. The results that are gained by examining the samples are used to identifying the lower and upper bounds covering 95 percent of the population by statistical methods. This conventional method is based on certain rules and recommendations and the application of this method is onerous and costly. On the other hand, reference intervals that are defined for the adult patient group are used for the pediatric patient group. In this study, an integrated software system is developed and presented, which includes artificial intelligence-based learning algorithms that are empowered by data mining methods, allowing each health institution to perform its own pediatric reference interval studies with its own available data. By using this system, experts can examine the characteristics of the data, perform experiments with learning algorithms, and display results efficiently. In the tests that were conducted during the development phase of the system, the data that is published by Canadian Laboratory Initiative on Pediatric Reference Intervals (CALIPER) and reference intervals that are published by CALIPER is used. The reference intervals that were generated by the developed system were compared with the reference intervals published by CALIPER, the results were examined. According to the results, it was observed that the reference intervals that were generated by the developed system are compatible with the reference intervals that are published by CALIPER. In addition, high-resolution reference intervals were generated for some biochemical tests.