Diagnosa Penyakit Preeklamsia Menggunakan Metode Dempster Shafer

Studi Kasus : RSU Bidadari

Abstract

Preeclampsia is one of the most serious complications in pregnancy, characterized by hypertension and proteinuria, and it poses a significant risk of maternal and fetal morbidity and mortality if not detected and managed promptly. Early detection is crucial, yet clinical diagnosis often faces challenges due to the variability of symptoms and uncertainty in medical decision-making. To address this issue, this study aims to develop an expert system for diagnosing preeclampsia by employing the Dempster-Shafer method, which is known for its ability to handle uncertainty and incomplete information in complex domains such as healthcare. A case study was conducted at Bidadari General Hospital, where data on clinical symptoms and patient medical records were collected and analyzed. The development process of the expert system followed systematic stages, including knowledge acquisition from obstetrics specialists, designing the knowledge base, constructing inference rules, and integrating the Dempster-Shafer algorithm for decision support. The system was subsequently tested using real-case scenarios of pregnant women suspected of having preeclampsia. Evaluation results demonstrated that the system achieved an accuracy rate of 92% in differentiating between preeclampsia and eclampsia, based on belief and plausibility measures combined with symptom analysis. These findings indicate that the proposed system can effectively support medical personnel by providing diagnostic recommendations with a high degree of reliability. In addition, the system offers efficiency in the clinical workflow by minimizing diagnostic errors and reducing delays in treatment initiation. Therefore, this expert system has the potential to become a valuable clinical decision support tool for early detection, risk assessment, and management of preeclampsia. Future development may focus on expanding the knowledge base, integrating real-time patient monitoring data, and enhancing usability to ensure broader applicability in diverse healthcare settings.