<![CDATA[Preeclampsia is a pregnancy complication that may increase the risk of maternal and fetal mortality if it is not detected and managed early. At the primary healthcare level, screening still relies on healthcare workers’ assessment of examination results. Therefore, a data-based approach is needed to support the identification of preeclampsia risk. This study aims to develop a website-based prototype that applies the K-Nearest Neighbor (KNN) algorithm to support the early detection of preeclampsia in pregnant women. The research method used was Research and Development (R&D) with a Prototype model, while data processing followed the CRISP-DM stages. The data were obtained from Puskesmas and consisted of age, parity, height, weight, body mass index, systolic blood pressure, diastolic blood pressure, Mean Arterial Pressure (MAP), random blood glucose, and urine protein. The modeling results showed that K=3 achieved an accuracy of 99.81%, precision of 100.00%, recall of 98.97%, and F1-score of 99.48%. System testing obtained 95% in Black Box Testing, 100% in White Box Testing, and 82.8% in user testing using the Post-Study System Usability Questionnaire (PSSUQ). The developed prototype is expected to help healthcare workers conduct more objective screening and support the diagnostic process at Puskesmas.]]>
Copyrights © 2026
