<![CDATA[The increasing number of patients seeking medical consultation at primary healthcare facilities often leads to prolonged waiting times and delayed preliminary diagnosis, particularly for common diseases with overlapping symptoms. At Puskesmas Karanganyar, diagnostic procedures remain largely dependent on direct consultation, limiting service efficiency under high patient volumes. This study aims to design and implement a web-based expert system to support early diagnosis by simulating clinical reasoning. The system employs a rule-based inference mechanism using Forward Chaining to identify potential diseases from patient-selected symptoms, while the Certainty Factor method is integrated to quantify diagnostic confidence. Unlike prior rule-based diagnostic systems, this study contributes an integrated multi-disease diagnostic framework tailored to primary healthcare workflows, combining transparent rule traceability with graded confidence representation to enhance interpretability and practical usability. Knowledge acquisition was conducted through expert interviews and literature review, resulting in a knowledge base of ten diseases and forty symptoms. The system was implemented using PHP and MySQL and is accessible across devices. Empirical evaluation through structured functional testing and user-oriented validation indicates that the system achieves consistent diagnostic outputs with 100% functional success across tested scenarios, average response time below 2 seconds, and positive usability feedback from healthcare staff, demonstrating operational reliability in real-world settings. The findings suggest that the proposed system provides fast, consistent, and informative preliminary diagnoses, supporting more efficient decision-making in primary healthcare services.]]>
