PathoGuide: A Hybrid AI Clinical Decision Support System for Evidence-Based Antimicrobial Prescribing in Resource-Constrained Settings

Antimicrobial resistance (AMR) poses a critical global health threat, disproportionately affecting resource-constrained healthcare systems. In Zimbabwe, clinicians frequently initiate empiric antibiotic therapy without access to real-time, localized resistance data, while laboratory and pharmacy workflows remain fragmented and largely paper-based. This study presents PathoGuide, a prototype clinical decision-support system designed to support evidence-informed antibiotic prescribing by linking clinical, laboratory, and pharmacy workflows in healthcare facilities in Mutare, Zimbabwe. Using a Design Science Research (DSR) approach, the study assessed workflow and antimicrobial-stewardship needs through two questionnaire groups comprising 71 healthcare professionals, developed a rule-based recommendation engine (the Pathogen-Empiric Treatment Engine, PETE), and conducted an initial acceptance assessment with a smaller subgroup of eight participants. Key findings revealed that 83.3% of clinicians lacked regular access to local resistance data, 100% rated inter-unit communication as Fair or Poor, and 63.5% received fewer than one antimicrobial stewardship training session annually despite 92.1% recognizing the need for more education. All participants rated PathoGuide as useful and 100% expressed positive usage intention. The system demonstrated strong potential to bridge critical prescribing infrastructure gaps in low- and middle-income country (LMIC) healthcare settings through integrated real-time decision support, data standardization, and embedded education. Keywords—Antimicrobial resistance; Antimicrobial stewardship; Clinical decision support systems; Evidence-based medicine; Hybrid systems; PathoGuide; Zimbabwe.