A Review on Artificial Intelligence and Point-of-Care Diagnostics to Combat Antimicrobial Resistance in Resource-Limited Healthcare Settings like Nigeria

Abstract

The global crisis of antimicrobial resistance (AMR) demands innovative diagnostic solutions, particularly in resource-limited settings. This paper examines the integration of artificial intelligence (AI) with point-of-care (POC) diagnostics for AMR detection in Nigerian healthcare systems. A systematic search of literature published between 2018 and 2024 was conducted across major databases including PubMed, Scopus, and Web of Science, yielding 127 relevant studies. Current evidence indicates that AI-enabled POC platforms demonstrate 92-97% accuracy in pathogen identification and can reduce diagnostic turnaround time from 48-72 hours to 2-4 hours. Machine learning algorithms, particularly deep neural networks and random forests, have shown promising results in predicting resistance patterns with 89% sensitivity and 94% specificity. Implementation challenges in Nigeria include limited infrastructure, with only 23% of healthcare facilities having adequate diagnostic capabilities, and a significant workforce shortage, with a ratio of 1 laboratory scientist to 20,000 patients. Economic analyses suggest that AI-POC integration could reduce diagnostic costs by 60% and decrease inappropriate antibiotic prescriptions by 40%. Literature indicates that AI-augmented POC diagnostics represent a viable solution for enhancing AMR surveillance and antimicrobial stewardship in Nigeria.