Objective:

To develop a clinical decision support system using artificial intelligence to accurately distinguish between multisystem inflammatory syndrome in children (MIS-C) and endemic typhus, thereby improving diagnostic accuracy and treatment outcomes.

Key Findings:

• AI-MET achieved 100% accuracy in classifying typhus and MIS-C, demonstrating its potential as a reliable diagnostic tool.
• A validation cohort of 111 additional MIS-C patients was classified with 99% accuracy, reinforcing the model's robustness.
• 24 out of 30 features differed statistically, but individual parameters were clinically irrelevant, highlighting the need for careful interpretation of AI outputs.

Interpretation:

Artificial intelligence can effectively differentiate between MIS-C and endemic typhus using readily available clinical features, aiding front-line providers in diagnosing febrile children in endemic areas and potentially improving patient outcomes.

Limitations:

• The study relied on a specific patient population from Texas Children's Hospital, which may limit generalizability to other regions.
• The overlap of clinical features between MIS-C and typhus may complicate diagnosis despite AI assistance, necessitating further validation in diverse populations.

Conclusion:

AI-MET represents a valuable tool for clinicians in diagnosing febrile children, potentially improving treatment outcomes for both MIS-C and typhus, and underscores the role of AI in enhancing clinical decision-making.