Artificial Intelligence for Rapid Meta-Analysis: Case Study on Ocular Toxicity of Hydroxychloroquine

Area of Science:

• Medical Informatics
• Evidence-Based Medicine
• Artificial Intelligence in Healthcare

Background:

• Clinical decision-making requires rapid access to evidence, especially during evolving health crises.
• Traditional meta-analysis is time-consuming, posing challenges in urgent situations.
• A novel approach, rapid meta-analysis (RMA), is proposed to balance speed and data quality.

Purpose of the Study:

• To evaluate the efficacy of RMA in generating timely clinical insights.
• To compare the processing time of RMA against traditional meta-analysis.
• To address the clinical question of ocular toxicity associated with hydroxychloroquine therapy.

Main Methods:

• Developed a rapid meta-analysis (RMA) approach integrating artificial intelligence (AI) and human analysis.
• Utilized AI to rapidly screen and extract data from relevant studies on hydroxychloroquine and ocular toxicity.
• Applied standard statistical methods for data analysis.

Main Results:

• RMA successfully generated a clinical result in under 30 minutes.
• Identified 11 studies on hydroxychloroquine's ocular toxicity, estimating an incidence of 3.4%.
• Acknowledged high heterogeneity across studies, necessitating careful interpretation.

Conclusions:

• Demonstrated that RMA, utilizing AI, can produce clinically relevant insights significantly faster than traditional meta-analysis.
• Highlights the potential of AI-powered RMA for efficient evidence synthesis in critical healthcare scenarios.