Evidence-based medicine and artificial intelligence in pediatric healthcare

Alberto Tozzi is responsible for the unit of research in preventive and predictive medicine at Bambino Gesù hospital in Rome.

Artificial intelligence is not an alternative method opposed to evidence-based medicine.

AI and evidence-based medicine should be integrated together.

AI algorithms must be evaluated in clinical cases through rigorous study frameworks.

AI can work with various data types including databases, audio, and video.

Artificial intelligence has fundamental statistical foundations.

AI algorithms can identify associations and correlations but cannot establish causation.

Privacy is a critical consideration when developing AI algorithms with patient data.

AI algorithms using photographs can identify genetic syndromes, with performance data from North American studies.

Algorithm performance must be monitored over time in clinical settings.

Clinical trials are necessary to validate AI algorithms before clinical deployment.

Clinical trials for AI algorithms may be conducted at single clinical centers.

Observational studies and clinical observation are fundamental for preliminary evaluation of AI algorithms.