AI promises to revolutionize healthcare by enhancing diagnosis, treatment, and patient safety. Yet, a critical gap exists between AI research and its real-world impact. We urgently need comprehensive clinical effectiveness evaluations to bridge this divide.
My colleagues at UC San Diego Health, UC Davis Health, and I published this article in the NEJM AI to establish boundaries for using AI in the healthcare setting, bringing together research and real-world evidence for comprehensive clinical effectiveness. AI is bound to revolutionize our industry that was traditionally adverse to change, yet it is essential to collaborate on the shared goal of safer, more effective, and equitable care for all patients. Christopher Longhurst Karandeep Singh Aneesh Chopra Ashish Atreja, MD, MPH
Our call to action:
(1) Move Beyond Model Validation: It's essential to validate AI models not just through simulations but in real-world clinical settings, akin to clinical trials for new drugs.
(2) Focus on Local Context: The importance of local health care contexts in AI model validation cannot be overstated. Different settings require tailored approaches for optimal outcomes.
(3) Implement Science Principles: Adopting implementation science principles will ensure that AI tools are effectively integrated into health care workflows, resulting in tangible improvements in patient outcomes.
A robust network of healthcare delivery organizations is crucial to focus on the clinical effectiveness of AI models in real-world settings. This collaborative approach will help us achieve safer, more effective, and equitable care for all patients.
#HealthcareInnovation #AIinHealthcare #ClinicalEffectiveness #PatientSafety #ImplementationScience #AIValidation #EquitableCare #HealthTech