Pilot of ClinicalKey AI in a Large Health Care System

Introduction 

ClinicalKey AI searches within a closed loop of resources such as journal abstracts, government publications, and particular Elsevier books and journals, and always provides a linked reference list. Our institution piloted ClinicalKey AI from July 2024 through September 2025. Survey results from January – June 2025 show different use cases of the tool, overall impressions of the tool, and comparison to other tools. 

Methods 

We reached out to a variety of departments to gauge interest and provide access. A survey was sent out monthly to all users. The survey included questions about previous AI use, usage of ClinicalKey AI, and several open-ended questions about the impact of their ClinicalKey AI use.  

Results 

We added 950 users by June 2025, and 39 responded to our survey. 31 indicated previous AI use. 33 used ClinicalKey AI for patient care, 28 for identifying resources, 18 for studying and/or exam prep, and 8 for evaluating a product or medical device. 32 respondents gave the tool a 4 or 5 out of 5 for overall impressions, and 31 wanted continued access after the pilot period. Comparisons to other tools were mixed, with some users preferring ClinicalKey AI to OpenEvidence and UptoDate, and others indicating that they would continue to use OpenEvidence or UptoDate instead. 

Discussion 

The majority of our users were among residents and physician faculty. The tool was well received with many users mentioning the speed and ease of use. Although our institution has decided to subscribe, continuing evaluation will be needed.