Objective:To develop an artificial intelligence(AI)algorithms tool for data collection in chronic hepatitis B,addressing the issue of low efficiency in traditional multicenter data collection.Methods:Based on internationally recognized data standards,this study applies mature AI techniques in the field of computer science,such as optical character recognition and natural language processing,to the data collection of real-world cohorts in chronic hepatitis B research.This tool enables the electronic collection of various data formats,including digitization of raw data in image format,structured processing,and automatic population of data into electronic case report forms(eCRFs)designed in the Research Electronic Data Capture(REDCap)system.Results:Experimental results demonstrate that the AI algorithmstool for data collection achieves the same average accuracy as manual data collection(P =0.23),with an accuracy rate of 98.66%,while reducing the time required by 75.49%(P<0.05)compared to manual collection.Conclusion:The AI algorithms tool for data collection developed in this study significantly improves the efficiency of research data collection and brings new developments to the real-world research data collection paradigm.
data collectionhepatitis Bartificial intelligence(AI)natural language processingoptical character recognition
NETL
NSTL
万方数据
