Investigating bias in the natural language processing of electronic health records
When clinicians describe patients in medical records, their language may reflect bias against certain groups of patients or conditions. This project will study the language of electronic health records to see if such biases can be detected and measured.