We present a protocol and associated metadata template for the extraction of text describing biomedical concepts in clinical case reports. The structured text values produced through this protocol can support deep analysis of thousands of clinical narratives.