A Preliminary Characterization of Canonicalized and Non-Canonicalized Section Headers Across Variable Clinical Note Types.

In the electronic health record, the majority of clinically relevant information is stored within clinical notes. Most clinical notes follow a set organizational structure composed of canonicalized section headers that facilitate clinical review and information gathering. Standardized section header terminologies such as the SecTag terminology permit the identification and standardization of headers to a canonicalized form. Although the SecTag terminology has been evaluated extensively for history & physical notes, the coverage of canonical section header terms has not been assessed across other note types. For this pilot study, we conducted a coverage study and characterization of canonical section headers across 5 common, clinical note types and a generalizability study of canonical section headers detected within two types of clinical notes from Penn Medicine.