The dataset can be downloaded by clicking the link below for data in Excel (XLSX) format:
Randomly selected de-identified example sentences from clinical notes for confirmed associations
Note: All sample sentences have been de-identified using several methods.
1. All numbers were changed to 1.
2. All patient names were replaced with garbled text.
3. MIST de-identification software was trained and run on the clinical notes to further garble identifiers (and some of the numbers were randomly re-assigned to numbers other than 1).
4. De-identification of clinical notes was validated and certified by Harvard Vanguard Medical Associates.
5. The randomly selected sentences below were hand-checked for protected health information, where any identifiers or pseudo-identifiers were replaced with "***" and unnecessary information was removed (indicated by "…").