The RPDR is happy to announce that we have a new way to search through the text of 130 million clinical notes in the RPDR.
This new functionality will allow researchers to query clinical notes for specific text terms and phrases in order to identify patient cohorts who have notes/reports that contain the searched text. Currently, the clinical notes available to search include LMR Notes, Cardiology, Discharge Summaries, Operative Reports, Pathology, Pulmonary, Endoscopy and Radiology Reports. Researchers will also have the option to refine a note search by using date constraints or choosing to eliminate reports with negated terms; this option will attempt to exclude notes with statements like “the patient does NOT have disease X”.
As with other RPDR item types, note search items can be combined with additional criteria (i.e. diagnoses, procedures, etc.) to construct queries for patient cohorts of interest.
In order to maintain patient privacy, there are several limitations. First, rare non-medical names cannot be included in a search string. Second, the RPDR automatically eliminates any data found in the personal information files of the patient from the search. Third, numbers cannot be used as search terms since it is difficult to distinguish them from medical record numbers and accession numbers found in reports. Finally, phrases longer than 7 words cannot be searched.
We have this available thanks to the Partners Big Data Commons. For existing patients, the data in the Commons is updated daily. Therefore, note searches will reflect the daily updates.
For further information, please see the Notes Search ontology and functionality in the RPDR Query Tool or email us at @email.