Any document that is uploaded into Pattern is analyzed using four methods in the following order:
1. Optical Character Recognition (OCR)
OCR is used to extract text records using a combination of OCR processes and tools.
2. Medical Language Processing
After text has been extracted, Pattern searches for any and all medical concepts and terms within that text. Pattern's medical linguistics engine supports misspellings, alternative names, and most shorthand.
3. Date Assignment
In order to get the most accurate information possible, Pattern scans all the extracted text and attempts to determine what date is most relevant for each piece of medical information. For example, a clinical note might have a "Visit Date" noted at the top, but then denote a history of a procedure in a prior year. Our date assignment is what determines that the historical procedure should not be aligned with the "Visit Date".
4. Lens Analysis
The unfiltered volume of medical information Pattern extracts would be overwhelming, and, like medical records, much of it would be irrelevant. The lens criteria distills the medical history of a case to an easily digestible curated dataset. The results of the lens criteria are available on the Case Criteria and Review pages.