Most inclusion/exclusion criteria are available only in unstructured ways (paper-based, unstructured PDFs, ...). So transfering these informations into structured data is time-consuming and not 100% possible (depending on the data sources connected to the HL7 interface).
So adding NLP features to the trial registry would ease the way, how trial personell can register trials. For example paper- or PDF-based study protocols could be scanned via OCR, analyzed via NLP and the trials dataset could be pre-filled with the mined criteria.
Most inclusion/exclusion criteria are available only in unstructured ways (paper-based, unstructured PDFs, ...). So transfering these informations into structured data is time-consuming and not 100% possible (depending on the data sources connected to the HL7 interface).
So adding NLP features to the trial registry would ease the way, how trial personell can register trials. For example paper- or PDF-based study protocols could be scanned via OCR, analyzed via NLP and the trials dataset could be pre-filled with the mined criteria.