What was already known on the topic	What this study added to our knowledge
Most clinical information in EHRs is stored as unstructured text, and this includes most eligibility criteria. NLP can be used successfully to extract eligibility criteria from trial descriptions (e.g., ClinicalTrials.gov). NLP can be used to extract information from unstructured clinical text with good accuracy.	Besides demographics, a large majority of eligibility criteria is only mentioned in unstructured clinical text (95.8% in our case). NLP can also be used to extract eligibility criteria from EHR unstructured clinical text with high precision and recall.