Clinicians often use the ‘notes’ section of the electronic health record (EHR) to document patient information. But the use of free-form text can leave data unstandardized – and important details may get lost or missed when providing patient care.

To solve this problem, healthcare organizations often turn to natural language processing (NLP) solutions built to scan and interpret this text.

But medical language is unique. NLP used in a healthcare setting should be trained using a robust, foundational clinical terminology in order to account for a variety of situations common in clinical notes. In this insight brief, we explore four examples of how clinical language is distinctive, including:

• Acronyms, abbreviations, and misspellings
• Use of negation
• Specialty-dependent language
• Time-based relationships