Automated Recognition and Post-Coordination of Complex Clinical Terms

One of the key tasks in integrating guideline-based decision support systems with the electronic patient record is the mapping of clinical terms contained in both guidelines and patient notes to a common, controlled terminology. However, a vocabulary of pre-coordinated terms cannot cover every possible variation - clinical terms are often highly compositional and complex. We present a rule-based approach for automated recognition and post-coordination of clinical terms using minimal, morpheme-based thesauri, neoclassical combining forms and part-of-speech analysis. The process integrates MetaMap with the open-source GATE framework.