Abstract

Clinical NLP is the application of text processing approaches on documents written by healthcare professionals in clinical settings, such as notes and reports in health records. The interest for clinical NLP is spurred by the need for real-time, largescale, and accurate information extraction from health records to support clinical care, e.g., through automated generation of a patient problem list, to support biomedical and health services research, e.g., through precise cohort identification, and to support public health practice, e.g., through disease surveillance. Clinical NLP can provide clinicians with critical patient case details, which are often locked within unstructured clinical texts and dispersed throughout a patient’s health record. Recently, patient-generated text such as posts on social media forums related to health have also received increased interest for population-based studies.

In this talk, I will describe work on automated extraction of clinically relevant information from clinical text, including semantic aspects such as negation, uncertainty and time information. I will give some examples from my experience in working with Swedish and English health record data, collaborating with clinicians and considerations needed when working with this type of data, as well as my recent experiences with working in the mental health domain, including studies on social media data.