Time Expressions in Mental Health Records for Symptom Onset Extraction

Natalia Viani,Rina Dutta,Ayunni Alawi,Rashmi Patel,André Bittar,Robert Stewart,Lucia Yin,Sumithra Velupillai,Joyce Kam

doi:10.18653/v1/w18-5621

Natalia Viani, Rina Dutta + Show 7 more

Open Access

https://doi.org/10.18653/v1/w18-5621

Copy DOI

Abstract

For psychiatric disorders such as schizophrenia, longer durations of untreated psychosis are associated with worse intervention outcomes. Data included in electronic health records (EHRs) can be useful for retrospective clinical studies, but much of this is stored as unstructured text which cannot be directly used in computation. Natural Language Processing (NLP) methods can be used to extract this data, in order to identify symptoms and treatments from mental health records, and temporally anchor the first emergence of these. We are developing an EHR corpus annotated with time expressions, clinical entities and their relations, to be used for NLP development. In this study, we focus on the first step, identifying time expressions in EHRs for patients with schizophrenia. We developed a gold standard corpus, compared this corpus to other related corpora in terms of content and time expression prevalence, and adapted two NLP systems for extracting time expressions. To the best of our knowledge, this is the first resource annotated for temporal entities in the mental health domain.

Highlights

Introduction and BackgroundFor psychiatric disorders such as schizophrenia, prolonged periods of time without treatment are associated with worse intervention outcomes (Kisely et al, 2006)
We focus on one subgoal: addressing the problem of accurately identifying time expressions in mental health records related to patients who have been diagnosed with schizophrenia
We queried the Clinical Record Interactive Search (CRIS) database for patients who had been documented with an ICD-10 code for this disease (F20*) or, if not documented with a structured code, we relied on the output of an Natural Language Processing (NLP) tool which extracts diagnoses from free text (Perera et al, 2016), resulting in 8,483 documents for 1,691 patients3

Summary

Introduction

For psychiatric disorders such as schizophrenia, prolonged periods of time without treatment are associated with worse intervention outcomes (Kisely et al, 2006). A longer DUP has been linked to poorer cognitive function at the time of first presentation (Lappin et al, 2007). Identifying and reducing the DUP could significantly improve both clinical and functional outcomes. Starting from this observation, there is an increasing interest in measuring the DUP across large clinical samples, to provide a quality measure for mental health services, and in developing international guidelines aimed at reducing this value, improving outcomes at different levels (Connor et al, 2013)

Objectives

Methods

Results

Discussion

Conclusion