Natural Language Processing (NLP) techniques have shown promise for extracting data from the free text of electronic health records (EHRs), but studies have consistently found that techniques do not readily generalize across application settings. Unfortunately, most of the focus in applying NLP to real use cases has remained on a paradigm of single, well-defined application settings, so that generalizability to unseen use cases remains implicitly unaddressed. We propose to explicitly account for unseen application settings by adopting an information retrieval (IR) perspective with the objective of patient-level cohort identification. To do so, we introduce layered language models, an IR framework that enables the reuse of NLP-produced artifacts. Our long term goal is to accelerate investigations of patient health and disease by providing robust, user- centric tools that are necessary to process, retrieve, and utilize the free text of EHRs. The main goal of this proposal is to accurately retrieve ad hoc, realistic cohorts from clinical text at Mayo Clinic and OHSU, establishing methods, resources, and evaluation for patient-level IR. We hypothesize that cohort identification can be addressed in a generalizable fashion by a new IR framework: layered language models. We will test this hypothesis through four specific aims.
In Aim 1, we will make medical NLP artifacts searchable in our layered language IR framework. This involves storing and indexing the NLP artifacts, as well as using statistical language models to retrieve documents based on text and its associated NLP artifacts.
In Aim 2, we deal with the practical setting of ad hoc cohort identification, moving to patient-level (rather than document-level) IR. To accurately handle patient cohorts in which qualifying evidence may be spread over multiple documents, we will develop and implement patient-level retrieval models that account for cross- document relational and temporal combinations of events.
In Aim 3, we will construct parallel IR test collections using EHR data from two sites;a diverse set of cohort queries written by multiple people toward various clinical or epidemiological ends;and assessments of which patients are relevant to which queries at both sites. Finally, in Aim 4, we refine and evaluate patient-level layered language IR on the ad hoc cohort identification task, making comparisons across the users, queries, optimization metrics, and institutions. We will draw additional extrinsic comparisons with pre-existing techniques, e.g., for cohorts from the Electronic Medical Records and Genonmics network. The expected outcomes of the proposed work are: (i) An open-source cohort identification tool, usable by clinicians and epidemiologists, that makes principled use of NLP artifacts for unseen queries;ii) A parallel test collection for cohort identification, includig two intra-institutional document collections, diverse test topics and user-produced text queries, and patient-level judgments of relevance to each query;and (iii) Validation of the reusability of medical NLP via the task of retrieving patient cohorts.

Public Health Relevance

With the widespread adoption of electronic medical records, one might expect that it would be simple for a medical expert to find things like patients in the community who suffer from asthma. Unfortunately, on top of lab tests, medications, and demographic information, there are observations that a physician writes down as text - which are difficult for a computer to understand. Therefore, we aim to process text so that a computer can understand enough of it, and then search that text along with the rest of a patient's medical record;this will allow clinicians or researchers to find and study patients groups of interest.

National Institute of Health (NIH)
National Library of Medicine (NLM)
Research Project (R01)
Project #
Application #
Study Section
Biomedical Computing and Health Informatics Study Section (BCHI)
Program Officer
Vanbiervliet, Alan
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Mayo Clinic, Rochester
United States
Zip Code
Wang, Liwei; Ruan, Xiaoyang; Yang, Ping et al. (2016) Comparison of Three Information Sources for Smoking Information in Electronic Health Records. Cancer Inform 15:237-242
Wang, Yanshan; Wu, Stephen; Li, Dingcheng et al. (2016) A Part-Of-Speech term weighting scheme for biomedical information retrieval. J Biomed Inform 63:379-389
Liu, Sijia; Liu, Hongfang; Chaudhary, Vipin et al. (2016) An Infinite Mixture Model for Coreference Resolution in Clinical Notes. AMIA Jt Summits Transl Sci Proc 2016:428-37
Kaggal, Vinod C; Elayavilli, Ravikumar Komandur; Mehrabi, Saeed et al. (2016) Toward a Learning Health-care System - Knowledge Delivery at the Point of Care Empowered by Big Data and NLP. Biomed Inform Insights 8:13-22
Rastegar-Mojarad, Majid; Komandur Elayavilli, Ravikumar; Liu, Hongfang (2016) BELTracker: evidence sentence retrieval for BEL statements. Database (Oxford) 2016:
Rastegar-Mojarad, Majid; Liu, Hongfang; Nambisan, Priya (2016) Using Social Media Data to Identify Potential Candidates for Drug Repurposing: A Feasibility Study. JMIR Res Protoc 5:e121
Li, Yanpeng; Liu, Hongfang (2015) Learning Semantic Tags from Big Data for Clinical Text Representation. AMIA Jt Summits Transl Sci Proc 2015:461-5
Mehrabi, Saeed; Krishnan, Anand; Roch, Alexandra M et al. (2015) Identification of Patients with Family History of Pancreatic Cancer--Investigation of an NLP System Portability. Stud Health Technol Inform 216:604-8
Li, Dingcheng; Rastegar Mojarad, Majid; Li, Yanpeng et al. (2015) A Frequency-based Strategy of Obtaining Sentences from Clinical Data Repository for Crowdsourcing. Stud Health Technol Inform 216:1033-4