Natural language processing of clinical notes enables early inborn error of immunity risk ascertainment
Background: There are now approximately 450 discrete inborn errors of immunity (IEI) described; however, diagnostic rates remain suboptimal. Use of structured health record data has proven useful for patient detection but may be augmented by natural language processing (NLP). Here we present a machi...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Elsevier
2024-05-01
|
Series: | Journal of Allergy and Clinical Immunology: Global |
Subjects: | |
Online Access: | http://www.sciencedirect.com/science/article/pii/S2772829324000201 |
_version_ | 1797289646064926720 |
---|---|
author | Kirk Roberts, PhD Aaron T. Chin, MD Klaus Loewy, MS Lisa Pompeii, PhD Harold Shin, MS Nicholas L. Rider, DO |
author_facet | Kirk Roberts, PhD Aaron T. Chin, MD Klaus Loewy, MS Lisa Pompeii, PhD Harold Shin, MS Nicholas L. Rider, DO |
author_sort | Kirk Roberts, PhD |
collection | DOAJ |
description | Background: There are now approximately 450 discrete inborn errors of immunity (IEI) described; however, diagnostic rates remain suboptimal. Use of structured health record data has proven useful for patient detection but may be augmented by natural language processing (NLP). Here we present a machine learning model that can distinguish patients from controls significantly in advance of ultimate diagnosis date. Objective: We sought to create an NLP machine learning algorithm that could identify IEI patients early during the disease course and shorten the diagnostic odyssey. Methods: Our approach involved extracting a large corpus of IEI patient clinical-note text from a major referral center’s electronic health record (EHR) system and a matched control corpus for comparison. We built text classifiers with simple machine learning methods and trained them on progressively longer time epochs before date of diagnosis. Results: The top performing NLP algorithm effectively distinguished cases from controls robustly 36 months before ultimate clinical diagnosis (area under precision recall curve > 0.95). Corpus analysis demonstrated that statistically enriched, IEI-relevant terms were evident 24+ months before diagnosis, validating that clinical notes can provide a signal for early prediction of IEI. Conclusion: Mining EHR notes with NLP holds promise for improving early IEI patient detection. |
first_indexed | 2024-03-07T19:08:08Z |
format | Article |
id | doaj.art-29e51555a77543379d8757b0f61c0338 |
institution | Directory Open Access Journal |
issn | 2772-8293 |
language | English |
last_indexed | 2024-03-07T19:08:08Z |
publishDate | 2024-05-01 |
publisher | Elsevier |
record_format | Article |
series | Journal of Allergy and Clinical Immunology: Global |
spelling | doaj.art-29e51555a77543379d8757b0f61c03382024-03-01T05:07:45ZengElsevierJournal of Allergy and Clinical Immunology: Global2772-82932024-05-0132100224Natural language processing of clinical notes enables early inborn error of immunity risk ascertainmentKirk Roberts, PhD0Aaron T. Chin, MD1Klaus Loewy, MS2Lisa Pompeii, PhD3Harold Shin, MS4Nicholas L. Rider, DO5McWilliams School of Biomedical Informatics, The University of Texas Health Science Center at Houston, Houston, TexDivision of Immunology, Allergy, and Rheumatology, University of California, Los Angeles, CalifTexas Children’s Hospital, Houston, TexDepartment of Patient Services, Cincinnati Children’s Hospital Medical Center, Cincinnati, OhioCollege of Osteopathic Medicine, Liberty University, Lynchburg, VaDivision of Health System & Implementation Science, Virginia Tech Carilion School of Medicine, Roanoke, Va; Section of Allergy and Immunology, Carilion Clinic, Roanoke, Va; Corresponding author: Nicholas L. Rider, DO, Virginia Tech Carilion School of Medicine, 1 Riverside Circle, 249 Roanoke, VA 24016.Background: There are now approximately 450 discrete inborn errors of immunity (IEI) described; however, diagnostic rates remain suboptimal. Use of structured health record data has proven useful for patient detection but may be augmented by natural language processing (NLP). Here we present a machine learning model that can distinguish patients from controls significantly in advance of ultimate diagnosis date. Objective: We sought to create an NLP machine learning algorithm that could identify IEI patients early during the disease course and shorten the diagnostic odyssey. Methods: Our approach involved extracting a large corpus of IEI patient clinical-note text from a major referral center’s electronic health record (EHR) system and a matched control corpus for comparison. We built text classifiers with simple machine learning methods and trained them on progressively longer time epochs before date of diagnosis. Results: The top performing NLP algorithm effectively distinguished cases from controls robustly 36 months before ultimate clinical diagnosis (area under precision recall curve > 0.95). Corpus analysis demonstrated that statistically enriched, IEI-relevant terms were evident 24+ months before diagnosis, validating that clinical notes can provide a signal for early prediction of IEI. Conclusion: Mining EHR notes with NLP holds promise for improving early IEI patient detection.http://www.sciencedirect.com/science/article/pii/S2772829324000201Natural language processingmachine learningtext mininginborn errors of immunityprimary immunodeficiencydiagnosis |
spellingShingle | Kirk Roberts, PhD Aaron T. Chin, MD Klaus Loewy, MS Lisa Pompeii, PhD Harold Shin, MS Nicholas L. Rider, DO Natural language processing of clinical notes enables early inborn error of immunity risk ascertainment Journal of Allergy and Clinical Immunology: Global Natural language processing machine learning text mining inborn errors of immunity primary immunodeficiency diagnosis |
title | Natural language processing of clinical notes enables early inborn error of immunity risk ascertainment |
title_full | Natural language processing of clinical notes enables early inborn error of immunity risk ascertainment |
title_fullStr | Natural language processing of clinical notes enables early inborn error of immunity risk ascertainment |
title_full_unstemmed | Natural language processing of clinical notes enables early inborn error of immunity risk ascertainment |
title_short | Natural language processing of clinical notes enables early inborn error of immunity risk ascertainment |
title_sort | natural language processing of clinical notes enables early inborn error of immunity risk ascertainment |
topic | Natural language processing machine learning text mining inborn errors of immunity primary immunodeficiency diagnosis |
url | http://www.sciencedirect.com/science/article/pii/S2772829324000201 |
work_keys_str_mv | AT kirkrobertsphd naturallanguageprocessingofclinicalnotesenablesearlyinbornerrorofimmunityriskascertainment AT aarontchinmd naturallanguageprocessingofclinicalnotesenablesearlyinbornerrorofimmunityriskascertainment AT klausloewyms naturallanguageprocessingofclinicalnotesenablesearlyinbornerrorofimmunityriskascertainment AT lisapompeiiphd naturallanguageprocessingofclinicalnotesenablesearlyinbornerrorofimmunityriskascertainment AT haroldshinms naturallanguageprocessingofclinicalnotesenablesearlyinbornerrorofimmunityriskascertainment AT nicholaslriderdo naturallanguageprocessingofclinicalnotesenablesearlyinbornerrorofimmunityriskascertainment |