{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,1]],"date-time":"2026-08-01T00:44:18Z","timestamp":1785545058745,"version":"3.56.0"},"reference-count":33,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["J Med Internet Res"],"abstract":"Abstract<\/jats:title>\n \n Background<\/jats:title>\n Multicenter electronic health records (EHRs) can support quality improvement and comparative effectiveness research in critical care. However, limitations of EHR-based research include challenges in abstracting key clinical variables, including a patient\u2019s level of consciousness.<\/jats:p>\n <\/jats:sec>\n \n Objective<\/jats:title>\n This study aimed to develop a natural language processing model to predict Glasgow Coma Scale (GCS) scores from daily EHR notes.<\/jats:p>\n <\/jats:sec>\n \n Methods<\/jats:title>\n The study included adult patients (aged \u226518 years) admitted to Mass General Brigham (MGB) hospitals (2017\u20102024) and patients from the Medical Information Mart for Intensive Care-III (MIMIC-III version 1.4; 2001\u20102012) database. A dataset of all patients from both institutions was split into training (70%) or hold-out test (30%) sets. Variables consisted of daily notes, age, sex, and admission type. A pooled ordinal regression model (ordinalNet) with an elastic net penalty was trained to predict the lowest daily level of consciousness across 3 classes of impairment: severe (GCS score 3\u20108), moderate (GCS score 9\u201012), and mild (GCS score 13\u201015), and a pooled linear model was trained to predict continuous GCS scores (3-15). Gold standard GCS was obtained from structured flowsheet data. External generalizability was assessed using a single-institution ordinal model trained on MGB and tested on MIMIC. Following post hoc calibration, the performance of the ordinal and linear models was evaluated on the hold-out test sets using the area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve (AUPRC) for the ordinal models and root mean square error and Pearson correlation coefficient for the linear models.<\/jats:p>\n <\/jats:sec>\n \n Results<\/jats:title>\n The modeling cohort included 145,897 patients (MGB: n=123,257, MIMIC: n=22,640), with 1,446,965 days of hospitalization between training and testing sets; the average age was 62 (SD 18) years, and the sex distribution was balanced. The pooled ordinalNet achieved an AUROC of 0.96 (95% CI 0.96\u20100.96) and an AUPRC of 0.77 (95% CI 0.76\u20100.77). The single-institution ordinal model achieved an AUROC of 0.90 (95% CI 0.89\u20100.90) and an AUPRC of 0.80 (95% CI 0.79\u20100.80). The pooled linear model achieved a root mean square error of 2.30 (95% CI 2.30\u20102.30) and a correlation of 0.76 (95% CI 0.76\u20100.76). Predictions for severe GCS were driven by terms indicating unresponsiveness and critical interventions, moderate GCS by intermediate alertness descriptors, and mild GCS by mentions of normal or awake behavior.<\/jats:p>\n <\/jats:sec>\n \n Conclusions<\/jats:title>\n Pooled ordinal and linear models can accurately predict GCS from unstructured data and can support large-scale phenotyping of neurological assessments for future critical care research.<\/jats:p>\n <\/jats:sec>","DOI":"10.2196\/81245","type":"journal-article","created":{"date-parts":[[2026,4,20]],"date-time":"2026-04-20T14:50:09Z","timestamp":1776696609000},"page":"e81245-e81245","source":"Crossref","is-referenced-by-count":0,"title":["Automated Prediction of Glasgow Coma Scale Scores From Unstructured Electronic Health Records Using Natural Language Processing: Development and Validation Study"],"prefix":"10.2196","volume":"28","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-7203-2832","authenticated-orcid":false,"given":"Marta","family":"Fernandes","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-4806-578X","authenticated-orcid":false,"given":"Niels","family":"Turley","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5041-8312","authenticated-orcid":false,"given":"Haoqi","family":"Sun","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5677-6954","authenticated-orcid":false,"given":"Shibani S","family":"Mukerji","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1191-1315","authenticated-orcid":false,"given":"Lidia M V R","family":"Moura","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4803-312X","authenticated-orcid":false,"given":"M Brandon","family":"Westover","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5252-5376","authenticated-orcid":false,"given":"Sahar F","family":"Zafar","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","published-online":{"date-parts":[[2026,6,29]]},"reference":[{"key":"R1","unstructured":"Jain S Margetis K Iverson LM . Glasgow Coma Scale. In: StatPearls [Internet]. StatPearls Publishing; 2025. URL: https:\/\/www.ncbi.nlm.nih.gov\/books\/NBK513298\/ [Accessed 16-06-2026]"},{"issue":"10","key":"R2","doi-asserted-by":"crossref","first-page":"818","DOI":"10.1097\/00003246-198510000-00009","volume":"13","author":"Knaus","journal-title":"Crit Care Med"},{"issue":"7","key":"R3","doi-asserted-by":"publisher","first-page":"707","DOI":"10.1007\/BF01709751","article-title":"The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction\/failure. On behalf of the working group on sepsis-related problems of the European Society of Intensive Care Medicine","volume":"22","author":"Vincent","journal-title":"Intensive Care Med"},{"issue":"24","key":"R4","doi-asserted-by":"publisher","first-page":"2957","DOI":"10.1001\/jama.270.24.2957","article-title":"A new Simplified Acute Physiology Score (SAPS II) based on a European\/North American multicenter study","volume":"270","author":"Le Gall","journal-title":"JAMA"},{"issue":"7641","key":"R5","doi-asserted-by":"publisher","first-page":"425","DOI":"10.1136\/bmj.39461.643438.25","article-title":"Predicting outcome after traumatic brain injury: practical prognostic models based on large cohort of international patients","volume":"336","author":"MRC CRASH Trial Collaborators","journal-title":"BMJ"},{"issue":"8","key":"R6","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pmed.0050165","article-title":"Predicting outcome after traumatic brain injury: development and international validation of prognostic scores based on admission characteristics","volume":"5","author":"Steyerberg","journal-title":"PLoS Med"},{"issue":"6","key":"R7","doi-asserted-by":"publisher","first-page":"962","DOI":"10.1097\/00005373-199412000-00016","article-title":"Comparison of mortality, morbidity, and severity of 59,713 head injured patients with 114,447 patients with extracranial injuries","volume":"37","author":"Gennarelli","journal-title":"J Trauma"},{"issue":"6","key":"R8","doi-asserted-by":"publisher","first-page":"1238","DOI":"10.1097\/01.ta.0000195593.60245.80","article-title":"Statistical validation of the Glasgow Coma Score","volume":"60","author":"Moore","journal-title":"J Trauma"},{"key":"R9","unstructured":"Field triage guideline revision: Glasgow Coma Scale: systematic review. Effective Health Care Program, Agency for Healthcare Research and Quality; 2016. URL: https:\/\/effectivehealthcare.ahrq.gov\/sites\/default\/files\/pdf\/field-triage-glascow_research-protocol.pdf [Accessed 16-06-2026]"},{"key":"R10","doi-asserted-by":"publisher","DOI":"10.1016\/j.wneu.2024.02.139","article-title":"The message of the Glasgow Coma Scale: a comprehensive bibliometric analysis and systematic review of clinical practice guidelines spanning the past 50 years","volume":"185","author":"Zhang","journal-title":"World Neurosurg"},{"issue":"3","key":"R11","first-page":"698","volume":"59","author":"Moore","journal-title":"J Trauma"},{"issue":"4","key":"R12","doi-asserted-by":"publisher","first-page":"373","DOI":"10.1111\/j.1547-5069.2008.00252.x","article-title":"Managing the common problem of missing data in trauma studies","volume":"40","author":"Rue","journal-title":"J Nurs Scholarsh"},{"issue":"5","key":"R13","doi-asserted-by":"publisher","first-page":"548","DOI":"10.1177\/0310057X0303100509","article-title":"Outcome prediction in a surgical ICU using automatically calculated SAPS II scores","volume":"31","author":"Engel","journal-title":"Anaesth Intensive Care"},{"issue":"8","key":"R14","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0253425","article-title":"Imputation strategies for missing baseline neurological assessment covariates after traumatic brain injury: a CENTER-TBI study","volume":"16","author":"Ercole","journal-title":"PLoS One"},{"key":"R15","volume":"2017","author":"Boag","journal-title":"AMIA Jt Summits Transl Sci Proc"},{"key":"R16","doi-asserted-by":"publisher","DOI":"10.1136\/bmj-2023-078378","article-title":"TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods","volume":"385","author":"Collins","journal-title":"BMJ"},{"key":"R17","doi-asserted-by":"publisher","DOI":"10.1038\/sdata.2016.35","article-title":"MIMIC-III, a freely accessible critical care database","volume":"3","author":"Johnson","journal-title":"Sci Data"},{"issue":"2","key":"R18","doi-asserted-by":"publisher","first-page":"301","DOI":"10.1111\/j.1467-9868.2005.00503.x","article-title":"Regularization and variable selection via the elastic net","volume":"67","author":"Zou","journal-title":"J R Stat Soc Series B Stat Methodol"},{"issue":"9","key":"R19","doi-asserted-by":"publisher","first-page":"1932","DOI":"10.1016\/j.injury.2017.05.038","article-title":"Differential effects of the Glasgow Coma Scale Score and its components: an analysis of 54,069 patients with traumatic brain injury","volume":"48","author":"Reith","journal-title":"Injury"},{"issue":"2","key":"R20","doi-asserted-by":"publisher","first-page":"417","DOI":"10.3171\/2023.11.JNS231425","article-title":"Performance of the IMPACT and CRASH prognostic models for traumatic brain injury in a contemporary multicenter cohort: a TRACK-TBI study","volume":"141","author":"Yue","journal-title":"J Neurosurg"},{"issue":"3","key":"R21","doi-asserted-by":"publisher","first-page":"260","DOI":"10.1097\/MRR.0b013e32835fd99a","article-title":"Prognostic value of time-related Glasgow Coma Scale components in severe traumatic brain injury: a prospective evaluation with respect to 1-year survival and functional outcome","volume":"36","author":"Kouloulas","journal-title":"Int J Rehabil Res"},{"issue":"4","key":"R22","doi-asserted-by":"publisher","first-page":"273","DOI":"10.1097\/HTR.0000000000000938","article-title":"The predictive value of the verbal Glasgow Coma Scale in traumatic brain injury: a systematic review","volume":"39","author":"Pisano","journal-title":"J Head Trauma Rehabil"},{"issue":"6","key":"R23","doi-asserted-by":"publisher","first-page":"1612","DOI":"10.3171\/2017.12.JNS172780","article-title":"Simplifying the use of prognostic information in traumatic brain injury. Part 1: the GCS-Pupils score: an extended index of clinical severity","volume":"128","author":"Brennan","journal-title":"J Neurosurg"},{"key":"R24","doi-asserted-by":"publisher","DOI":"10.1016\/j.jocn.2024.110994","article-title":"50 years of the Glasgow Coma Scale: a historical perspective","volume":"133","author":"Wells","journal-title":"J Clin Neurosci"},{"issue":"1178","key":"R25","doi-asserted-by":"publisher","first-page":"1237","DOI":"10.1093\/postmj\/qgad071","article-title":"Glasgow Coma Scale on admission as predictor of neurological sequelae at discharge and acute respiratory failure in patients with heatstroke","volume":"99","author":"Chen","journal-title":"Postgrad Med J"},{"issue":"6","key":"R26","doi-asserted-by":"publisher","first-page":"755","DOI":"10.1016\/j.jamcollsurg.2017.08.020","article-title":"Inclusion of highest Glasgow Coma Scale motor component score in mortality risk adjustment for benchmarking of trauma center performance","volume":"225","author":"Gomez","journal-title":"J Am Coll Surg"},{"key":"R27","doi-asserted-by":"publisher","DOI":"10.3389\/fneur.2025.1607946","article-title":"Associations between Glasgow Coma Scale trajectories and 28-day survival rate in patients with sepsis-associated encephalopathy: insights from longitudinal group trajectory modeling","volume":"16","author":"Shen","journal-title":"Front Neurol"},{"key":"R28","volume":"2017","author":"Butler","journal-title":"AMIA Jt Summits Transl Sci Proc"},{"issue":"1","key":"R29","doi-asserted-by":"publisher","DOI":"10.1186\/s13054-025-05677-0","article-title":"Discovery of data quality issues in electronic health records: profound consequences for critical care medicine applications - a systematized review","volume":"30","author":"de Andrade","journal-title":"Crit Care"},{"key":"R30","doi-asserted-by":"publisher","DOI":"10.2196\/42615","article-title":"Digital health data quality issues: systematic review","volume":"25","author":"Syed","journal-title":"J Med Internet Res"},{"issue":"3","key":"R31","doi-asserted-by":"publisher","DOI":"10.13063\/2327-9214.1035","article-title":"Strategies for handling missing data in electronic health record derived data","volume":"1","author":"Wells","journal-title":"EGEMS (Wash DC)"},{"key":"R32","doi-asserted-by":"publisher","DOI":"10.2196\/57615","article-title":"Data quality-driven improvement in health care: systematic literature review","volume":"26","author":"Lighterness","journal-title":"J Med Internet Res"},{"key":"R33","unstructured":"Automated prediction of Glasgow Coma Scale scores from unstructured electronic health records: a natural language processing approach 1.0.0. Brain Data Science Platform. URL: https:\/\/bdsp.io\/content\/gd2l2o4nrl4dl9sok4bd\/1.0.0\/%20(https:\/\/doi.org\/10.60508\/3fgr-8g80 [Accessed 25-06-2026]"}],"container-title":["Journal of Medical Internet Research"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.jmir.org\/2026\/1\/e81245","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,31]],"date-time":"2026-07-31T22:26:33Z","timestamp":1785536793000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.jmir.org\/2026\/1\/e81245"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,29]]},"references-count":33,"URL":"https:\/\/doi.org\/10.2196\/81245","relation":{},"ISSN":["1438-8871"],"issn-type":[{"value":"1438-8871","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,6,29]]},"article-number":"v28i10e81245"}}