Sara Knox, Stephanie Aghamoosa, Paul M. Heider, Maxwell Cutty, Andrew Wright, Dmitry Scherbakov, Gabriel Hood, Sara A. Nolin, Jihad S. Obeid
{"title":"利用电子健康记录对阿尔茨海默病和相关痴呆症进行表型分析的人工智能方法","authors":"Sara Knox, Stephanie Aghamoosa, Paul M. Heider, Maxwell Cutty, Andrew Wright, Dmitry Scherbakov, Gabriel Hood, Sara A. Nolin, Jihad S. Obeid","doi":"10.1002/trc2.70089","DOIUrl":null,"url":null,"abstract":"<div>\n \n \n <section>\n \n <h3> INTRODUCTION</h3>\n \n <p>The current standard electronic (e-)phenotype for identifying patients with Alzheimer's disease and related dementias (ADRD) from medical claims data yields suboptimal diagnostic accuracy. This study leveraged artificial intelligence (AI)–based text-classification methods to improve the identification of patients with dementia due to ADRD using clinical notes from electronic health records (EHRs).</p>\n </section>\n \n <section>\n \n <h3> METHODS</h3>\n \n <p>EHR data for patients aged ≥ 64 (<i>N</i> = 4000) from an academic medical center were used. The cohort included 1000 patients with ADRD per the Chronic Conditions Warehouse (CCW) algorithm for ADRD (i.e., at least one ADRD International Classification of Diseases, Tenth Revision codes [ICD-10 code]) and 3000 matched controls without ADRD (i.e., no CCW codes). We trained several AI-based text-classification models, including bag-of-words models, deep learning, and large language models (LLMs), to make ADRD determinations from clinical notes. The performance of each model was evaluated against “gold standard” manual chart review.</p>\n </section>\n \n <section>\n \n <h3> RESULTS</h3>\n \n <p>A foundational LLM derived from Llama 2 demonstrated superior performance in identifying patients with ADRD (area under the curve [AUC] = 0.9534, <i>F</i><sub>1</sub> score 0.8571) compared to both the current standard CCW algorithm (AUC = 0.8482, <i>F</i><sub>1</sub> score 0.8323, although only the AUC was statistically significantly different) and other AI-based models. Several of the AI-based models, including convolutional neural networks, also outperformed the CCW algorithm.</p>\n </section>\n \n <section>\n \n <h3> DISCUSSION</h3>\n \n <p>These findings highlight the potential of AI-based text-classification methods to optimize the automated identification of patients with ADRD using rich EHR data. However, the success of this approach depends on the quality of clinical notes, and more work is needed to refine and validate these methods across more diverse data sets.</p>\n </section>\n \n <section>\n \n <h3> Highlights</h3>\n \n <div>\n <ul>\n \n <li>The current e-phenotype for patients with Alzheimer's disease and related dementias (ADRD) in electronic health records has suboptimal diagnostic accuracy.</li>\n \n <li>The study used artificial intelligence (AI)–based text classification methods to improve the detection of patients with ADRD.</li>\n \n <li>AI-based models, including convolutional neural networks, outperformed the Chronic Conditions Warehouse algorithm.</li>\n </ul>\n </div>\n </section>\n </div>","PeriodicalId":53225,"journal":{"name":"Alzheimer''s and Dementia: Translational Research and Clinical Interventions","volume":"11 2","pages":""},"PeriodicalIF":4.9000,"publicationDate":"2025-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/trc2.70089","citationCount":"0","resultStr":"{\"title\":\"AI approaches for phenotyping Alzheimer's disease and related dementias using electronic health records\",\"authors\":\"Sara Knox, Stephanie Aghamoosa, Paul M. Heider, Maxwell Cutty, Andrew Wright, Dmitry Scherbakov, Gabriel Hood, Sara A. Nolin, Jihad S. Obeid\",\"doi\":\"10.1002/trc2.70089\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div>\\n \\n \\n <section>\\n \\n <h3> INTRODUCTION</h3>\\n \\n <p>The current standard electronic (e-)phenotype for identifying patients with Alzheimer's disease and related dementias (ADRD) from medical claims data yields suboptimal diagnostic accuracy. This study leveraged artificial intelligence (AI)–based text-classification methods to improve the identification of patients with dementia due to ADRD using clinical notes from electronic health records (EHRs).</p>\\n </section>\\n \\n <section>\\n \\n <h3> METHODS</h3>\\n \\n <p>EHR data for patients aged ≥ 64 (<i>N</i> = 4000) from an academic medical center were used. The cohort included 1000 patients with ADRD per the Chronic Conditions Warehouse (CCW) algorithm for ADRD (i.e., at least one ADRD International Classification of Diseases, Tenth Revision codes [ICD-10 code]) and 3000 matched controls without ADRD (i.e., no CCW codes). We trained several AI-based text-classification models, including bag-of-words models, deep learning, and large language models (LLMs), to make ADRD determinations from clinical notes. The performance of each model was evaluated against “gold standard” manual chart review.</p>\\n </section>\\n \\n <section>\\n \\n <h3> RESULTS</h3>\\n \\n <p>A foundational LLM derived from Llama 2 demonstrated superior performance in identifying patients with ADRD (area under the curve [AUC] = 0.9534, <i>F</i><sub>1</sub> score 0.8571) compared to both the current standard CCW algorithm (AUC = 0.8482, <i>F</i><sub>1</sub> score 0.8323, although only the AUC was statistically significantly different) and other AI-based models. Several of the AI-based models, including convolutional neural networks, also outperformed the CCW algorithm.</p>\\n </section>\\n \\n <section>\\n \\n <h3> DISCUSSION</h3>\\n \\n <p>These findings highlight the potential of AI-based text-classification methods to optimize the automated identification of patients with ADRD using rich EHR data. However, the success of this approach depends on the quality of clinical notes, and more work is needed to refine and validate these methods across more diverse data sets.</p>\\n </section>\\n \\n <section>\\n \\n <h3> Highlights</h3>\\n \\n <div>\\n <ul>\\n \\n <li>The current e-phenotype for patients with Alzheimer's disease and related dementias (ADRD) in electronic health records has suboptimal diagnostic accuracy.</li>\\n \\n <li>The study used artificial intelligence (AI)–based text classification methods to improve the detection of patients with ADRD.</li>\\n \\n <li>AI-based models, including convolutional neural networks, outperformed the Chronic Conditions Warehouse algorithm.</li>\\n </ul>\\n </div>\\n </section>\\n </div>\",\"PeriodicalId\":53225,\"journal\":{\"name\":\"Alzheimer''s and Dementia: Translational Research and Clinical Interventions\",\"volume\":\"11 2\",\"pages\":\"\"},\"PeriodicalIF\":4.9000,\"publicationDate\":\"2025-04-24\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://onlinelibrary.wiley.com/doi/epdf/10.1002/trc2.70089\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Alzheimer''s and Dementia: Translational Research and Clinical Interventions\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://onlinelibrary.wiley.com/doi/10.1002/trc2.70089\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"CLINICAL NEUROLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Alzheimer''s and Dementia: Translational Research and Clinical Interventions","FirstCategoryId":"1085","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1002/trc2.70089","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"CLINICAL NEUROLOGY","Score":null,"Total":0}
AI approaches for phenotyping Alzheimer's disease and related dementias using electronic health records
INTRODUCTION
The current standard electronic (e-)phenotype for identifying patients with Alzheimer's disease and related dementias (ADRD) from medical claims data yields suboptimal diagnostic accuracy. This study leveraged artificial intelligence (AI)–based text-classification methods to improve the identification of patients with dementia due to ADRD using clinical notes from electronic health records (EHRs).
METHODS
EHR data for patients aged ≥ 64 (N = 4000) from an academic medical center were used. The cohort included 1000 patients with ADRD per the Chronic Conditions Warehouse (CCW) algorithm for ADRD (i.e., at least one ADRD International Classification of Diseases, Tenth Revision codes [ICD-10 code]) and 3000 matched controls without ADRD (i.e., no CCW codes). We trained several AI-based text-classification models, including bag-of-words models, deep learning, and large language models (LLMs), to make ADRD determinations from clinical notes. The performance of each model was evaluated against “gold standard” manual chart review.
RESULTS
A foundational LLM derived from Llama 2 demonstrated superior performance in identifying patients with ADRD (area under the curve [AUC] = 0.9534, F1 score 0.8571) compared to both the current standard CCW algorithm (AUC = 0.8482, F1 score 0.8323, although only the AUC was statistically significantly different) and other AI-based models. Several of the AI-based models, including convolutional neural networks, also outperformed the CCW algorithm.
DISCUSSION
These findings highlight the potential of AI-based text-classification methods to optimize the automated identification of patients with ADRD using rich EHR data. However, the success of this approach depends on the quality of clinical notes, and more work is needed to refine and validate these methods across more diverse data sets.
Highlights
The current e-phenotype for patients with Alzheimer's disease and related dementias (ADRD) in electronic health records has suboptimal diagnostic accuracy.
The study used artificial intelligence (AI)–based text classification methods to improve the detection of patients with ADRD.
AI-based models, including convolutional neural networks, outperformed the Chronic Conditions Warehouse algorithm.
期刊介绍：
Alzheimer''s & Dementia: Translational Research & Clinical Interventions (TRCI) is a peer-reviewed, open access,journal from the Alzheimer''s Association®. The journal seeks to bridge the full scope of explorations between basic research on drug discovery and clinical studies, validating putative therapies for aging-related chronic brain conditions that affect cognition, motor functions, and other behavioral or clinical symptoms associated with all forms dementia and Alzheimer''s disease. The journal will publish findings from diverse domains of research and disciplines to accelerate the conversion of abstract facts into practical knowledge: specifically, to translate what is learned at the bench into bedside applications. The journal seeks to publish articles that go beyond a singular emphasis on either basic drug discovery research or clinical research. Rather, an important theme of articles will be the linkages between and among the various discrete steps in the complex continuum of therapy development. For rapid communication among a multidisciplinary research audience involving the range of therapeutic interventions, TRCI will consider only original contributions that include feature length research articles, systematic reviews, meta-analyses, brief reports, narrative reviews, commentaries, letters, perspectives, and research news that would advance wide range of interventions to ameliorate symptoms or alter the progression of chronic neurocognitive disorders such as dementia and Alzheimer''s disease. The journal will publish on topics related to medicine, geriatrics, neuroscience, neurophysiology, neurology, psychiatry, clinical psychology, bioinformatics, pharmaco-genetics, regulatory issues, health economics, pharmacoeconomics, and public health policy as these apply to preclinical and clinical research on therapeutics.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
