Predicting Serum Levels of Lithium-Treated Patients: A Supervised Machine Learning Approach

Routine monitoring of lithium levels is common clinical practice. This is because the lithium prediction strategies available developed by previous studies are still limited due to insufficient prediction performance. Thus, we used machine learning approaches to predict lithium concentration in a la...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Chih-Wei Hsu, Shang-Ying Tsai, Liang-Jen Wang, Chih-Sung Liang, Andre F. Carvalho, Marco Solmi, Eduard Vieta, Pao-Yen Lin, Chien-An Hu, Hung-Yu Kao
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	bipolar disorder lithium machine learning random forest support vector machine therapeutic drug monitoring Biology (General) QH301-705.5
Acceso en línea:	https://doaj.org/article/916e42c8d5484171be0701de89439476
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:916e42c8d5484171be0701de89439476
record_format	dspace
spelling	oai:doaj.org-article:916e42c8d5484171be0701de894394762021-11-25T16:49:00ZPredicting Serum Levels of Lithium-Treated Patients: A Supervised Machine Learning Approach10.3390/biomedicines91115582227-9059https://doaj.org/article/916e42c8d5484171be0701de894394762021-10-01T00:00:00Zhttps://www.mdpi.com/2227-9059/9/11/1558https://doaj.org/toc/2227-9059Routine monitoring of lithium levels is common clinical practice. This is because the lithium prediction strategies available developed by previous studies are still limited due to insufficient prediction performance. Thus, we used machine learning approaches to predict lithium concentration in a large real-world dataset. Real-world data from multicenter electronic medical records were used in different machine learning algorithms to predict: (1) whether the serum level was 0.6–1.2 mmol/L or 0.0–0.6 mmol/L (binary prediction), and (2) its concentration value (continuous prediction). We developed models from 1505 samples through 5-fold cross-validation and used 204 independent samples to test their performance by evaluating their accuracy. Moreover, we ranked the most important clinical features in different models and reconstructed three reduced models with fewer clinical features. For binary and continuous predictions, the average accuracy of these models was 0.70–0.73 and 0.68–0.75, respectively. Seven features were listed as important features related to serum lithium levels of 0.6–1.2 mmol/L or higher lithium concentration, namely older age, lower systolic blood pressure, higher daily and last doses of lithium prescription, concomitant psychotropic drugs with valproic acid and -pine drugs, and comorbid substance-related disorders. After reducing the features in the three new predictive models, the binary or continuous models still had an average accuracy of 0.67–0.74. Machine learning processes complex clinical data and provides a potential tool for predicting lithium concentration. This may help in clinical decision-making and reduce the frequency of serum level monitoring.Chih-Wei HsuShang-Ying TsaiLiang-Jen WangChih-Sung LiangAndre F. CarvalhoMarco SolmiEduard VietaPao-Yen LinChien-An HuHung-Yu KaoMDPI AGarticlebipolar disorderlithiummachine learningrandom forestsupport vector machinetherapeutic drug monitoringBiology (General)QH301-705.5ENBiomedicines, Vol 9, Iss 1558, p 1558 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	bipolar disorder lithium machine learning random forest support vector machine therapeutic drug monitoring Biology (General) QH301-705.5
spellingShingle	bipolar disorder lithium machine learning random forest support vector machine therapeutic drug monitoring Biology (General) QH301-705.5 Chih-Wei Hsu Shang-Ying Tsai Liang-Jen Wang Chih-Sung Liang Andre F. Carvalho Marco Solmi Eduard Vieta Pao-Yen Lin Chien-An Hu Hung-Yu Kao Predicting Serum Levels of Lithium-Treated Patients: A Supervised Machine Learning Approach
description	Routine monitoring of lithium levels is common clinical practice. This is because the lithium prediction strategies available developed by previous studies are still limited due to insufficient prediction performance. Thus, we used machine learning approaches to predict lithium concentration in a large real-world dataset. Real-world data from multicenter electronic medical records were used in different machine learning algorithms to predict: (1) whether the serum level was 0.6–1.2 mmol/L or 0.0–0.6 mmol/L (binary prediction), and (2) its concentration value (continuous prediction). We developed models from 1505 samples through 5-fold cross-validation and used 204 independent samples to test their performance by evaluating their accuracy. Moreover, we ranked the most important clinical features in different models and reconstructed three reduced models with fewer clinical features. For binary and continuous predictions, the average accuracy of these models was 0.70–0.73 and 0.68–0.75, respectively. Seven features were listed as important features related to serum lithium levels of 0.6–1.2 mmol/L or higher lithium concentration, namely older age, lower systolic blood pressure, higher daily and last doses of lithium prescription, concomitant psychotropic drugs with valproic acid and -pine drugs, and comorbid substance-related disorders. After reducing the features in the three new predictive models, the binary or continuous models still had an average accuracy of 0.67–0.74. Machine learning processes complex clinical data and provides a potential tool for predicting lithium concentration. This may help in clinical decision-making and reduce the frequency of serum level monitoring.
format	article
author	Chih-Wei Hsu Shang-Ying Tsai Liang-Jen Wang Chih-Sung Liang Andre F. Carvalho Marco Solmi Eduard Vieta Pao-Yen Lin Chien-An Hu Hung-Yu Kao
author_facet	Chih-Wei Hsu Shang-Ying Tsai Liang-Jen Wang Chih-Sung Liang Andre F. Carvalho Marco Solmi Eduard Vieta Pao-Yen Lin Chien-An Hu Hung-Yu Kao
author_sort	Chih-Wei Hsu
title	Predicting Serum Levels of Lithium-Treated Patients: A Supervised Machine Learning Approach
title_short	Predicting Serum Levels of Lithium-Treated Patients: A Supervised Machine Learning Approach
title_full	Predicting Serum Levels of Lithium-Treated Patients: A Supervised Machine Learning Approach
title_fullStr	Predicting Serum Levels of Lithium-Treated Patients: A Supervised Machine Learning Approach
title_full_unstemmed	Predicting Serum Levels of Lithium-Treated Patients: A Supervised Machine Learning Approach
title_sort	predicting serum levels of lithium-treated patients: a supervised machine learning approach
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/916e42c8d5484171be0701de89439476
work_keys_str_mv	AT chihweihsu predictingserumlevelsoflithiumtreatedpatientsasupervisedmachinelearningapproach AT shangyingtsai predictingserumlevelsoflithiumtreatedpatientsasupervisedmachinelearningapproach AT liangjenwang predictingserumlevelsoflithiumtreatedpatientsasupervisedmachinelearningapproach AT chihsungliang predictingserumlevelsoflithiumtreatedpatientsasupervisedmachinelearningapproach AT andrefcarvalho predictingserumlevelsoflithiumtreatedpatientsasupervisedmachinelearningapproach AT marcosolmi predictingserumlevelsoflithiumtreatedpatientsasupervisedmachinelearningapproach AT eduardvieta predictingserumlevelsoflithiumtreatedpatientsasupervisedmachinelearningapproach AT paoyenlin predictingserumlevelsoflithiumtreatedpatientsasupervisedmachinelearningapproach AT chienanhu predictingserumlevelsoflithiumtreatedpatientsasupervisedmachinelearningapproach AT hungyukao predictingserumlevelsoflithiumtreatedpatientsasupervisedmachinelearningapproach
_version_	1718412969881108480

Predicting Serum Levels of Lithium-Treated Patients: A Supervised Machine Learning Approach

Ejemplares similares