Search

Yujia Liu,1 Shangyuan Ye,2 Xianchao Xiao,1 Chenglin Sun,1 Gang Wang,1 Guixia Wang,1 Bo Zhang3 1Department of Endocrinology and Metabolism, The First Hospital of Jilin University, Changchun, Jilin 130021, People’s Republic of China; 2Department of Population Medicine, Harvard Pilgrim Health Care and Harvard Medical School, Boston, MA, USA; 3Department of Neurology and ICCTR Biostatistics and Research Design Center, Boston Children’s Hospital and Harvard Medical School, Boston, MA, USACorrespondence: Bo ZhangDepartment of Neurology and ICCTR Biostatistics and Research Design Center, Boston Children’s Hospital and Harvard Medical School, 21 Autumn Street, Boston, MA 02115, USAEmail bo.zhang@childrens.harvard.edu Guixia WangDepartment of Endocrinology and Metabolism, The First Hospital of Jilin University, 71 Xinmin Street, Changchun, Jilin 130021, People’s Republic of ChinaEmail gwang168@jlu.edu.cnBackground: This study proposes the use of machine learning algorithms to improve the accuracy of type 2 diabetes predictions using non-invasive risk score systems.Methods: We evaluated and compared the prediction accuracies of existing non-invasive risk score systems using the data from the REACTION study (Risk Evaluation of Cancers in Chinese Diabetic Individuals: A Longitudinal Study). Two simple risk scores were established on the bases of logistic regression. Machine learning techniques (ensemble methods) were used to improve prediction accuracies by combining the individual score systems.Results: Existing score systems from Western populations performed worse than the scores from Eastern populations in general. The two newly established score systems performed better than most existing scores systems but a little worse than the Chinese score system. Using ensemble methods with model selection algorithms yielded better prediction accuracy than all the simple score systems.Conclusion: Our proposed machine learning methods can be used to improve the accuracy of screening the undiagnosed type 2 diabetes and identifying the high-risk patients.Keywords: type 2 diabetes, risk score, machine learning, voting, stacking, prediction

Yujia Liu,1 Shangyuan Ye,2 Xianchao Xiao,1 Tong Zhou,1 Shuo Yang,1 Gang Wang,1 Chenglin Sun,1 Bo Zhang,2 Guixia Wang1 1Department of Endocrinology and Metabolism, The First Hospital of Jilin University, Changchun, Jilin, People’s Republic of China; 2Department of Population and Quantitative Health Sciences, University of Massachusetts Medical School, Worcester, MA, USACorrespondence: Bo ZhangDepartment of Population and Quantitative Health Sciences, University of Massachusetts Medical School, Worcester, MA, USAEmail bo.zhang@umassmed.eduGuixia WangDepartment of Endocrinology and Metabolism, The First Hospital of Jilin University, Changchun, Jilin, People’s Republic of ChinaEmail gwang168@jlu.edu.cnObjective: To study the association between glucose metabolism disorders and hepatotropic virus infection.Methods: A cross-sectional analysis was performed using data from the REACTION study (Risk Evaluation of Cancers in Chinese Diabetic Individuals: A Longitudinal Study). Outcomes of the analysis were test results of kidney function, liver function, lipid metabolism, and the prevalence of hepatitis B virus (HBV) infection and potential hepatitis C virus (HCV) infection (positive hepatitis C virus antibody) among individuals with and without diabetes mellitus (DM) or pre-diabetes mellitus (pre-DM).Results: Of the 10,080 patients who participated in the study, 7665 eligible subjects were included in the analysis. There was no significant difference in the prevalence of HBV infection between DM and normal subjects, pre-DM and normal subjects, and DM or pre-DM and normal subjects (p-values of 0.9180, 0.8154, and 0.6448, respectively). There was also no significant difference in the prevalence of potential HCV infection between DM and normal subjects, pre-DM and normal subjects, and DM or pre-DM and normal subjects (p-values of 0.1190, 0.0591, and 0.5591, respectively). Lipid metabolism showed a significant difference between DM or pre-DM subjects and normal subjects (p-values were less than 0.0221 in all cases). Multiple logistic regression analysis revealed hypertension as the leading significant variable associated with DM, pre-DM, and both. Other significant factors included gender, body mass index, age, and alanine aminotransferase.Conclusion: No significant association was detected between DM or pre-DM and HBV or potential HCV infection. Significant association was detected between lipid metabolism disorders and DM, but this association was absent in pre-DM patients when adjusting for other factors.Keywords: diabetes, pre-diabetes, lipid metabolism disorder, hepatitis B, hepatitis C