scholarly journals An Innovative Artificial Intelligence–Based App for the Diagnosis of Gestational Diabetes Mellitus (GDM-AI): Development Study (Preprint)

2020 ◽  
Author(s):  
Jiayi Shen ◽  
Jiebin Chen ◽  
Zequan Zheng ◽  
Jiabin Zheng ◽  
Zherui Liu ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Diabetes Mellitus ◽  
Blood Glucose ◽  
Gestational Diabetes ◽  
Pregnant Women ◽  
Fasting Blood Glucose ◽  
External Validation ◽  
Validation Dataset ◽  
Gradient Boosting ◽  
Local Hospital

BACKGROUND Gestational diabetes mellitus (GDM) can cause adverse consequences to both mothers and their newborns. However, pregnant women living in low- and middle-income areas or countries often fail to receive early clinical interventions at local medical facilities due to restricted availability of GDM diagnosis. The outstanding performance of artificial intelligence (AI) in disease diagnosis in previous studies demonstrates its promising applications in GDM diagnosis. OBJECTIVE This study aims to investigate the implementation of a well-performing AI algorithm in GDM diagnosis in a setting, which requires fewer medical equipment and staff and to establish an app based on the AI algorithm. This study also explores possible progress if our app is widely used. METHODS An AI model that included 9 algorithms was trained on 12,304 pregnant outpatients with their consent who received a test for GDM in the obstetrics and gynecology department of the First Affiliated Hospital of Jinan University, a local hospital in South China, between November 2010 and October 2017. GDM was diagnosed according to American Diabetes Association (ADA) 2011 diagnostic criteria. Age and fasting blood glucose were chosen as critical parameters. For validation, we performed k-fold cross-validation (k=5) for the internal dataset and an external validation dataset that included 1655 cases from the Prince of Wales Hospital, the affiliated teaching hospital of the Chinese University of Hong Kong, a non-local hospital. Accuracy, sensitivity, and other criteria were calculated for each algorithm. RESULTS The areas under the receiver operating characteristic curve (AUROC) of external validation dataset for support vector machine (SVM), random forest, AdaBoost, k-nearest neighbors (kNN), naive Bayes (NB), decision tree, logistic regression (LR), eXtreme gradient boosting (XGBoost), and gradient boosting decision tree (GBDT) were 0.780, 0.657, 0.736, 0.669, 0.774, 0.614, 0.769, 0.742, and 0.757, respectively. SVM also retained high performance in other criteria. The specificity for SVM retained 100% in the external validation set with an accuracy of 88.7%. CONCLUSIONS Our prospective and multicenter study is the first clinical study that supports the GDM diagnosis for pregnant women in resource-limited areas, using only fasting blood glucose value, patients’ age, and a smartphone connected to the internet. Our study proved that SVM can achieve accurate diagnosis with less operation cost and higher efficacy. Our study (referred to as GDM-AI study, ie, the study of AI-based diagnosis of GDM) also shows our app has a promising future in improving the quality of maternal health for pregnant women, precision medicine, and long-distance medical care. We recommend future work should expand the dataset scope and replicate the process to validate the performance of the AI algorithms.

scholarly journals An Innovative Artificial Intelligence–Based App for the Diagnosis of Gestational Diabetes Mellitus (GDM-AI): Development Study

10.2196/21573 ◽  
2020 ◽  
Vol 22 (9) ◽  
pp. e21573 ◽  
Author(s):  
Jiayi Shen ◽  
Jiebin Chen ◽  
Zequan Zheng ◽  
Jiabin Zheng ◽  
Zherui Liu ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Diabetes Mellitus ◽  
Blood Glucose ◽  
Gestational Diabetes ◽  
Pregnant Women ◽  
Fasting Blood Glucose ◽  
External Validation ◽  
Validation Dataset ◽  
Gradient Boosting ◽  
Local Hospital

Background Gestational diabetes mellitus (GDM) can cause adverse consequences to both mothers and their newborns. However, pregnant women living in low- and middle-income areas or countries often fail to receive early clinical interventions at local medical facilities due to restricted availability of GDM diagnosis. The outstanding performance of artificial intelligence (AI) in disease diagnosis in previous studies demonstrates its promising applications in GDM diagnosis. Objective This study aims to investigate the implementation of a well-performing AI algorithm in GDM diagnosis in a setting, which requires fewer medical equipment and staff and to establish an app based on the AI algorithm. This study also explores possible progress if our app is widely used. Methods An AI model that included 9 algorithms was trained on 12,304 pregnant outpatients with their consent who received a test for GDM in the obstetrics and gynecology department of the First Affiliated Hospital of Jinan University, a local hospital in South China, between November 2010 and October 2017. GDM was diagnosed according to American Diabetes Association (ADA) 2011 diagnostic criteria. Age and fasting blood glucose were chosen as critical parameters. For validation, we performed k-fold cross-validation (k=5) for the internal dataset and an external validation dataset that included 1655 cases from the Prince of Wales Hospital, the affiliated teaching hospital of the Chinese University of Hong Kong, a non-local hospital. Accuracy, sensitivity, and other criteria were calculated for each algorithm. Results The areas under the receiver operating characteristic curve (AUROC) of external validation dataset for support vector machine (SVM), random forest, AdaBoost, k-nearest neighbors (kNN), naive Bayes (NB), decision tree, logistic regression (LR), eXtreme gradient boosting (XGBoost), and gradient boosting decision tree (GBDT) were 0.780, 0.657, 0.736, 0.669, 0.774, 0.614, 0.769, 0.742, and 0.757, respectively. SVM also retained high performance in other criteria. The specificity for SVM retained 100% in the external validation set with an accuracy of 88.7%. Conclusions Our prospective and multicenter study is the first clinical study that supports the GDM diagnosis for pregnant women in resource-limited areas, using only fasting blood glucose value, patients’ age, and a smartphone connected to the internet. Our study proved that SVM can achieve accurate diagnosis with less operation cost and higher efficacy. Our study (referred to as GDM-AI study, ie, the study of AI-based diagnosis of GDM) also shows our app has a promising future in improving the quality of maternal health for pregnant women, precision medicine, and long-distance medical care. We recommend future work should expand the dataset scope and replicate the process to validate the performance of the AI algorithms.

scholarly journals Spontaneous Abortions and Gestational Diabetes Mellitus- A Study from Pakistan

2021 ◽  
Vol 2 (2) ◽  
pp. 58-63
Author(s):  
Aasia Kanwal ◽  
Asma Salam ◽  
Aisha Bashir
Keyword(s):  
Diabetes Mellitus ◽  
Blood Glucose ◽  
Gestational Diabetes ◽  
Gestational Diabetes Mellitus ◽  
Pregnant Women ◽  
Fasting Blood Glucose ◽  
Categorical Variables ◽  
Spontaneous Abortions ◽  
Rural And Urban ◽  
Increased Risk

Background: Gestational diabetes mellitus leads to adverse pregnancy outcomes. Objectives: The objective of the study was to explore the relationship of spontaneous abortions with gestational diabetes mellitus in pregnant women from rural and urban Lahore. Methods: This cross-sectional study was conducted at University of Health Sciences, Lahore in 2019. Among 60 pregnant women sampled, 30 had gestational diabetes mellitus (GDM) and 30 were normal pregnant controls. Pregnant women were sampled from different hospitals of rural and urban areas of Lahore. Independent sample t-test was applied for analyzing the data. Chi- square test was used to analyze the categorical variables. Association of fasting blood glucose (FBG) and abortions was checked. Odd ratio and relative risk were calculated. Results: Mean fasting blood sugar levels were significantly higher in GDM group (105 mg/dL) as compared to non-GDM group (80.50 mg/dL) at p<0.001. The proportions of the women with increased number of abortions had significantly higher blood glucose levels (OR 5.091, 95% CI, RR 1.27). Conclusions: Gestational diabetes mellitus is associated with an increased risk of spontaneous abortions.

scholarly journals Association of WISP1/CCN4 with Risk of Overweight and Gestational Diabetes Mellitus in Chinese Pregnant Women

2020 ◽  
Vol 2020 ◽  
pp. 1-10
Author(s):  
Lei Liu ◽  
Jiajin Hu ◽  
Liu Yang ◽  
Ningning Wang ◽  
Yang Liu ◽  
...  
Keyword(s):  
Diabetes Mellitus ◽  
Blood Glucose ◽  
Gestational Diabetes ◽  
Gestational Diabetes Mellitus ◽  
Pregnant Women ◽  
Fasting Blood Glucose ◽  
Adverse Outcomes ◽  
Obese Women ◽  
Complement C1q

Background. Obese women with gestational diabetes mellitus (GDM) have a higher risk of adverse outcomes than women with obesity or GDM alone. Our study is aimed at investigating the discriminatory power of circulatory Wnt1-inducible signaling pathway protein-1 (WISP1), a novel adipocytokine, on the copresence of prepregnancy overweight/obesity and GDM and at clarifying the relationship between the WISP1 level and clinical cardiometabolic parameters. Methods. A total of 313 participants were screened from a multicenter prospective prebirth cohort: Born in Shenyang Cohort Study (BISCS). Subjects were examined with a 2×2 factorial design for body mass index BMI≥24 and GDM. Between 24 and 28 weeks of pregnancy, follow-up individuals underwent an OGTT and blood sampling for cardiometabolic characterization. Results. We observed that the WISP1 levels were elevated in prepregnancy overweight/obesity patients with GDM, compared with nonoverweight subjects with normal blood glucose (3.45±0.89 vs. 2.91±0.75 ng/mL). Multilogistic regression analyses after adjustments for potential confounding factors revealed that WISP1 was a strong and independent risk factor for prepregnancy overweight/obesity with GDM (all ORs>1). In addition, the results of the ROC analysis indicated that WISP1 exhibited the capability to identify individuals with prepregnancy overweight/obesity and GDM (all AUC>0.5). Finally, univariate and multivariate linear regression showed that WISP1 level was positively and independently correlated with fasting blood glucose, systolic blood pressure, and aspartate aminotransferase and was negatively correlated with HDL-C and complement C1q. Conclusions. WISP1 may be critical for the prediction, diagnosis, and therapeutic strategies against obesity and GDM in pregnant women.

scholarly journals Serum cobalamin status of pregnant women suffering from gestational diabetes mellitus.

2020 ◽  
Vol 27 (05) ◽  
pp. 1004-1010
Author(s):  
Saher Fatima ◽  
Sadia Saeed ◽  
Syeda Fariha Hasnny ◽  
Nathumal Maheshwari ◽  
Urooj Tabassum ◽  
...  
Keyword(s):  
Diabetes Mellitus ◽  
Blood Glucose ◽  
Glycemic Control ◽  
Gestational Diabetes ◽  
Gestational Diabetes Mellitus ◽  
Pregnant Women ◽  
Fasting Blood Glucose ◽  
Inverse Correlation ◽  
Cobalamin Deficiency ◽  
Assay Method

Objectives: Determining serum cobalamin levels in Pregnant Women suffering from Gestational Diabetes mellitus (GDM) presenting at our tertiary care hospital. Study Design: Case control study. Setting: Department of Gynecology and Medicine, SMBB Medical College Layari General Hospital Karachi. Period: January 2016 to April 2017. Material & Methods: Sample of 100 pregnant women in 2nd and 3rd trimester was selected into; 50 controls and 50 GDM cases through convenient sampling. GDM was defined as pregnant women with fasting blood sugar ≥100 mg/dL. 5 ml blood was collected; 3 ml put into EDTA tubes for complete blood counts and 2 ml for sera. Blood glucose was estimated by hexokinase method, HbA1c by colorimetric method and cobalamin by ECLIA assay method. SPSS software 21.0 (IBM, Inc USA) was used for data analysis using Student t-test and Chi-square test (P ≤ 0.05). Results: Age of control was 30.43±1.49 years and 29.95±1.27 years in cases. Gestational age was 33.67±2.69 weeks in controls and 34.75±2.53 weeks in cases. Control and cases shows serum cobalamin levels of 316.34± 113.77 pg/ml and 253.5±121.32 pg/ml respectively (P=0.009). Serum cobalamin deficiency was noted in 68% of cases and 40% of controls (P<0.05). Glycemic control was bad in majority of cases. Serum cobalamin shows inverse correlation with random blood glucose, fasting blood glucose and Glycated HbA1. Conclusion: We found low serum cobalamin levels in pregnant women suffering from gestational diabetes mellitus that showed inverse correlation with random and fasting blood glucose and glycemic control.

scholarly journals Predictive Value of First-Trimester Glycosylated Hemoglobin Levels in Gestational Diabetes Mellitus: A Chinese Population Cohort Study

2021 ◽  
Vol 2021 ◽  
pp. 1-6
Author(s):  
Jianbin Sun ◽  
Sanbao Chai ◽  
Xin Zhao ◽  
Ning Yuan ◽  
Jing Du ◽  
...  
Keyword(s):  
Diabetes Mellitus ◽  
Blood Glucose ◽  
Gestational Diabetes ◽  
Pregnant Women ◽  
Predictive Value ◽  
Statistical Difference ◽  
Glycosylated Hemoglobin ◽  
Fasting Blood Glucose ◽  
First Trimester ◽  
Hba1c Levels

This study was aimed at exploring the predictive value of first-trimester glycosylated hemoglobin (HbA1c) levels in the diagnosis of gestational diabetes mellitus (GDM). A total of 744 pregnant women registered at the Peking University International Hospital between March 2017 and March 2019 were included in this study. Data on personal characteristics and biochemical indicators of the pregnant women were collected during the first trimester. The International Association of Diabetes and Pregnancy Study Groups has adopted specific diagnostic criteria as the gold standard for the diagnosis of GDM. Receiver operating characteristic (ROC) curve statistics were used to assess the predictive value of first-trimester HbA1c levels in the diagnosis of GDM. HbA1c levels in the first trimester were significantly higher in the GDM group than in the non-GDM group ( 5.23 % ± 0.29 % vs. 5.06 ± 0.28 % , P < 0.05 ). The first-trimester HbA1c level was an independent risk factor for gestational diabetes. The area under the ROC curve (AUC) of HbA1c for GDM was 0.655 (95% confidence interval 0.620-0.689, P < 0.001 ). The positive likelihood ratio was the highest at HbA 1 c = 5.9 % , sensitivity was 2.78, and specificity was 99.83%. There was no statistical difference in AUC between fasting blood glucose and HbA1c ( P = 0.407 ). First-trimester HbA1c levels can be used to predict GDM. The risk of GDM was significantly increased in pregnant women with first ‐ trimester   HbA 1 c   levels > 5.9 % . There was no statistical difference between first-trimester HbA1c and fasting blood glucose levels in predicting GDM.

scholarly journals GESTATIONAL DIABETES MELLITUS

2012 ◽  
Vol 19 (04) ◽  
pp. 462-468
Author(s):  
M. IKRAM ◽  
SYED HAIDER HASAN ALAM ◽  
SHAFQAT MUKHTAR ◽  
M. Saeed
Keyword(s):  
Diabetes Mellitus ◽  
Blood Glucose ◽  
Gestational Diabetes ◽  
Glucose Tolerance ◽  
Gestational Diabetes Mellitus ◽  
Glucose Tolerance Test ◽  
Fasting Blood Glucose ◽  
Tolerance Test ◽  
Oral Glucose ◽  
Oral Glucose Tolerance

Introduction: Gestational diabetes mellitus is common disorder in pregnancy. It is associated with adverse pregnancy outcome. There is no consensus regarding the optimal approach to screening of gestational diabetes mellitus. The present study has tried toobserve the value of fasting blood glucose in screening of gestational diabetes. Objective: To determine the frequency of patients in whomfasting blood glucose and 100gm glucose tolerance show agreement for screening of gestational diabetes mellitus at 24 -28 wks. Studydesign: Comparative cross sectional study. Settings: The study was conducted at Gynecology and Obstetrics department Shaikh ZayedFederal Post Graduate Institute Lahore. Duration of study with dates: 6 months from 12Nov 2010 to 11 May 2011. Material and method: Thestudy included 135 booked patients with positive family history of diabetes mellitus. All patients underwent fasting blood glucose at 24-28 weeksof gestation, regardless of results of fasting blood glucose on next visit they underwent 100g oral glucose tolerance test (OGTT). The agreementbetween fasting blood glucose and 100g oral glucose tolerance test was calculated in frequency and percentages. Results: The mean age ofwomen in studied population was 27.15±3.70.Out of 135 patients 86.7 %( 117) showed agreement between results of fasting blood glucose and100g OGTT while 13.31 %( 18) showed no agreement between both of the tests. Conclusions: Fasting blood glucose is a good screeningoption for gestational diabetes mellitus along with positive history. It provides a simple, cheap and more practical test for screening of gestationaldiabetes mellitus. However diagnostic confirmation with 100g OGTT should be done.

scholarly journals The Role of Untimed Blood Glucose in Screening for Gestational Diabetes Mellitus in a High Prevalent Diabetic Population

Scientifica ◽  
2016 ◽  
Vol 2016 ◽  
pp. 1-6 ◽  
Author(s):  
Sarah Cuschieri ◽  
Johann Craus ◽  
Charles Savona-Ventura
Keyword(s):  
Blood Glucose ◽  
Gestational Diabetes ◽  
Pregnant Women ◽  
Fasting Blood Glucose ◽  
Diabetes Type 2 ◽  
Obstetric Outcomes ◽  
Oral Glucose ◽  
Stepwise Logistic Regression ◽  
Operating Characteristics ◽  
Screening Protocol

Global prevalence increase of diabetes type 2 and gestational diabetes (GDM) has led to increased awareness and screening of pregnant women for GDM. Ideally screening for GDM should be done by an oral glucose tolerance test (oGTT), which is laborious and time consuming. A randomized glucose test incorporated with anthropomorphic characteristics may be an appropriate cost-effective combined clinical and biochemical screening protocol for clinical practice as well as cutting down on oGTTs. A retrospective observational study was performed on a randomized sample of pregnant women who required an OGTT during their pregnancy. Biochemical and anthropomorphic data along with obstetric outcomes were statistically analyzed. Backward stepwise logistic regression and receiver operating characteristics curves were used to obtain a suitable predictor for GDM without an oGTT and formulate a screening protocol. Significant GDM predictive variables were fasting blood glucose (p=0.0001) and random blood glucose (p=0.012). Different RBG and FBG cutoff points with anthropomorphic characteristics were compared to carbohydrate metabolic status to diagnose GDM without oGTT, leading to a screening protocol. A screening protocol incorporating IADPSG diagnostic criteria, BMI, and different RBG and FBG criteria would help predict GDM among high-risk populations earlier and reduce the need for oGTT test.

scholarly journals The effectiveness of Information-Motivation and Behavioral skills (IMB) model of self-care in early pregnancy to prevent gestational diabetes mellitus in Iranian overweight and obese women: A Randomized Controlled Trial

2020 ◽  
Author(s):  
Narges Sadat Motahari-Tabari ◽  
Mahbobeh Faramarzi ◽  
Marjan Ahmad Shirvani ◽  
Afsaneh Bakhtiari ◽  
Shabnam Omidvar ◽  
...  
Keyword(s):  
Blood Glucose ◽  
Gestational Diabetes ◽  
Pregnant Women ◽  
Controlled Trial ◽  
Fasting Blood Glucose ◽  
Intervention Group ◽  
Tolerance Test ◽  
Obese Women ◽  
Imb Model ◽  
Behavioral Skills

Abstract Background: Gestational diabetes is one of the most common metabolic dysfunction in pregnancy and as overweight and obesity are of the major risk factors, this study aimed to determine the effect of Information-Motivation and Behavioral skills (IMB) model-based counseling on preventing gestational diabetes in overweight and obese pregnant women. Methods: A randomised controlled trial (RCT) was conducted involving pregnant women who are overweight (BMI >25 to 29.9 k/gm2) or obese (BMI >30 k/gm2), at the 12 to 16 weeks gestation and recruited from the Prenatal Clinic of Rohani Educational Hospital in Babol medical university, Iran, women in the intervention group will receive a program informed four sessions by the Information-Motivation and Behavioral skills. The control group received routine care. Blood glucose was measured before and 8 weeks after the intervention. Descriptive and inferential statistics including mean, standard deviation, frequency, t-test, chi-square and ANCOVA were used. Results: The prevalence of gestational diabetes was lower in the intervention group than the routine care group (10% and 29.9%, respectively, RR=0.33, CI 95% (0.15- 0.74) p=0.004) as well as mean fasting blood glucose (d=0.28, P=0.07), and glucose tolerance test at the first and second hour (d= 0.41 and d=0.73, respectively, p<0.01). Conclusions: Our data suggest that women that IMB model-based counseling on self-care in early pregnancy in overweight and obese pregnant women can be effective in preventing gestational diabetes. Keywords: Gestational Diabetes, Obesity, Overweight, Information-Motivation-Behavioral Skills Model Name of the registry: Comparison of the effectiveness of counseling based on health promoting behaviors on fasting blood glucose and glucose tolerance test in pregnant and overweight and obese women IRCT registration number: IRCT20120125008822N3Registration date: 2018-07-05, 1397/04/14Registration timing: prospective https://en.irct.ir/trial/32150

Sign in / Sign up

Export Citation Format

Share Document