Machine Learning Models in Type 2 Diabetes Risk Prediction: Results from a Cross-sectional Retrospective Study in Chinese Adults.

Type 2 diabetes mellitus (T2DM) has become a prevalent health problem in China, especially in urban areas. Early prevention strategies are needed to reduce the associated mortality and morbidity. We applied the combination of rules and different machine learning techniques to assess the risk of development of T2DM in an urban Chinese adult population. A retrospective analysis was performed on 8000 people with non-diabetes and 3845 people with T2DM in Nanjing. Multilayer Perceptron (MLP), AdaBoost (AD), Trees Random Forest (TRF), Support Vector Machine (SVM), and Gradient Tree Boosting (GTB) machine learning techniques with 10 cross validation methods were used with the proposed model for the prediction of the risk of development of T2DM. The performance of these models was evaluated with accuracy, precision, sensitivity, specificity, and area under receiver operating characteristic (ROC) curve (AUC). After comparison, the prediction accuracy of the different five machine models was 0.87, 0.86, 0.86, 0.86 and 0.86 respectively. The combination model using the same voting weight of each component was built on T2DM, which was performed better than individual models. The findings indicate that, combining machine learning models could provide an accurate assessment model for T2DM risk prediction.

Comparison of Glycemic Variability in Chinese T2DM Patients Treated with Exenatide or Insulin Glargine: A Randomized Controlled Trial.

INTRODUCTION: Increasing the frequency of blood glucose monitoring aids the evaluation of glycemic variability and blood glucose control by antidiabetic drugs. It remains unclear, however, whether GLP-1 receptor agonists or basal insulin has a better effect on glycemic variability in type 2 diabetes mellitus (T2DM) patients who are inadequately controlled by metformin. We used a continuous glucose monitoring system (CGMS) to compare patients on a GLP-1 receptor agonist with patients on basal insulin in terms of glycemic variability. METHODS: This prospective randomized study assigned T2DM patients treated with metformin (N = 39) to either exenatide treatment or insulin glargine treatment for 16 weeks. Glycemic variability was assessed using a CGMS; hemoglobin A1c (HbA1c), ß-cell function, weight, body mass index (BMI), and waist circumference were also evaluated. RESULTS: Mean blood glucose level, continuous overlapping net glycemic action, mean amplitude of glycemic excursions, percentage of the time that the blood glucose value was > 10.0 mmol/L, and highest blood glucose level (P  < 0.01-0.05) significantly decreased in both groups. Standard deviation of the mean glucose value, largest amplitude of glycemic excursions, and waist circumference significantly decreased for those treated with exenatide (P  < 0.05), while no changes were observed with insulin glargine treatment. Percentage of the time that the blood glucose value was > 7.8 mmol/L decreased after insulin glargine use (P  < 0.05) but not with the exenatide intervention. Similar decreases in fasting blood glucose and HbA1c and increases in the 1/homeostasis model assessment of insulin resistance, disposition index 30, and disposition index 120 were observed in both groups (P  < 0.01-0.05). Reductions in weight and BMI were greater with exenatide than with insulin glargine treatment (P  < 0.05). CONCLUSIONS: In overweight and obese patients with T2DM inadequately controlled by metformin, exenatide and insulin glargine have similar efficacies in terms of glycemic variability, HbA1c alleviation, and ß-cell function, but exenatide has a greater effect on body weight and BMI.

Retrospective Examination of Q Fever Endocarditis: An Underdiagnosed Disease in the Mainland of China.

BACKGROUND: Q fever endocarditis, a chronic illness caused by Coxiella burnetii, can be fatal if misdiagnosed or left untreated. Despite a relatively high positive rate of Q fever serology in healthy individuals in the mainland of China, very few cases of Q fever endocarditis have been reported. This study summarized cases of Q fever endocarditis among blood culture negative endocarditis (BCNE) patients and discussed factors attributing to the low diagnostic rate. METHODS: We identified confirmed cases of Q fever endocarditis among 637 consecutive patients with infective endocarditis (IE) in the Peking Union Medical College Hospital between 2006 and 2016. The clinical findings for each confirmed case were recorded. BCNE patients were also examined and each BCNE patient's Q fever risk factors were identified. The risk factors and presence of Q fever serologic testing between BCNE patients suspected and unsuspected of Q fever were compared using the Chi-squared or Chi-squared with Yates' correction for continuity. RESULTS: Among the IE patients examined, there were 147 BCNE patients, of whom only 11 patients (7.5%) were suspected of Q fever and undergone serological testing for C. burnetii. Six out of 11 suspected cases were diagnosed as Q fever endocarditis. For the remaining136 BCNE patients, none of them was suspected of Q fever nor underwent relevant testing. Risk factors for Q fever endocarditis were comparable between suspected and unsuspected patients, with the most common risk factors being valvulopathy in both groups. However, significantly more patients had consulted the Infectious Diseases Division and undergone comprehensive diagnostic tests in the suspected group than the unsuspected group (100% vs. 63%, P= 0.03). CONCLUSIONS: Q fever endocarditis is a serious yet treatable condition. Lacking awareness of the disease may prevent BCNE patients from being identified, despite having Q fever risk factors. Increasing awareness and guideline adherence are crucial in avoiding misdiagnosing and missed diagnosing of the disease.