Early identification of individuals at high risk of diabetes is crucial for implementing early intervention strategies. However, algorithms specific to elderly Chinese adults are lacking. The aim of this study is to build effective prediction models based on machine learning (ML) for the risk of type 2 diabetes mellitus (T2DM) in Chinese elderly. A retrospective cohort study was conducted using the health screening data of adults older than 65 years in Wuhan, China from 2018 to 2020. With a strict data filtration, 127,031 records from the eligible participants were utilized. Overall, 8298 participants were diagnosed with incident T2DM during the 2-year follow-up (2019–2020). The dataset was randomly split into training set (n = 101,625) and test set (n = 25,406). We developed prediction models based on four ML algorithms: logistic regression (LR), decision tree (DT), random forest (RF), and extreme gradient boosting (XGBoost). Using LASSO regression, 21 prediction features were selected. The Random under-sampling (RUS) was applied to address the class imbalance, and the Shapley Additive Explanations (SHAP) was used to calculate and visualize feature importance. Model performance was evaluated by the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and accuracy. The XGBoost model achieved the best performance (AUC = 0.7805, sensitivity = 0.6452, specificity = 0.7577, accuracy = 0.7503). Fasting plasma glucose (FPG), education, exercise, gender, and waist circumference (WC) were the top five important predictors. This study showed that XGBoost model can be applied to screen individuals at high risk of T2DM in the early phrase, which has the strong potential for intelligent prevention and control of diabetes. The key features could also be useful for developing targeted diabetes prevention interventions.

Identifying people with a high risk of developing diabetes among those with prediabetes may facilitate the implementation of a targeted lifestyle and pharmacological interventions. We aimed to establish machine learning models based on demographic and clinical characteristics to predict the risk of incident diabetes. We used data from the free medical examination service project for elderly people who were 65 years or older to develop logistic regression (LR), decision tree (DT), random forest (RF), and extreme gradient boosting (XGBoost) machine learning models for the follow-up results of 2019 and 2020 and performed internal validation. The receiver operating characteristic (ROC), sensitivity, specificity, accuracy, and F1 score were used to select the model with better performance. The average annual progression rate to diabetes in prediabetic elderly people was 14.21%. Each model was trained using eight features and one outcome variable from 9607 prediabetic individuals, and the performance of the models was assessed in 2402 prediabetes patients. The predictive ability of four models in the first year was better than in the second year. The XGBoost model performed relatively efficiently (ROC: 0.6742 for 2019 and 0.6707 for 2020). We established and compared four machine learning models to predict the risk of progression from prediabetes to diabetes. Although there was little difference in the performance of the four models, the XGBoost model had a relatively good ROC value, which might perform well in future exploration in this field.

The relationship between hyperuricemia and chronic kidney disease (CKD) has been investigated extensively. However, studies on elderly individuals are still limited. Moreover, there is no consensus on whether hyperuricemia or elevated serum uric acid (SUA) within the normal range is correlated with the new onset of CKD and whether there are differences between males and females. We included 39039 elderly diabetic patients without CKD at baseline from a community-based cohort in Wuhan, China. The outcome event was the new onset of CKD (defined as an estimated glomerular filtration rate < 60 mL/min/1.73 m²). Multivariate Cox models were used to assess the adjusted hazard ratio (HR). During the 2-year follow-up period, 3162 (8.10%) patients with diabetes developed new-onset CKD. The optimal cutoff value of SUA for incident CKD was 347.4 μmol/L. The adjusted HRs of hyperuricemia for new-onset CKD were 1.925 (1.724-2.150) and 1.676 (1.520-1.848) for males and females, respectively. The risk of developing CKD increased across the Q4 group up to 2.242 times for their counterparts in the lowest SUA quartile, independent of age, sex, diabetes duration, obesity, hypertension, systolic blood pressure, diastolic blood pressure, smoking, drinking, dyslipidemia, triglyceride, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and fasting plasma glucose. Hyperuricemia is an independent predictor of incident CKD. Elevated SUA was linearly correlated with CKD in elderly patients with diabetes, showing a relatively higher intensity among males compared with that among females. The optimal cutoff value of SUA for the risk of new-onset CKD in elderly patients with diabetes was 347.4 μmol/L.

Background: Chronic obstructive pulmonary disease (COPD) has been an important public health issue in China. This study aimed to analyze the temporal trends in mortality and years of life lost (YLL) from COPD, and explore the effects of age, period, and cohort in Wuhan, China from 2010 to 2019. Methods: Data were collected from the cause of death surveillance system in Wuhan. Age-standardized mortality rate (ASMR), age-standardized YLL rate (ASYR) and the estimated annual percent changes (EAPC) were calculated to evaluate the temporal trends. The age-period-cohort (APC) model was adopted to estimate the age, period, and cohort effects. Results: From 2010 to 2019, COPD accounted for 26,051.15 deaths and 394,659.58 person years YLL in Wuhan. Recently, the death burden of COPD in Wuhan has somewhat improved, especially after 2015, with declining trends in ASMR and ASYR. Additionally, the ASMR and ASYR of COPD was higher in males. And these of males showed the overall upward trends, with EAPCs of 1.06 (0.13, 2.00) and 1.21 (0.12, 2.31), respectively, while females showed downward trends since 2010. According to APC model, the age effect of COPD increased with age, and the cohort risk ratios (RRs) followed the overall downward trends. Period RRs for the Wuhan population generally tended to rise and then fall, with females showing a clear downward trend after 2015, while period RRs for males maintained an upward trend throughout the study period. Conclusions: Recently, the death burden from COPD in Wuhan has improved, especially after 2015, with improvements in ASMR, ASYR and period RRs. Sex differences still exist. COPD posed a greater threat to the elderly, especially males. Public health managers should continue to execute more targeted programs to lessen the death burden of COPD in Wuhan.

Background: Brain cancer is one of the worst types of cancer worldwide. Understanding the epidemiology of CNS cancer is critical for properly allocating healthcare resources. Methods: We collected data on CNS cancer deaths in Wuhan, China, during 2010–2019. We constructed the cause-eliminated life tables to calculate life expectancy (LE), mortality, and years of life lost (YLLs) by age and sex. The BAPC model was used to forecast the future trends of age-standardized mortality rate (ASMR). Decomposition analysis was adopted to explore the contribution of population growth, population aging, and age-specific mortality to the change in total CNS cancer deaths. Results: In 2019, the ASMR of CNS cancer was 3.75, and the ASYR was 135.70 in Wuhan, China. ASMR was expected to decrease to 3.43 in 2024. The age distribution of deaths due to CNS cancer was concentrated in the middle-aged and older population, with a peak in the 65–69 age group. Caidian, Jianghan, and Qingshan had the greatest ASMRs in 2019 in Wuhan, with ASMRs of 6.32, 4.78, and 4.75, respectively. Population aging is critical to the change in total CNS cancer deaths. Conclusion: We analyzed the current status, temporal trends, and gender and age distributions of the burden of CNS cancer in Wuhan, during 2010–2019, providing a valuable reference for better lessening the CNS cancer burden.

Background: Liver cancer has caused a heavy burden worldwide. This study aimed to estimate the trends in the mortality and years of life lost (YLL) due to liver cancer and decompose the total deaths into three contributors: population growth, population aging, and mortality change. Methods: Our study used data from the cause-of-death surveillance system in Wuhan. The mortality and YLL rates were standardized according to the sixth national population census in China. This study calculated the estimated annual percentage change (EAPC) to estimate the trends in the age-standardized mortality rate (ASMR) and age-standardized YLL rate (ASYR). Meanwhile, a decomposition analysis was used to explore the effect of population growth, population aging, and age-specific mortality change on the change in liver cancer deaths. Results: The ASMR of liver cancer declined at an annual rate of 4.6% from 30.87 per 100,000 people in 2010 to 20.29 per 100,000 people in 2019, while the ASYR was at an annual rate of 5.6% from 969.35 per 100,000 people in 2010 to 581.82 per 100,000 people in 2019. Similar downward trends were seen in men and women. The decomposition analysis found that total deaths number changed by −12.42% from 2010 to 2019, of which population growth and population aging caused the total death numbers to increase by 9.75% and 21.15%, while the age-specific mortality change caused the total death numbers to decrease by 43.32%. Conclusion: Although the ASMR of liver cancer has declined in recent years in Wuhan, it still causes a heavy burden with the increasing population and rapid population aging and remains an essential public health issue. The government should take measures to reduce the burden of liver cancer, especially among men.