Background: Synthetic electronic health records (EHRs) generated by large language models (LLMs) offer potential for clinical education and model training while addressing privacy concerns. However, performance variations and demographic biases in these models remain underexplored, posing risks to equitable health care. Objective: This study aimed to systematically assess the performance of various LLMs in generating synthetic EHRs and to critically evaluate the presence of gender and racial biases in the generated outputs. We focused on assessing the completeness and representativeness of these EHRs across 20 diseases with varying demographic prevalence. Methods: A framework was developed to generate 140,000 synthetic EHRs using 10 standardized prompts across 7 LLMs. The electronic health record performance score (EPS) was introduced to quantify completeness, while the statistical parity difference (SPD) was proposed to assess the degree and direction of demographic bias. Chi-square tests were used to evaluate the presence of bias across demographic groups. Results: Larger models exhibited superior performance but heightened biases. The Yi-34B achieved the highest EPS (96.8), while smaller models (Qwen-1.8B: EPS=63.35) underperformed. Sex polarization emerged: female-dominated diseases (eg, multiple sclerosis) saw amplified female representation in outputs (Qwen-14B: 973/1000, 97.3% female vs 564,424/744,778, 75.78% real; SPD=+21.50%), while balanced diseases and male-dominated diseases skewed the male group (eg, hypertension Llama 2-13 B: 957/1000, 95.7% male vs 79,540,040/152,466,669, 52.17% real; SPD=+43.50%). Racial bias patterns revealed that some models overestimated the representation of White (eg, Yi-6B: mean SPD +14.40%, SD 16.22%) or Black groups (eg, Yi-34B: mean SPD +14.90%, SD 27.16%), while most models systematically underestimated the representation of Hispanic (average SPD across 7 models is –11.93%, SD 8.36%) and Asian groups (average SPD across 7 models is –0.77%, SD 11.99%). Conclusions: Larger models, such as Yi-34B, Qwen-14B, and Llama 2 to 13 B, showed improved performance in generating more comprehensive EHRs, as reflected in higher EPS values. However, this increased performance was accompanied by a notable escalation in both gender and racial biases, highlighting a performance-bias trade-off. The study identified 4 key findings as follows: (1) as model size increased, EHR generation improved, but demographic biases also became more pronounced; (2) biases were observed across all models, not just the larger ones; (3) gender bias closely aligned with real-world disease prevalence, while racial bias was evident in only a subset of diseases; and (4) racial biases varied, with some diseases showing overrepresentation of White or Black populations and underrepresentation of Hispanic and Asian groups. These findings underline the need for effective bias mitigation strategies and the development of benchmarks to ensure fairness in artificial intelligence applications for health care.

Background: Pioglitazone may have a protective effect against cardiovascular disease risk among type 2 diabetes (T2D) patients, but evidence from China is lacking. This study investigated the association using electronic health records (EHR) data from a Chinese cohort of T2D patients.Methods: All T2D patients from the First Affiliated Hospital of Nanjing Medical University who were prescribed at least one oral antidiabetic drug and were aged ≥18 years between 1 July 2005 and 30 June 2017 were eligible for inclusion (n = 71 783). Pioglitazone use was determined in 6‐month study intervals, with outcome events of myocardial infarction (MI), ischemic stroke, and heart failure. Poisson regression was used to estimate adjusted rate ratios (RRs) with 95% confidence intervals (CIs).Results: In multivariable analysis adjusted for potential confounders, pioglitazone use, compared with no use, was associated with a significant 39% decreased risk of MI (RR = 0.61; 95% CI = 0.42‐0.90; P = 0.012). Pioglitazone use was also associated with a non‐significant reduction in risk of heart failure or stroke. When MI, heart failure, and stroke were combined as a composite outcome, pioglitazone use was associated with a 30% decrease in risk (RR = 0.70; 95% CI = 0.56‐0.88; P = 0.002).Conclusions: This study demonstrates that applying informatics tools to a large EHR database could be a good way to efficiently conduct clinical observational research. In addition, the findings validated the favorable effect of pioglitazone on the risk of MI among T2D patients in China, with the use of pioglitazone decreasing the risk of MI among those with T2D.

DNA methylation plays a crucial role in human diseases pathogenesis. Substantial experimental evidence from clinical and biological studies has confirmed numerous methylation-disease associations, which provide valuable prior knowledge for advancing precision medicine through biomarker discovery and disease subtyping. To systematically mine reliable methylation prior knowledge from known DNA methylation-disease associations and develop robust computational methods for precision medicine applications, we propose MethPriorGCN. By integrating layer attention mechanisms and feature weighting mechanisms, MethPriorGCN not only identified reliable methylation digital biomarkers but also achieved superior disease subtype classification accuracy.

Analysis of the related risks of disease provides a scientific basis for disease prevention and treatment, hospital management, and policy formulation by the changes in disease spectrum of patients in hospital. Retrospective analysis was made to the first diagnosis, age, gender, daily average cost of hospitalized patients, and other factors in the First Affiliated Hospital of Nanjing Medical University during 2006–2013. The top 4 cases were as follows: cardiovascular disease, malignant tumors, lung infections, and noninsulin dependent diabetes mellitus. By the age of disease analysis, we found a younger age trend of cardiovascular disease, and the age of onset of cancer or diabetes was somewhat postponed. The average daily cost of hospitalization and the average daily cost of the main noncommunicable diseases were both on the rise. Noncommunicable diseases occupy an increasingly important position in the constitution of the disease, and they caused an increasing medical burden. People should pay attention to health from the aspects of lifestyle changing. Hospitals should focus on building the appropriate discipline. On the other hand, an integrated government response is required to tackle key risks. Multiple interventions are needed to lower the burden of these diseases and to improve national health.

Background: Patients with type 2 diabetes (T2D) and hypertension are at increased risk of adverse cardiovascular (CV) events. However, real-world evidence comparing the CV effectiveness and safety of major hypoglycemic drug classes remains limited in this population. This multicenter pooled analysis aims to directly compare the CV outcomes and safety profiles of these key agents in patients with T2D and hypertension. Methods: We analyzed electronic health records from two databases in a cohort study of T2D patients with hypertension who had initiated metformin as first-line therapy. Propensity score matching (PSM) and Cox proportional hazards models were used to compare the risks of 3-/4-point major adverse cardiovascular events (MACE) and safety outcomes across drug classes added to metformin: insulin, sulfonylureas (SUs), glucagon-like peptide-1 receptor agonists (GLP-1 RAs), dipeptidyl peptidase-4 inhibitors (DPP4is), glinides, acarbose, and sodium-glucose transporter 2 inhibitors (SGLT2is). Results: Compared with insulin, GLP-1 RAs, DPP4is, and glinides were associated with a lower risk of 3-point MACE (HR: 0.48 [0.31–0.76], 0.70 [0.57–0.85], and 0.70 [0.52–0.94], respectively). SUs were associated with a higher risk of 3-point MACE compared with DPP4is (HR: 1.30 [1.06–1.59]). DPP4is, GLP-1 RAs, and glinides showed a lower risk of 3-point MACE compared with acarbose (HR: 0.62 [0.51–0.76], 0.47 [0.29–0.75], and 0.59 [0.43–0.81], respectively). Similar patterns were observed for 4-point MACE. For safety outcomes, DPP4is were associated with a reduced risk of chronic kidney disease, while insulin use was associated with reduced risks of inflammatory polyarthritis and insomnia. However, DPP4is were associated with higher risks of coronary atherosclerotic diseases and hypertensive heart disease. Conclusions: This study highlights the differential cardiovascular effectiveness and safety profiles of hypoglycemic therapies in real-world settings, providing valuable insights for optimizing T2D management, particularly in patients with comorbid hypertension.