Heart failure (HF) is a complex and heterogeneous syndrome caused by diverse factors, such as atrial fibrillation, diabetes, and pulmonary hypertension. The intricate pathophysiology of HF, coupled with variability in patient demographics and presentations, poses significant challenges to the effectiveness of existing deep learning models in HF risk prediction. In this paper, we propose a novel deep neural network, PDSNet, which leverages a new dual patient-disease spatial similarity strategy to improve HF risk prediction using short electronic health records. First, we develop ontology graphs to capture hierarchical relationships between patients based on HF-related symptoms and causes; Then, a bipartite graph model is utilised to learn spatial similarities between patients with similar hospital visit histories; Finally, we design a transformer-based architecture to integrate both temporal and spatial dynamics for predicting future hospital visits associated with HF risk. We benchmarked the PDSNet on predicting HF risk for 7,346 patients from the MIMIC-III dataset. Compared to seven state-of-the-art deep learning methods, our PDSNet model achieved improvements of 2%-12% in the area under the curve (AUC) score and 3%-18% in the F1 score. These findings highlight the promising potential of our proposed PDSNet to provide accurate and robust HF risk predictions, paving the way for efficient clinical decision support and personalized HF management.

Over the past few decades, osmotherapy has commonly been employed to reduce intracranial pressure in post-stroke oedema. However, evaluating the effectiveness of osmotherapy has been challenging due to the difficulties in clinical intracranial pressure measurement. As a result, there are no established guidelines regarding the selection of administration protocol parameters. Considering that the infusion of osmotic agents can also give rise to various side effects, the effectiveness of osmotherapy has remained a subject of debate. In previous studies, we proposed the first mathematical model for the investigation of osmotherapy and validated the model with clinical intracranial pressure data. The physiological parameters vary among patients and such variations can result in the failure of osmotherapy. Here, we propose an AI-assisted in silico trial for further investigation of the optimisation of administration protocols. The proposed deep neural network predicts intracranial pressure evolution over osmotherapy episodes. The effects of the parameters and the choice of dose of osmotic agents are investigated using the model. In addition, clinical stratifications of patients are related to a brain model for the first time for the optimisation of treatment of different patient groups. This provides an alternative approach to tackle clinical challenges with in silico trials supported by both mathematical/physical laws and patient-specific biomedical information.

Dear Editor, We thank Dr. Ser for her detailed reading of our study,1 and starting this important dialogue on the importance of dementia in our aging global population. We fully recognize her point that the use of ICD-9 coding for determining a diagnosis of dementia will lead to an underdiagnosis. Indeed, the cited studies from Canada and Australia illustrate the problem of undercoding specifically on dementia. It is fair to assume that there would be a similar issue for the local data from China. We were not able to identify studies investigating the reliability of coding specifically for dementia in our locality. Nevertheless, based on a local study that calculated the trends of prevalence of dementia in the Hong Kong city based on census data,2 an estimated 100 000 people above the age of 60 suffers from dementia in 2009 and it is projected to increase to approximately 330 000 patients in 2039, which is around 4.7% of the Hong Kong population similar to the incidence rate of 4.74% calculated in our study.

INTRODUCTION: The present study examined the gender-specific prognostic value of blood pressure (BP) and its variability in the prediction of dementia risk and developed a score system for risk stratification. MATERIALS AND METHODS: This was a retrospective, observational population-based cohort study of patients admitted to government-funded family medicine clinics in Hong Kong between January 1, 2000 and March 31, 2002 with at least 3 blood pressure measurements. Gender-specific risk scores for dementia were developed and tested. RESULTS: The study consisted of 74 855 patients, of whom 3550 patients (incidence rate: 4.74%) developed dementia over a median follow-up of 112 months (IQR= [59.8-168]). Nonlinear associations between diastolic/systolic BP measurements and the time to dementia presentation were identified. Gender-specific dichotomized clinical scores were developed for males (age, hypertension, diastolic and systolic BP and their measures of variability) and females (age, prior cardiovascular, respiratory, gastrointestinal diseases, diabetes mellitus, hypertension, stroke, mean corpuscular volume, monocyte, neutrophil, urea, creatinine, diastolic and systolic BP and their measures of variability). They showed high predictive strengths for both male (hazard ratio [HR]: 12.83, 95% confidence interval [CI]: 11.15-14.33, P value < .0001) and female patients (HR: 26.56, 95% CI: 14.44-32.86, P value < .0001). The constructed gender-specific scores outperformed the simplified systems without considering BP variability (C-statistic: 0.91 vs 0.82), demonstrating the importance of BP variability in dementia development. CONCLUSION: Gender-specific clinical risk scores incorporating BP variability can accurately predict incident dementia and can be applied clinically for early disease detection and optimized patient management.

In this paper, a concept of fuzzy semientropy is proposed to quantify the downside uncertainty. Several properties of fuzzy semientropy are identified and interpreted. By quantifying the downside risk with the use of semientropy, two mean-semi-entropy portfolio selection models are formulated, and a fuzzy simulation-based genetic algorithm is designed to solve the models to optimality. We carry out comparative analyses among the fuzzy mean-entropy models and the fuzzy mean-semi-entropy models and demonstrate that the mean-semi-entropy models can significantly improve the dispersion of investment. Several illustrative examples using stock dataset from the real-world financial market (China Shanghai Stock Exchange) also show the effectiveness of the models.