Fusion of sequential visits and medical ontology for mortality prediction.

The goal of mortality prediction task is to predict the future death risk of patients according to their previous Electronic Healthcare Records (EHR). The main challenge of mortality prediction is how to design an accurate and robust predictive model with sequential, multivariate, sparse and irregular EHR data. In addition, the performance of model may be affected by lack of sufficient information of some patients with rare diseases in EHRs. To address these challenges, we propose a model to fuse Sequential visits and Medical Ontology to predict patients' death risk. SeMO not only learns reasonable embeddings for medical concepts from sequential and irregular visits, but also exploits medical ontology to improve the prediction performance. With integration of multivariate features, SeMO learns robust representations of medical codes, mitigating data insufficiency and insightful sequential dependencies among patient's visits. Experimental results on real world datasets prove that the proposed SeMO improves the prediction performance compared with the baseline approaches. Our model achieves an precision of up to 0.975. Compared with RNN, the precision has been improved up to 2.204%.

Prediction and optimization of regional land-use patterns considering nonpoint-source pollution control under conditions of uncertainty.

Socioeconomic development, leading to significant changes in land-use patterns, has further influenced the output of regional nonpoint-source (NPS) pollution. Multiple uncertainties exist in the processes of land-use changes and NPS pollution export. These uncertainties can deeply affect the management of regional land-use patterns and control of NPS pollution. In this research, an integrated land-use prediction and optimization (ILUPO) model based on system dynamics, export coefficient, interval linear programming, and fuzzy parameter programming models was proposed. The ILUPO model can provide future land-use patterns and NPS pollution loads, and also help optimize the patterns under multiple pollution reduction scenarios. Interval and fuzzy uncertainties in the processes of land-use changes and NPS pollution output can be effectively addressed. The developed model was applied to a water source area in the central part of northern Guangdong Province in South China. For the prediction period 2020-2030 under the high-speed development scenario, results show that cropland area would decrease, while grassland and waterbody areas would increase. In contrast, these three types of land-use would show opposite variation trends under the low-speed development scenario. Construction land area would decrease, while forestland area would increase under both the low-speed and high-speed development scenarios. Variation of the predicted land-use patterns would lead to an increase of total nitrogen loads under each of the scenario, while the total phosphorus loads would show relatively complex variation trends. Regional land-use patterns should be further optimized to mitigate NPS pollution. However, the pollution loads in the study area cannot be reduced by >5% through land-use adjustment. Because cropland would still be the critical source of NPS pollution after optimization, strictly controlling the areas of cropland would be important for the management of such pollution in the research area. In addition, certain areas of grassland and waterbody would need to be converted into cropland and construction land to balance the economic benefit of the system and NPS pollution control. Multiple results obtained from the model under different scenarios of pollution reduction targets and α-cut levels can provide decision-making supports for the local policy makers. The developed ILUPO model can yield insights useful for the planning and adjustment of regional land-use patterns while considering NPS pollution control under conditions of uncertainty.

Management optimization of nonpoint source pollution considering the risk of exceeding criteria under uncertainty.

Management of nonpoint source (NPS) pollution is highly important in watershed water environmental and ecological security. However, the many complexities and uncertainties that exist in the processes of export and management of NPS pollution exert substantial influences on the reliability of multiple management practices. This study developed an inexact multiobjective possibilistic mean-variance mixed-integer programming (IMPMMP) model for NPS pollution management through optimization of watershed land use pattern and livestock production structure. By coupling interval parameter programming, mixed-integer programming, multiobjective programming, and an export coefficient model within a general possibilistic mean-variance model framework, the IMPMMP model deals effectively with system uncertainties and complexities. Moreover, the risk of exceeding criteria (REC) in NPS pollution management systems can be considered. The proposed IMPMMP model was applied to a real-world case study in the Xinfengjiang Reservoir watershed in South China. Results showed that the preference of decision makers regarding land use adjustment plays a decisive role in determining model feasibility. The area provided for each land use type that could be adjusted has to reach a certain threshold to achieve the goals of reduced pollution load and REC control. The NPS pollution loads after optimization would be exported primarily from different land uses and the human population. Compared with NPS nitrogen pollution management, it is more difficult to reduce the NPS phosphorus load and to manage the corresponding REC through adjustment of the land use pattern and livestock production structure. Moreover, it is difficult to simultaneously reduce the NPS nitrogen and phosphorus pollution loads and REC in each subbasin. The model, which can provide policy makers with a series of schemes for optimization of land use pattern and livestock production structure, has satisfactory applicability and could be used for watershed NPS pollution management.

Screening and Identification of Potential Hub Genes in Myocardial Infarction Through Bioinformatics Analysis.

BACKGROUND: Myocardial infarction (MI) is a common cause of death worldwide. It is characterized by coronary artery occlusion that causes ischemia and hypoxia of myocardial cells, leading to irreversible myocardial damage. MATERIALS AND METHODS: To explore potential targets for treatment of MI, we reorganized and analyzed two microarray datasets (GSE4648 and GSE775). The GEO2R tool was used to screen for differentially expressed genes (DEGs) between infarcted and normal myocardium. We used the Database for Annotation, Visualization and Integrated Discovery (DAVID) to perform Gene Ontology functional annotation analysis (GO analysis) and the Kyoto Encyclopedia of Genes and Genomes for pathway enrichment analysis (KEGG analysis). We examined protein-protein interactions to characterize the relationship between differentially expressed genes, and we screened potential hub genes according to the degree of connection. PCR and Western blotting were used to identify the core genes. RESULTS: At different times of infarction, a total of 35 genes showed upregulation at all times; however, none of the genes showed downregulation at all 3 times. Similarly, 10 hub genes with high degrees of connectivity were identified. In vivo and in vitro experiments suggested that expression levels of MMP-9 increased at various times after myocardial infarction and that expression increased in a variety of cells simultaneously. CONCLUSION: Expression levels of MMP-9 increase throughout the course of acute myocardial infarction, and this expression has both positive and negative sides. Further studies are needed to explore the role of MMP-9 in MI treatment. The potential values of Il6, Spp1, Ptgs2, Serpine1, Plaur, Cxcl5, Lgals3, Serpinb2, and Cd14 are also worth exploring.

The reliability and validity of Chinese version of SF36 v2 in aging patients with chronic heart failure.

BACKGROUND: Chronic heart failure (CHF), a major public health problem worldwide, seriously limits health-related quality of life (HRQOL). How to evaluate HRQOL in older patients with CHF remains a problem. AIM: To evaluate the reliability and validity of the Chinese version of the Medical Outcomes Study Short Form version 2 (SF-36v2) in CHF patients. METHODS: From September 2012 to June 2014, we assessed QOL using the SF-36v2 in 171 aging participants with CHF in four cardiology departments. Convergent and discriminant validity, factorial validity, sensitivity among different NYHA classes and between different age groups, and reliability were determined using standard measurement methods. RESULTS: A total of 150 participants completed a structured questionnaire including general information and the Chinese SF-36v2; 132 questionnaires were considered valid, while 21 patients refused to take part. 25 of the 50 participants invited to complete the 2-week test-retest questionnaires returned completed questionnaires. The internal consistency reliability (Cronbach's α) of the total SF-36v2 was 0.92 (range 0.74-0.93). All hypothesized item-subscale correlations showed satisfactory convergent and discriminant validity. Sensitivity was measured in different NYHA classes and age groups. Comparison of different NYHA classes showed statistical significance, but there was no significant difference between age groups. DISCUSSION: We confirmed the SF-36v2 as a valid instrument for evaluating HRQOL Chinese CHF patients. Both reliability and validity were strongly satisfactory, but there was divergence in understanding subscales such as "social functioning" because of differing cultural background. CONCLUSIONS: The reliability, validity, and sensitivity of SF-36v2 in aging patients with CHF were acceptable.