Spatial And Temporal Distribution Of Cardiovascular Diseases And Remote Sensing Diagnosis Of Response Relation With Its Environmental Health Factors

Cardiovascular disease is one of severe disease threatening human life and health.The distribution of cardiovascular diseases mortality is spatial clusters.Eastern Europe and Central Asia had a high mortality rate of cardiovascular diseases.And its spatial distribution characteristics are influenced by many factors such as geographical environment,climatic factors,atmospheric quality and people's living habits.Remote sensing technology(RS)has faster speed to get the data,a shorter period than any other technology,and the data have more information and large range.It can provide data support for the monitoring of cardiovascular diseases environmental factors on a global scale.Geographic information system technology(GIS)can visually display diseases and various risk factors on a map,and study the spatial distribution of diseases and relationship between disease and environmental risk factors with various spatial analysis methods.The application of GIS can provide technical support for cardiovascular disease research and the theory of remote sensing diagnosis of environmental health provides a new idea on describing that cardiovascular diseases,the response of environmental factors and predicting cardiovascular diseases.In this paper,RS and GIS were used to analyze the temporal and spatial distribution characteristics of global cardiovascular disease mortality.Then we select the environmental risk factors of cardiovascular disease mortality and study the response characteristics based on the spatial and temporal distribution of mortality.At last,we build a prediction model of cardiovascular disease mortality based on the environmental risk factors of cardiovascular disease mortality.The main contents and conclusion of this study are as follows:(1)Extraction of spatial and temporal distribution characteristics of cardiovascular disease mortality.With analyzing the temporal and spatial distribution characteristics of cardiovascular mortality in 2000,2004,2008 and 2012,we get the results that cardiovascular mortality remained at a high level from 2000 to 2012,with an annual global mortality rate of over 250(per 100,000).The high incidence of cardiovascular disease mortality is concentrated in Eastern Europe and Central Asia.Oceania and Latin America have lower rates of cardiovascular mortality.In addition,the distribution of cardiovascular mortality has spatial autocorrelation,with more than 14countries having a high concentration of positive correlation each year.From 2000 to2012,they have spatial and temporal aggregation every year,and the main aggregation areas were the Mediterranean coast,Central Asia,South Asia and Southeast Asia.(2)The remote sensing diagnosis of cardiovascular disease mortality response to environmental risk factors.The temporal and spatial distribution of cardiovascular mortality was used to select the environmental risk factors of cardiovascular diseases,using geographic statistical analysis and single factor analysis.The result shows that within a certain range,atmospheric pollution factors have an impact on the mortality distribution of cardiovascular diseases.The mortality of cardiovascular diseases is firstly decreased and then increase with the increase in air pollution factors'concentration.When the concentration of SO₂ column is greater than 0.9Du or the total amount of O₃ column is greater than 320Du,the mortality of cardiovascular diseases will reduce.The mortality of cardiovascular diseases is relatively high in the regions where the annual average temperature is in the range of the 10-20?and the annual average air pressure is between 950-1000 h Pa.Annual average and the maximum wind speed in the diurnal range has a significant influence on cardiovascular disease mortality distribution.The mortality of cardiovascular diseases is relatively high in the regions where the annual average wind speed is 2-3 m/s and the highest average annual temperature is 20-23?in diurnal range.(3)Selection of environmental risk factors of cardiovascular disease mortality.Generalized Additive Model(GAM)was used to select the environmental risk factors affecting cardiovascular diseases.With stepwise regression analysis based on Akaike Information Criterion(AIC)generalized additive model,we get results that the concentration of NO₂ columns,O₃ columns,annual average temperature,annual mean air pressure,the annual average wind speed was significantly associated with cardiovascular disease mortality rates,and regional parameters,concentration of ozone column and the annual maximum diurnal range has a certain correlation with cardiovascular disease,but not significant.(4)The remote sensing diagnostic model for cardiovascular disease mortality prediction.Using OMI data,meteorological data and historical cardiovascular mortality data,a global cardiovascular mortality prediction model was established.The six risk factors of NO₂ column concentration,O₃ column total volume,annual average temperature,annual average air pressure,annual average wind speed and annual maximum daily range and the historical cardiovascular disease mortality are as the explanatory factors of the prediction model.It establishes the tandem grayscale Back Propagation(BP)neutral network model.The model fits well.Moreover,from the error analysis results and the actual prediction verification results,the accuracy of the tandem grayscale BP cardiovascular disease mortality model is higher than that of the pure BP neural network model based on environmental factors.The main innovation points of this study are:(1)The temporal and spatial distribution of cardiovascular mortality and the response between mortality and environmental health risk factors were analyzed on a global scale.We explore the spatial distribution characteristics,autocorrelation and aggregation of cardiovascular diseases on the global scale.And we extract the response of cardiovascular diseases to environmental risk factors based on the environmental risk factors,such as air pollution factors and meteorological factors;(2)Proposed a remote sensing diagnosis algorithm for cardiovascular mortality risk factors selection,based on regional parameter.We make AIC as an indicator parameter of this algorithm.And the algorithm has the advantage of generalized additive function;(3)Established a remote sensing diagnosis cardiovascular mortality prediction model based on tandem gray BP neural network model.We want to make the prediction model have less volatility and Greater accuracy.So we combined gray scale model and the neutral network in series.And the cardiovascular disease prediction model will be established based on it and the environmental health factors and historical disease data.