Spatiotemporal heterogeneity of influencing factors of pulmonary tuberculosis incidence rate in China based on PCA-GTWR model
10.3760/cma.j.cn341190-20231115-00420
- VernacularTitle:基于PCA-GTWR模型的中国肺结核发病率影响因素时空异质性研究
- Author:
Mengmeng DAI
1
;
Yayi WANG
;
Peiji LI
;
Yingbo LIU
Author Information
1. 中国药科大学理学院生物统计系，南京 211198
- Keywords:
Tuberculosis,pulmonary;
Incidence;
Root cause analysis;
Principal component analysis;
Spatiotemporal analysis
- From:
Chinese Journal of Primary Medicine and Pharmacy
2024;31(5):753-759
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To study the spatiotemporal heterogeneity of the incidence rate of pulmonary tuberculosis and its macro influencing factors in China, so as to provide a reference for the relevant departments to formulate prevention and control policies.Methods:In virtue of the complexity of influencing factors, a new method combining principal component analysis (PCA) with geographically and temporally weighted regression (GTWR) was proposed to analyze the spatiotemporal heterogeneity of influencing factors for pulmonary tuberculosis. Using the data of incidence rate of pulmonary tuberculosis and gross regional product (secondary indicators) of 31 provinces (excluding Hong Kong, Macao, and Taiwan) in China from 2010 to 2019, a macro influencing factor indicator system was established by the PCA scores of 21 secondary indicators quantified to determine four primary indicators: comprehensive economy, medical security, cultural and educational transportation, and resources and environment. Based on the indicator system, PCA-ordinary least squares (PCA-OLS) model, PCA-geographically weighted regression (PCA-GWR) model, and PCA-GTWR model were constructed.Results:Three models passed F-test with F-values of 58.74, 196.62, and 1 202.90 respectively (all P < 0.05), indicating that the impact of the primary indicators on the incidence of tuberculosis is statistically significant. The mean squared error (0.01), the mean absolute error (0.08), the mean absolute percentage error (0.02), and the corrected Akaike information criterion (-358.76) of PCA-GTWR were lower than those of PCA-OLS (0.13, 0.28, 0.07, 258.38) and PCA-GWR (0.06, 0.15, 0.03, 23.41). Meanwhile, the determination coefficient (0.95) of PCA-GTWR was higher than that of PCA-OLS (0.44) and PCA-GWR (0.77), indicating the goodness of fit of the model is the best. And the PCA-GTWR model showed that the comprehensive economy, medical security, cultural and educational transportation, and resources and environment had significant spatiotemporal heterogeneity on the incidence rate of pulmonary tuberculosis according to the distribution of regression coefficients. Conclusion:It is necessary to comprehensively consider various factors and formulate detailed and overall prevention and control measures for pulmonary tuberculosis according to local conditions.
