Study On The Spatial Characteristic Of HIV/AIDS Infection Among IDUs And The Influence Factors Of The Infection In China

Objective:To describe and analyze the HIV/AIDS infection among injection drugs users (IDUs) in China from the year of1995to2011. In the case of possible cluster of infection and data, explore the influence factors of infection at individual level and macroscopic level.Method:Using the data of HIV infected and AIDS patients through injecting drug between2005and2011, also analyze demographic characteristics of injection drug users. Spatial correlation analysis (provincial level and country level) and median center of hot spots (country level) were conducted by Arcgis software.Collection of an annual (2011) sentinel monitoring data among community-based recruitment drug user, using chi-square analysis and three-level logistic model to analyze the impact of the AIDS epidemic factors of China’s drug user, computing variable parameter estimates and the value of OR.GWR model was established to explore the impact of social-economic factors on HIV/AIDS infection among IDUs. Accumulative identified numbers of infection cases among IDUs per100,000person in each province from2007to2011was defined as the dependent variable. To reduce the correlation among dozens of social-economic variables, four macroscopic-factors were extracted from social-economic variables at the provincial level by using principal component analysis, reflecting the level of the economic situation, transportation, public security and health:care welfare level respectively. All the macroscopic-factors were introduced into the GWR model as independent variable to explain the impact of macroscopic-factors on identified HIV/AIDS infections among IDUs and get the provincial coefficients of four macroscopic-factors.Results:The distribution of HIV/AIDS infection among IDUs in China during the period between the1995and2011was not random. The general autocorrelation test was conducted for the accumulative number of HIV/AIDS cases during the study period, the cluster was observed (Moran’s1=0.066, Z-value=32.629and P<0.05), and the cluster also existed in infection among IDUs by years. The hotspots mainly located at Xinjiang Uygur Automous Region, Yunnan province, Guangxi Zhuang Autonomous Region and Sichuan province etc. The western hotspots were confined to Xinjiang Uygur Automous Region, while the southwestern hotspots moved from west to east during1995between2003. Since2004. the western hotspots continuous located at Xinjiang Uygur Automous Region, and the trend of southwestern hotspots moved from border regions to inland was observed.42,011drug user who were recruited from community and the HIV prevalence was5.01%(95%CI:4.80%,5.22%). Null model analysis showed that HIV-infected drug users show cluster in country and province level. Further analysis through three levels of Logistic model showed infected with hepatitis C (OR=6.404,95%CI:5.411,7.580), sharing needles (OR=OR=4.043,95%CI:3.545,4.618), injection drug use (OR=1.736,95%CI:1.391,2.167), minorities(OR=1.728,95%CI:1.446,2.065) and other demographic characteristics and behavioral factors are risk factors of community drug users infected with HIV.HIV/AIDS cases density among IDUs were spatially clustered in China (Z value=0.102and P<0.05). Further GWR model showed that economic status had a negative effect on the identified HIV infections among IDUs in most provinces, and transportation was another negative macroscopic-factor for the local identified cases density among IDUs. Social security and health care services were positive factors for HIV infections. After fitting geographically weighted regression model, using global autocorrelation test to analyze the value of residual(the difference between predicted the number of HIV/AIDS infections and Realistic value) and there was no spatial cluster observed (Moran’s I=0.030, Z value=1.970and.P=0.56. The adjusted R Square=62.5%,Condition Number less than30, which indicated that fitness of geographically weighted regression model is good.Conclusions:The distribution of HIV/AIDS infection among IDUs in China during the period between the1995and2011was not random, and hotspots moved from border regions to inland was observed in recent years. When analyze the influence factors of HIV/AIDS infection among IDUs, The multilevel model can be used to deal with the AIDS sentinel surveillance data of China, but still need to continue to explore the influencing factors at the population level. Economy, transportation, public security and safety, health had different response to HIV/AIDS infection among injecting drug user in China. Geographically Weighted Regression models can be used to analyze China’s macroscopic-factor of HIV/AIDS infection prevalence among injecting drug users.