Factors associated with needle sharing among people who inject drugs in Yunnan, China: a combined network and regression analysis

Background Network analyses have been widely utilized to evaluate large datasets, but have not yet been used to explore factors associated with risk behaviours. In combination with traditional regression analysis, network analyses may provide useful information and highlight key factors for reducing needle sharing behaviours among people who inject drugs (PWID). Methods Sociodemographic data, and information on injection behaviour and sexual practices were collected from a cross-sectional survey that was conducted with PWID in five prefectures of Yunnan province, China. A combination of logistic regression and correlation network analyses were used to explore key factors for reducing needle-sharing behaviours among PWID. Results In a total of 1 049 PWID, 37.5 % had a history of needle or syringe sharing. The logistic analysis showed that Zhaotong, Qujing, Dehong, or Lincang residents, diazepam use, longer injection duration, needle reuse, and infection with HIV, viral hepatitis, tuberculosis and/or malaria were independently associated with needle sharing. The correlation network analyses showed that, compared to PWID who had never shared needles, PWID who did share needles would achieve harm reduction goals faster and more permanently. HIV serostatus and marital status were found to be closely associated with other risk factors. By combining regression analyses with network analyses, it was shown that PWID who are HIV seropositive will be an ideal target group for harm reduction programs. Conclusion Needle-sharing behaviours are common among PWID in Yunnan, and harm reduction programs may help PWID who are HIV seropositive reduce risk behaviours and prevent blood borne diseases. Electronic supplementary material The online version of this article (doi:10.1186/s40249-016-0169-y) contains supplementary material, which is available to authorized users.