Abstract: Objective To analyze ecological associated factors of liver cancer deaths and to provide evidences for rational allocation of health resources by health administrators.Methods The study involved 161 counties or districts across China covered by National Disease Surveillance System.The collected data include:liver cancer mortality,hepatitis B and hepatitis C incidence,prevalence of hepatitis B virus (HBV) serological markers,prevalence of behavior risk factors,and relevant social and economic indicators.A three-level extra-Poisson variance model (liver cancer deaths as the first level,161 study sites as the second level,and the province/autonomous region/municipality directly under the central government as the third level) was adopted to analyze ecological factors associated with liver cancer deaths.SAS 9.4 and MLwiN 2.30 software were used in statistical analyses;the three-level extra-Poisson variance random intercept model was fitted.Results There were gender and between age differences in liver cancer mortality and interactive effect of age and gender on liver cancer mortality for the counties or districts in a same province/autonomous region/municipality.The random intercept of multilevel model was of statistical significance at county/district level and province/autonomous region/municipality level,suggesting that there were significant difference in liver cancer mortality between counties or districts and between provinces/autonomous regions/municipalities.The results of multiple factors analysis revealed that liver cancer mortality was independently and ecologically associated with average wage of on-the-job workers,the prevalence of HBV surface antigen (HBsAg),reported incidence rate of hepatitis B,reported incidence rate of hepatitis C,gender,age,and the interaction between gender and age.Conclusion Higher average wage of on-the-job workers is reversely correlated with liver cancer mortaltiy;while prevalence rate of HBsAg,reported incidence rate of hepatitis B and hepatitis C,male gender,and elder age are positively correlated with liver cancer mortality.