Table 4.

Sensitivity analyses: HRs for mortality of all cancers per 10 μg/m3 of PM2.5

SensitivitynHR (95% CI)P
Without using geospatial height in exposure estimation4,5461.20 (1.06–1.36)0.005
Time-varying annual mean exposure instead of baseline year exposure4,5141.21 (1.10–1.33)<0.001
Without control of competing diseases4,5311.21 (1.10–1.33)<0.001
Exclusion of subjects as potential loss to follow-up4,5311.23 (1.12–1.35)<0.001
Exclusion of deaths for the first 5 years3,6971.23 (1.11–1.37)<0.001
Exclusion of deaths for the first 7 years2,7061.14 (1.04–1.25)0.015
Exclusion of subjects who had moved during the follow-up period4,1921.22 (1.11–1.35)<0.001
Exclusion of subjects who had hospitalization record in 1998–20002,6871.25 (1.12–1.41)<0.001
Exclusion of subjects with self-reported preexisting respiratory or cardiometabolic diseases at baseline2,5821.20 (1.05–1.37)0.006
Adjustment for alcohol intake
 All malignant4,5311.22 (1.11–1.34)<0.001
 All digestive organs1,7341.22 (1.05–1.42)0.01
Random effects with adjustment of autocorrelation at planning areas (i.e., Tertiary Planning Units)4,5311.22 (1.11–1.34)<0.001
Competing risks4,5311.14 (1.04–1.25)0.007