Separate Cox regression models linking physical activity to mortality due to cancer in 2,867 Hong Kong elderly men and women.

aMode1: Adjusted for age (continuous), education level (no education, primary, secondary/matriculation, and university or above), Hong Kong ladder (bModel 2: Further adjusted for BMI (continuous), frailty index (continuous), living arrangement (live alone/live with spouse/others), and level of leisure time physical activity (continuous)/housework (continuous)*.*Level of leisure time physical activity was adjusted in the models for the relationship between non-leisure time physical activity and all-cause mortality only while level of housework was adjusted in the models for the relationship between leisure time physical activity and all-cause mortality only.Men: 1st tertile, PASE nd tertile, PASE 75.7143-rd tertile, PASE ≥111.3333/Women: 1st tertile, PASE nd tertile PASE, 69.5714-rd tertile 3, PASE ≥95.5476.