Table 3.

Relative Hazards Ratio (HR) of Consistent Tea Drinking Habit between 60 Years Old and the 1998 Wave Survey Time On 7-Year Mortality, CLHLS, 1998–2005

Model	Infrequently drinking consistently [n=6,143]	Drinking inconsistently [n=740]	Frequently drinking consistently [n=1,748]	P for linear trend
Model 1	1	1.00 (0.91 to 1.01)	0.93(0.87 to1.00)	0.053
Model 2	1	0.99 (0.90 to 1.09)	0.91 (0.85 to 0.96)	0.010
Model 3	1	0.97(0.88 to 1.07)	0.90 (0.84 to 0.97)	0.042

Note: Values in the cell are hazards ratio (HR) estimated based on Cox semi-parametric proportional hazard model, and values in the bracket are 95% confidence interval (CI) of the HR.

Variables in model 1: only tea consumption status

Variables in model 2: + demographic variables and socioeconomic status (SES).

Variables in model 3: +health activities and health status