Skip to main content
. 2003 Nov 8;327(7423):1072. doi: 10.1136/bmj.327.7423.1072

Table 3.

Estimates (95% confidence intervals) of incidence rate ratios for Australian citizens arriving from overseas and known to have venous thromboembolism (1981-99)

Model type exp(β1) (days 0-7)* exp(β2) (days 8-14) exp(β3) (days 15-21) exp(βs) (sex) exp(βa) (age)
Full 5.17 (3.21 to 8.31) 2.42 (1.31 to 4.49) 1.38 (0.64 to 2.98) 1.18 (0.68 to 2.06) 1.01 (0.99 to 1.02)
Parsimonious 5.61 (3.94 to 7.97) 2.63 (1.55 to 4.45)
*

First week is 7.5 days as time to event is calculated from day 0.5.

Variables testing for age and sex dependence of incidence rate ratio and for excess risk in third week are not statistically significant. Variables (β3, βs, βa) are dropped from final (parsimonious) model.