Skip to main content

Table 3 Hazard ratios for region comparisons in Cox regression models of CVD incidence and fatalities

From: Geographical variation in cardiovascular incidence: results from the British Women's Heart and Health Study

  N South England Midlands/
Wales
North
England
Scotland
Hazard ratios for incident non-fatal CVD events (95% CI)
Using subset with no missing values for any considered factor
Model 1 2253* 1 0.79 (0.50, 1.26) 0.97 (0.69, 1.37) 0.70 (0.41, 1.19)
Model 2** 2253* 1 0.71 (0.44, 1.14) 0.87 (0.61, 1.25) 0.62 (0.35, 1.12)
Using imputed data for missing values (10 imputations)
Model 1 4281*$ 1 0.64 (0.63, 1.23) 1.06 (0.82, 1.36) 0.92 (0.64, 1.33)
Model 2 4281*$ 1 0.82 (0.58, 1.15) 0.99 (0.76, 1.29) 0.86 (0.58, 1.28)
Hazard ratios for incident fatal CVD events (95% CI)
Using subset with no missing values for any considered factor
Model 1 2256 1 0.83 (0.38, 1.83) 1.19 (0.67, 2.10) 1.59 (0.78, 3.25)
Model 2 2256 1 0.63 (0.28, 1.41) 0.98 (0.54, 1.76) 1.07 (0.48, 2.37)
Using imputed data for missing values (10 imputations)
Model 1 4284$ 1 1.41 (0.91, 2.18) 1.53 (1.06, 2.20) 1.63 (1.01, 2.63)
Model 2 4284$ 1 1.13 (0.72, 1.77) 1.19 (0.82, 1.74) 1.09 (0.66, 1.82)
Hazard ratios for combined incident fatal/non-fatal CVD events (95% CI)
Using subset with no missing values for any considered factor
Model 1 2253* 1 0.77 (0.51, 1.16) 1.01 (0.75, 1.36) 0.92 (0.60, 1.41)
Model 2** 2253* 1 0.66 (0.43, 1.00) 0.87 (0.64, 1.19) 0.75 (0.47, 1.19)
Using imputed data for missing values (10 imputations)
Model 1 4281*$ 1 0.98 (0.75, 1.28) 1.20 (0.97, 1.48) 1.14 (0.85, 1.53)
Model 2 4281*$ 1 0.85 (0.64, 1.12) 1.05 (0.84, 1.31) 0.93 (0.68, 1.27)
  1. * excludes 3 with event on date of baseline questionnaire
  2. $ excludes 1 with missing date of birth and 1 with missing date for baseline questionnaire
  3. ** some evidence for violation of proportional hazards (global test in non-fatal Model 2a, p = 0.04; in combined Model 2a, p = 0.03)
  4. Model 1: unadjusted
  5. Model 2: adjusted for risk factors as described in Table 1 (except diabetes, where all cases were merged to give a binary variable)