Table 1. Excessa mortality rate ratios (RRs) and 95% confidence intervals (CI) for the available prognostic factors by ethnicity.
|
Univariate analysis |
|||||||||
|
South Asian women (n=1037) |
Non-South Asian women (n=50201) |
||||||||
| Risk factor | n | (%) | RR | (95% CI) | n | (%) | RR | (95% CI) | Heterogeneityb |
| Age (years) | |||||||||
| <45 | 483 | (47) | 1 | 11 035 | (22) | 1 | |||
| 45–64 | 419 | (40) | 0.80 | (0.61, 1.07) | 17 397 | (35) | 0.94 | (0.90, 0.99) | |
| ⩾65 | 135 | (13) | 1.13 | (0.73, 1.75) | 21 769 | (43) | 1.40 | (1.33, 1.46) | |
| Linear trendc | P=0.63 | P<0.001 | P=0.56 | ||||||
| Period of diagnosis | |||||||||
| 1986–1989 | 556 | (54) | 1 | 30 324 | (60) | 1 | |||
| 1990–1993 | 481 | (46) | 0.72 | (0.55, 0.94) | 19 877 | (40) | 0.71 | (0.68, 0.74) | |
| Heterogeneityd | P=0.02 | P<0.001 | P=0.53 | ||||||
| Stage | |||||||||
| 1 (local) | 358 | (35) | 1 | 23 346 | (46) | 1 | |||
| 2 | 132 | (13) | 2.23 | (1.37, 3.62) | 4710 | (9) | 1.88 | (1.74, 2.03) | |
| 3 | 240 | (23) | 3.09 | (2.07, 4.60) | 11 002 | (22) | 2.95 | (2.80, 3.11) | |
| 4 (metastases) | 99 | (10) | 8.38 | (5.54, 12.68) | 4500 | (9) | 12.46 | (11.80, 13.15) | |
| NK | 208 | (20) | 1.54 | (0.95, 2.51) | 6643 | (13) | 1.70 | (1.59, 1.83) | |
| Linear trende | P<0.001 | P<0.001 | P=0.06 | ||||||
| Socioeconomic Deprivation | |||||||||
| 1 (affluent) | 145 | (14) | 1 | 13 151 | (26) | 1 | |||
| 2 | 150 | (14) | 1.39 | (0.81, 2.38) | 11 875 | (24) | 1.10 | (1.04, 1.16) | |
| 3 | 205 | (20) | 1.29 | (0.77, 2.16) | 10 311 | (21) | 1.22 | (1.15, 1.29) | |
| 4 | 234 | (23) | 1.53 | (0.94, 2.50) | 8716 | (17) | 1.35 | (1.28, 1.43) | |
| 5 (deprived) | 291 | (28) | 1.71 | (1.07, 2.74) | 5734 | (11) | 1.46 | (1.37, 1.56) | |
| NK | 12 | (1) | 2.35 | (0.83, 6.65) | 414 | (1) | 1.18 | (0.95, 1.47) | |
| Linear trende | P=0.02 | P<0.001 | P=0.86 | ||||||
NK=not known.
a
The excess mortality RRs for the categories of the risk factors in the table were computed from relative survival models, each including one of the risk factors under study and fitted separately in South Asian and non-South Asian women. The RRs represent the estimated effects that each separate factor has on excess mortality rates.
b
Significance of the likelihood ratio test for heterogeneity between South Asian and non-South Asian women.
c
Significance of the likelihood ratio test for linear trend in the category-specific rates.
d
Significance of the likelihood ratio test for heterogeneity between the category-specific rates.
e
Significance of the likelihood ratio test for linear trend computed excluding the NK category.