Table 1.
Associations between completeness of ascertainment of stage and “advanced stage” at presentation (defined as stages III-IV) and other variables in men with prostate cancer, 1998-2006
| Variable and categories | All patients | Patients with stage information (% of all patients) | Predictors of completeness of stage ascertainment | Patients with stage III-IV (% of all cases with stage information) | Predictors of advanced stage (III-IV) | |||
|---|---|---|---|---|---|---|---|---|
| Odds ratio (95% CI)* | P value* | Odds ratio (95% CI)* | P value* | |||||
| Diagnosis period (years): | ||||||||
| 1998-2000 | 7899 | 2697 (34.1) | 2.58 (2.50 to 2.67) | <0.001 | 865 (32.1) | 0.86 (0.82 to 0.90) | <0.001 | |
| 2001-3 | 9860 | 5297 (53.7) | 1487 (28.1) | |||||
| 2004-6 | 10 211 | 7922 (77.6) | 2043 (25.8) | |||||
| All | 27 970 | 15 916 (56.9) | — | — | 4395 (27.6) | — | — | |
| Age group (years): | ||||||||
| 51-60 | 2632 | 1570 (59.7) | 0.88 (0.87 to 0.91) | <0.001 | 391 (24.9) | 1.15 (1.10 to 1.19) | <0.001 | |
| 61-70 | 8574 | 5085 (59.3) | 1367 (26.9) | |||||
| 71-80 | 11 156 | 6364 (57.0) | 1617 (25.4) | |||||
| >80 | 5608 | 2897 (51.7) | 1020 (35.2) | |||||
| All | 27 970 | 15 916 (56.9) | — | — | 4395 (27.6) | — | — | |
| Deprivation group: | ||||||||
| 1 (affluent) | 7235 | 4002 (55.3) | 0.99 (0.97 to 1.01) | 0.277 | 997 (24.6) | 1.06 (1.03 to 1.09) | <0.001 | |
| 2 | 7716 | 4557 (59.1) | 1303 (28.2) | |||||
| 3 | 6769 | 3999 (59.1) | 1158 (28.7) | |||||
| 4 | 4753 | 2479 (52.2) | 738 (29.4) | |||||
| 5 (deprived) | 1497 | 879 (58.7) | 250 (28.2) | |||||
| All | 27 970 | 15 916 (56.9) | — | — | 4395 (27.6) | — | — | |
*From binary logistic regression models (stage ascertainment v no ascertainment; and advanced stage v non-advanced stage, respectively) with diagnosis period, age group, and deprivation group entered as continuous variables in respective, otherwise unadjusted, models. Odds ratios denote increase in probability of outcome of interest (either completeness of stage ascertainment or advanced stage) derived by moving from one ordinal category to one immediately higher—for example, by moving from diagnosis period 1998-2000 to diagnosis period 2001-3, or by moving from 2001-3 to 2004-6, etc. P values from these models serve as test for linear trend.