Results of logistic regression analysis within the derivation group and calculation of simplified prediction rule coefficients
Covariate . | b . | Odds ratio, exp(b) (95% CI) . | P . | Simplified coefficient . |
|---|---|---|---|---|
| x1: B symptoms | b1 = 1.916 | 6.8 (2.2-21.1) | .001 | b1s = 8 |
| x2: Clinical stage III/IV prior to BMB | b2 = 1.433 | 4.2 (1.7-10.3) | .002 | b2s = 6 |
| x3: Anemia | b3 = 1.316 | 3.7 (1.5-9.3) | .005 | b3s = 5 |
| x4: Leukocytes < 6 × 109/L | b4 = 1.368 | 3.9 (1.8-8.7) | .001 | b4s = 5 |
| x5: Age ≥ 35 y | b5 = 0.774 | 2.2 (1.03-4.6) | .04 | b5s = 3 |
| x6: Iliac/inguinal involvement | b6 = 0.763 | 2.1 (1.02-4.5) | .04 | b6s = 3 |
| Constant | b0 = -6.792 | NA | NA | b0s = - 8* |
Covariate . | b . | Odds ratio, exp(b) (95% CI) . | P . | Simplified coefficient . |
|---|---|---|---|---|
| x1: B symptoms | b1 = 1.916 | 6.8 (2.2-21.1) | .001 | b1s = 8 |
| x2: Clinical stage III/IV prior to BMB | b2 = 1.433 | 4.2 (1.7-10.3) | .002 | b2s = 6 |
| x3: Anemia | b3 = 1.316 | 3.7 (1.5-9.3) | .005 | b3s = 5 |
| x4: Leukocytes < 6 × 109/L | b4 = 1.368 | 3.9 (1.8-8.7) | .001 | b4s = 5 |
| x5: Age ≥ 35 y | b5 = 0.774 | 2.2 (1.03-4.6) | .04 | b5s = 3 |
| x6: Iliac/inguinal involvement | b6 = 0.763 | 2.1 (1.02-4.5) | .04 | b6s = 3 |
| Constant | b0 = -6.792 | NA | NA | b0s = - 8* |
CI indicates confidence interval; BMB, bone marrow biopsy; NA, not applicable.
The simplification of b0 was different than the simplification of b1 to b6. The value of - 6.792 was rounded to - 6.75 and quadrupled to - 27. To set the optimal cutoff score of Zs to 0 (for reasons of convenience), we added 19 in all Z scores, augmenting b0 from - 27 to - 8.