Table 4.
Final model for predicting 5-year breast cancer in the NHS cohorts and BWHS cohort
| Predicting 5-year risk of breast cancer | NHS development cohort | NHS validation cohort | BWHS | |||
|---|---|---|---|---|---|---|
| Outcome being predicted | HR | P value | HR | P value | HR | P value |
| n = 37,628 | n = 18,980 | n = 13,247 | ||||
| N for that outcome | 712 | 365 | 223 | |||
| Factors in our final model | ||||||
| Age (per year increase) | 1.01 | 0.10 | 1.01 | 0.41 | 1.02 | 0.15 |
| Highest self-reported Body Mass Index (BMI) in past 10 years: (per kg/m2 increase) | 1.03 | < 0.001 | 1.02 | 0.008 | 1.02 | 0.01 |
| Average alcohol use per day (highest average use in past 10 years): None | 1 | – | 1 | – | 1 | – |
| 1–4.9 g/day | 1.12 | 0.27 | 0.99 | 0.95 | 0.79 | 0.17 |
| 5–14.9 g/day | 1.20 | 0.09 | 0.99 | 0.97 | 0.95 | 0.81 |
| 15+ g/day | 1.29 | 0.02 | 1.11 | 0.52 | 1.28 | 0.23 |
| Age at menopause (years), < 45 | 0.91 | 0.57 | 0.64 | 0.07 | 1.08 | 0.74 |
| 45–49 | 1 | – | 1 | – | 1 | – |
| 50–54 | 1.21 | 0.04 | 1.18 | 0.20 | 1.15 | 0.44 |
| 55+ | 1.32 | 0.05 | 1.13 | 0.53 | 1.08 | 0.78 |
| Mammogram in past 2 years | No | 1 | – | 1 | – | 1 | – |
| Yes | 0.91 | 0.46 | 0.87 | 0.40 | 0.68 | 0.03 |
| Number of breast biopsies, None | 1 | – | 1 | – | 1 | – |
| 1 | 1.38 | < 0.001 | 1.51 | < 0.001 | 1.28 | 0.12 |
| 2+ | 1.31 | 0.16 | 1.64 | 0.04 | 1.14 | 0.53 |
| Age at first birth (years) and parity, Nulliparous | 1.25 | 0.23 | 0.98 | 0.95 | 1.10 | 0.66 |
| < 25, 1–2 children | 1 | – | 1 | – | 1 | – |
| < 25, 3 + children | 1.06 | 0.62 | 1.16 | 0.40 | 1.02 | 0.93 |
| 25–29, 1–2 children | 1.07 | 0.63 | 1.12 | 0.58 | 1.37 | 0.15 |
| 25–29, 3 + children | 1.04 | 0.76 | 1.08 | 0.68 | 1.60 | 0.17 |
| 30+, 1–2 children | 1.37 | 0.07 | 2.02 | 0.002 | 1.33 | 0.27 |
| 30+, 3 + children | 1.06 | 0.82 | 0.61 | 0.31 | 2.92 | 0.07 |
| First-degree relatives with history of breast cancer and age at diagnosis, None | 1 | – | 1 | – | 1 | – |
| 1 and age < 50 | 1.43 | 0.03 | 1.30 | 0.28 | 1.21 | 0.56 |
| 1 and age 50+ | 1.30 | 0.02 | 1.31 | 0.08 | 1.30 | 0.14 |
| 2+ and at least one age < 50 | 2.19 | 0.003 | 3.27 | < 0.001 | 1.60 | 0.51 |
| 2+ and age 50+ | 2.43 | < 0.001 | 1.53 | 0.31 | 0.39 | 0.35 |
| Postmenopausal hormone use, Never | 1 | – | 1 | – | 1 | – |
| Current estrogen + progestin user < 5 years | 2.28 | 0.07 | 1.84 | 0.38 | 2.44 | 0.13 |
| Current estrogen + progestin user 5+ years | 2.60 | < 0.001 | 2.87 | < 0.001 | 2.31 | 0.07 |
| Current estrogen-alone user < 5 years | 1.60 | 0.15 | 1.42 | 0.50 | 1.88 | 0.14 |
| Current estrogen-alone user 5+ years | 1.36 | 0.02 | 1.33 | 0.13 | 0.98 | 0.95 |
| Past estrogen + progestin user < 5 years | 0.91 | 0.47 | 0.82 | 0.28 | 0.98 | 0.91 |
| Past estrogen + progestin user 5+ years | 1.17 | 0.17 | 1.18 | 0.30 | 1.37 | 0.25 |
| Past estrogen-alone user < 5 years | 1.01 | 0.93 | 0.78 | 0.31 | 1.04 | 0.84 |
| Past estrogen-alone user 5+ years | 0.77 | 0.06 | 1.08 | 0.64 | 0.70 | 0.15 |
| C-index (95% CI) | 0.610 (0.589–0.631) | 0.639 (0.611–0.667) | 0.617 (0.582–0.652) | |||
| C-index (95% CI) when using risk factor regression coefficients from the NHS development cohort in the NHS and BWHS validation cohorts | – | 0.612 (0.583–0.641) | 0.573 (0.536–0.611) | |||
NHS = Nurses’ Health Study, BWHS = Black Women’s Health Study