Table 4.
Univariate and multivariate cox regression models for variables predicting risk of death in the patient cohort
| Cox regression model | Univariate | Multivariate | ||
|---|---|---|---|---|
| Variable | HR (95% CI) | p-value | HR (95% CI) | p-value |
| Age | ||||
| ≤ 50 years (ref) | 1 | – | 1 | – |
| 51–60 | 1.25(1.16–1.36) | < 0.0001 | 1.17(1.04–1.31) | 0.0099 |
| 61–70 | 1.36(1.26–1.47) | < 0.0001 | 1.11(0.98–1.27) | 0.0935 |
| ≥ 70 | 2.05(1.88–2.22) | < 0.0001 | 1.53(1.31–1.79) | < 0.0001 |
| Sex | ||||
| Female (ref.) | 1 | – | ||
| Male | 1.15(0.90–1.46) | 0.2751 | ||
| Race | ||||
| White (ref.) | 1 | – | 1 | – |
| Black | 1.09(1.02–1.16) | 0.0167 | 1.04(0.94–1.15) | 0.4849 |
| Other | 0.76(0.66–0.88) | 0.0002 | 0.78(0.63–0.96) | 0.0216 |
| Ethnicity | ||||
| Non-Hispanic (ref.) | 1 | – | 1 | – |
| Hispanic | 0.65(0.58–0.73) | < 0.0001 | 0.72(0.60–0.86) | 0.0003 |
| Insurance status | ||||
| Private insurance/managed care (ref.) | 1 | – | 1 | – |
| Not insured | 1.37(1.23–1.52) | < 0.0001 | 1.38(1.18–1.61) | < 0.0001 |
| Medicaid | 1.17(1.08–1.27) | 0.0002 | 1.28(1.14–1.43) | < 0.0001 |
| Medicare | 1.57(1.48–1.68) | < 0.0001 | 1.20(1.07–1.34) | 0.0013 |
| Median income quartiles (2012–2016) | ||||
| > $63,333 (ref.) | 1 | – | 1 | – |
| $50,354–$63,332 | 1.12(1.04–1.21) | 0.0022 | 1.10(0.98–1.23) | 0.1031 |
| $40,227–$50,353 | 1.16(1.07–1.25) | 0.0002 | 1.11(0.98–1.26) | 0.0924 |
| < $40,227 | 1.17(1.08–1.26) | < 0.0001 | 1.22(1.06–1.40) | 0.0058 |
| Percent no high school degree quartiles (2012–2016) | ||||
| < 6.3% (ref.) | 1 | – | 1 | – |
| 6.3%–10.8% | 1.08(1.00–1.17) | 0.053 | 1.02(0.90–1.14) | 0.7999 |
| 10.9%–17.5% | 1.16(1.07–1.26) | 0.0004 | 1.05(0.92–1.20) | 0.4946 |
| > 17.6% | 1.05(0.97–1.14) | 0.2211 | 0.90(0.77–1.05) | 0.1804 |
| Facility type | ||||
| Academic/research program (ref.) | 1 | – | 1 | – |
| Community cancer program | 1.21(1.09–1.34) | 0.0002 | 1.06(0.92–1.22) | 0.3907 |
| Comprehensive community cancer program | 1.24(1.16–1.32) | < 0.0001 | 1.15(1.05–1.26) | 0.0018 |
| Integrated network cancer program | 1.26(1.16–1.37) | < 0.0001 | 1.21(1.08–1.36) | 0.0012 |
| Charlson Deyo score | ||||
| 0 (ref.) | 1 | – | 1 | – |
| 1 | 1.26(1.17–1.36) | < 0.0001 | 1.13(1.02–1.26) | 0.0249 |
| 2 | 1.65(1.45–1.88) | < 0.0001 | 1.32(1.09–1.60) | 0.0041 |
| ≥ 3 | 1.92(1.63–2.25) | < 0.0001 | 1.74(1.39–2.18) | < 0.0001 |
| Year of diagnosis | ||||
| 2010 | 1.32(1.18–1.49) | < 0.0001 | ||
| 2011 | 1.23(1.10–1.38) | 0.0004 | ||
| 2012 | 1.16(1.04–1.31) | 0.0108 | ||
| 2013 | 1.15(1.03–1.30) | 0.0156 | ||
| 2014 | 1.19(1.06–1.33) | 0.0033 | ||
| 2015 | 1.10(0.98–1.24) | 0.0965 | ||
| 2016 | 1.10(0.98–1.24) | 0.1067 | ||
| 2017 | 1.00(1.00–1.00) | |||
| 2018 | 1.13(1.01–1.27) | 0.0385 | ||
| 2019 | 1.00(1.00–1.00) | |||
| 2020 (ref.) | 1 | – | ||
| Year of diagnosis (regrouped) | ||||
| 2010–2011 | 1.20(1.10–1.30) | < 0.0001 | 1.25(1.08–1.45) | 0.0029 |
| 2012–2013 | 1.09(1.00–1.18) | 0.044 | 1.05(0.91–1.22) | 0.4896 |
| 2014–2015 | 1.08(0.99–1.16) | 0.0774 | 1.20(1.03–1.39) | 0.0164 |
| 2016–2017 | 1.04(0.94–1.15) | 0.5058 | 1.02(0.87–1.19) | 0.8482 |
| 2018–2020 (ref.) | 1 | – | 1 | – |
| Histology | ||||
| Ductal (ref.) | 1 | – | ||
| Lobular | 1.03(0.92–1.14) | 0.6592 | ||
| Other | 1.20(1.13–1.27) | < 0.0001 | ||
| Grade | ||||
| 1 (ref.) | 1 | – | 1 | |
| 2 | 1.15(1.03–1.28) | 0.0124 | 1.04(0.91–1.19) | 0.547 |
| 3 | 1.40(1.26–1.56) | < 0.0001 | 1.16(1.00–1.35) | 0.0551 |
| Tumor size | ||||
| > 3 cm(ref.) | 1 | – | ||
| 2–3 cm | 1.02(0.93–1.11) | 0.73 | ||
| 1–2 cm | 1.02(0.93–1.12) | 0.6889 | ||
| < 1 cm | 0.99(0.89–1.11) | 0.8897 | ||
| Lympho-vascular invasion | ||||
| 0 (ref.) | 1 | – | ||
| 1 | 1.01(0.92–1.11) | 0.8672 | ||
| Breast cancer subtype | ||||
| HR( − )/HER2( − ) (ref.) | 1 | – | 1 | - |
| HR( − )/HER2( +) | 0.52(0.47–0.58) | < 0.0001 | 0.58(0.51–0.66) | < 0.0001 |
| HR( +)/HER2( − ) | 0.51(0.48–0.55) | < 0.0001 | 0.54(0.49–0.60) | < 0.0001 |
| HR( +)/HER2( +) | 0.41(0.37–0.45) | < 0.0001 | 0.43(0.38–0.49) | < 0.0001 |
| Location of extracranial metastatic sites | ||||
| Only brain (ref.) | 1 | – | 1 | – |
| Bone | 0.96(0.88–1.05) | 0.3799 | 1.31(1.15–1.49) | < 0.0001 |
| Bone + liver | 1.22(1.09–1.37) | 0.0005 | 1.96(1.67–2.31) | < 0.0001 |
| Bone + liver + lung | 1.44(1.30–1.60) | < 0.0001 | 2.06(1.78–2.38) | < 0.0001 |
| Bone + lung | 1.07(0.97–1.19) | 0.1576 | 1.41(1.21–1.63) | < 0.0001 |
| Liver | 1.27(1.06–1.52) | 0.0107 | 1.88(1.45–2.45) | < 0.0001 |
| Liver + lung | 1.73(1.48–2.02) | < 0.0001 | 1.97(1.59–2.44) | < 0.0001 |
| Lung | 1.34(1.20–1.50) | < 0.0001 | 1.31(1.12–1.53) | 0.0009 |
| Other | 1.22(1.11–1.35) | < 0.0001 | 1.85(1.58–2.18) | < 0.0001 |
| Treatment combination | ||||
| Treatment for both (ref.) | 1 | – | 1 | – |
| Treatment for breast cancer only | 0.99(0.93–1.06) | 0.7528 | 0.93(0.85–1.01) | 0.092 |
| Treatment for brain metastasis only | 2.42(2.21–2.65) | < 0.0001 | 2.30(2.00–2.63) | < 0.0001 |
| No treatment for both | 3.14(2.91–3.38) | < 0.0001 | 2.65(2.36–2.98) | < 0.0001 |
The model included fourteen explanatory variables (age, race, ethnicity, insurance status, median household income quartile 2012–2016, percent of no high school degree, Charlson Deyo Score, histology, grade, breast cancer subtype, metastasis location sites, treatment combinations, and year of diagnosis)
aUnivariate logistic regressions ran first. Sex, tumor size, and lympho-vascular invasion all not significant so not included in multivariate model. Histology (p = 0.1024) was eliminated by backward elimination. Model set at 0.1 cutoff