. 2022 Feb 17;8(4):579–586. doi: 10.1001/jamaoncol.2021.7656

Table 1. Estimated Hazard Ratios for Relapse-Free Interval From a Multivariate Model.

Variable	Hazard ratio (95% CI)	P value
Age, y
≤40	1 [Reference]
41-50	0.47 (0.35-0.65)	<.001
51-60	0.35 (0.26-0.48)	<.001
61-70	0.44 (0.32-0.62)	<.001
≥71	0.46 (0.28-0.75)	.002
Race^a
Asian	1.02 (0.67-1.56)	.92
Black	1.39 (1.05-1.84)	.02
White	1 [Reference]
Unknown or not reported	0.71 (0.43-1.17)	.18
Insurance type^b
Private	1 [Reference]
Medicare	0.90 (0.66-1.24)	.52
Medicaid	0.91 (0.62-1.35)	.64
None	1.05 (0.65-1.70)	.83
International participants	0.82 (0.41-1.64)	.57
Other or unknown	1.30 (0.71-2.38)	.40
Neighborhood Deprivation Index
Quartile 1, lowest deprivation	1 [Reference]
Quartile 2	1.11 (0.85-1.46)	.44
Quartile 3	1.06 (0.81-1.40)	.65
Quartile 4, highest deprivation	1.03 (0.79-1.36)	.81
Unknown	1.24 (0.62-2.45)	.54
Tumor size in the largest dimension, cm
≤2	1 [Reference]
>2	2.08 (1.75-2.48)	<.001
Histologic grade of tumor
Low	1 [Reference]
Intermediate	1.65 (1.29-2.12)	<.001
High	2.11 (1.57-2.83)	<.001
Unknown	2.45 (1.52-3.94)	.002
Early discontinuation of endocrine therapy^c	1.05 (0.73-1.51)	.80
No endocrine therapy	2.12 (1.21-3.73)	.009
Breast cancer recurrence test score^d	NA	<.001

Abbreviation: NA, not applicable.

^{^a}

Other categories in race included Native American (n = 39), Pacific Islander (n = 30), multirace (n = 10), and other/unknown (n = 353).

^{^b}

Insurance status and type were only determined for US participants. Private refers to insurance within the US.

^{^c}

Time-dependent covariate: patients analyzed as entering the early discontinuation state 150 days after treatment was last given.

^{^d}

Recurrence score (determined with Oncotype DX, Genomic Health Inc) was used as a continuous variable in the model; therefore, no hazard ratio is reported.