Table 2.

Final multivariable model of overall survival in all patients by treatment cohort

Covariate	HR (95% CI)	P value
Treatment: sipuleucel-T + ARTAa vs ARTA monotherapyb	0.72 (0.648, 0.793)	< 0.01
Age: continuous variable	1.04 (1.040, 1.050)	< 0.01
Race: Black vs White	0.89 (0.798, 0.990)	0.03
Race: other vs White	0.76 (0.649, 0.894)	< 0.01
Medicare and Medicaid: both vs either	1.36 (1.241, 1.490)	< 0.01
Charlson Comorbidity Indexc: continuous variable (0–8)	1.05 (1.045, 1.062)	< 0.01
Opioid use around index dated: chronic vs not chronic	1.72 (1.589, 1.873)	< 0.01
Number of metastatic sites: > 1 vs ≤ 1	1.49 (1.375, 1.608)	< 0.01
Skeletal-related events around index date: any vs none	1.4 (1.300, 1.511)	< 0.01
Corticosteroid proportion of days coverede within 6 months after index datef: continuous variable (calculated, 0–1)	1.01 (1.007, 1.011)	< 0.01

ARTA androgen receptor-targeting agent, CI confidence interval, HR hazard ratio

^aPatients received sipuleucel-T followed by an ARTA or an ARTA followed by sipuleucel-T (Fig. 3). The switch occurred within 6 months of starting the first treatment

^bPatients received ARTA monotherapy

^cCharlson Comorbidity Index score was assigned on the basis of claims in the year before the index date. A score of 0 indicates that no comorbidities were found; worse comorbidities are indicated by higher scores, with a maximum possible score of 33

^dChronic opioid use is defined as two or more 30-day prescriptions within 60 days before or after the index date

^eProportion of days covered refers to the number of days of supply of corticosteroids divided by the difference in the number of days alive in the study and the number of days spent in an inpatient or skilled nursing facility care

^fExcludes corticosteroid use concomitant with abiraterone