Background: Skeletal related events (SREs) are a known complication for the 80% of men with metastatic prostate cancer who have bone metastases. Previous studies have demonstrated that bone modifying agents (BMAs) such as zoledronic acid and denosumab reduce SREs in men with metastatic castration-resistant prostate cancer who have bone metastases and are now recommended by national guidelines. We sought to investigate factors associated with use of BMAs in Veterans with CRPC across the Veterans Health Administration (VA). Methods: Using the VA Corporate Data Warehouse, consisting of aggregated medical record data from 130 facilities, we used an algorithm previously published to identify men with a diagnosis of castration-resistant prostate cancer (CRPC) based on rising prostate specific antigen (PSA) levels while on androgen deprivation therapy and who received systemic treatment for CRPC with one of the commonly used therapies: abiraterone, enzalutamide, docetaxel, ketoconazole between 2010 and 2017. To account for clustering among facilities, we used a multilevel multivariable logistic regression to determine the association of patient and disease-specific variables on the odds of a patient receiving a BMA after they started treatment for CRPC. Results: Of 4,998 patients with CRPC in our cohort, 2223 (44%) received either zoledronic acid or denosumab at some point after they were initiated on treatment for CRPC. After adjusting for other variables and accounting for a facility, the odds of receiving a BMA decreased by 3% for every additional year of age (odds ratio [OR] 0.97, 95% confidence interval [CI] 0.96-0.98), and decreased significantly with increasing comorbid conditions (OR 0.94, 95% CI 0.72-0.98 for Charlson Comorbidity Index [CCI] of 1; OR 0.69, 95% CI 0.59-0.81 for CCI 2+). Patients who were Black had 25% lower odds of receiving a BMA than patients who were White (OR 0.75, 95% CI 0.65-0.87). PSA at time of CRPC treatment start had a small but not significant effect on receipt of a BMA (OR 1.04, 95% CI 1.00-1.08) for every unit increase of PSA on the log scale. PSA doubling time was not associated with receipt of a BMA. The presence of a diagnosis code for bone metastases was far lower than expected in this cohort of patients with CRPC (40.7%), and thus was not included in the model. We did not expect the presence of bone metastases to vary significantly among the other independent variables. Conclusions: Despite most patients with CRPC historically having bone metastases, less than half of patients with CRPC received a BMA. Patients who are older, had more comorbidities, or were Black were less likely to receive a BMA after starting treatment for CRPC. Understanding factors that lead to different patterns of treatment can guide initiatives toward more guideline-concordant care.