Background: Bone metastases occur in up to 90% of men with metastatic prostate cancer and cause skeletal-related events (SREs) such as fracture and spinal cord compression. Although bone modifying agents (BMAs) have been shown to decrease the risk of SREs in men with metastatic castration-resistant prostate cancer (mCRPC), BMAs are not widely used in this patient population. In this retrospective study, we report the prescribing patterns of BMAs and predictors of BMA use in a large national cohort. Methods: We used the VA corporate data warehouse to identify patients with mCRPC who were treated with first-line mCRPC therapy (abiraterone, enzalutamide, docetaxel, or ketoconazole) between 2010 and 2017. BMA prescribing frequency and patterns were analyzed. Multivariable regression analysis was used to evaluate clinical and disease-specific factors associated with BMA use which are presented as adjusted odds ratios (OR) with 95% confidence intervals (CI). Results: In the final cohort of 4,079 patients, the median age at mCRPC diagnosis was 73 years (IQR 66-81). 29% of patients in the cohort were Black. Diagnosis codes identified bone metastases at mCRPC diagnosis in 48% of the cohort (ICD-9 and 10 coding for sites of metastases are known to underestimate the actual incidence of metastases). Only 48% of patients received a BMA following initiation of treatment for mCRPC, 47% of which had already received a BMA 6 months prior to mCRPC diagnosis. For those who were new BMA starts, the median time to BMA administration from mCRPC treatment was 3.4 months (IQR 1.2-8.8). Factors associated with BMA use were having an ICD-9 or 10 diagnosis code for bone metastases at time of mCRPC treatment (OR 1.28, 95% CI 1.10-1.48), concurrent corticosteroid use (OR 1.90, 95% CI 1.53-2.38), and docetaxel use as first-line mCRPC therapy (OR 1.78, 95% CI 1.45-2.18). Factors associated with decreased BMA use were Charlson comorbidity index score of ≥ 2 (OR 0.75, 95% CI 0.62-0.90), advancing age (OR 0.86 per 10 years, 95% CI 0.79-0.94), and decreased eGFR (eGFR 30-59, OR 0.81, 95% CI 0.68-0.96; eGFR 0-29, OR 0.23, 95% CI 0.14-0.37). Having a diagnosis code for a SRE before mCRPC diagnosis (e.g., bone fracture and cord compression) did not impact BMA use, but the incidence of those events was low. Conclusions: Less than half of this VA cohort received a BMA after starting therapy for mCRPC. Key factors associated with BMA use were corticosteroid use and first-line docetaxel use, which likely represented more aggressive disease. Notably, patients who were older and had more comorbid conditions were less likely to receive a BMA despite being highest risk for frailty and SREs. Future quality improvement efforts aimed at increasing BMA use may address the needs of the geriatric mCRPC patient population.