Background: Novel efficacy endpoints are needed that correlate with overall survival [OS] and can describe real world outcomes. Methods: We mined national VA data (VINCI) using a novel set of equations validated in > 10,000 patients that estimate simultaneously occurring exponential rates of tumor growth [g] and regression [d] using data gathered while a patient receives cancer treatment. We have previously established that g is highly correlated with OS and can esti mate doubling time (dt). Importantly, since the equations include time as a variable, this approach is ideal for real world analyses where re-assessments depend on the practitioner and are highly variable. To validate g in a real-world cohort, we collected cases of PC, demographics, treatments and outcomes from VINCI and compared parameters by receipt of chemotherapies for PC. Results: 5,890 Veterans were treated with abiraterone [ABI], enzalutamide [ENZA] or both. Median age was 75 years including 2,596 Veterans > 80 years, and 23% identified as African American [AA]. PSA values beyond the initial measurement were available in 88% of patients with little clinical difference between those with and those without serial PSA values. g and d could be estimated in 83-85%f Veterans with p-values for fits < 0.1 in all and < 0.001 in the majority. g values for Veterans receiving either ABI [0.0038d^-1; dt 182d] or ENZA [0.0040d^-1; dt 173d] in first line were indistinguishable (p = 0,27), suggesting comparable efficacy. Consistent with the clinical bias, in second line, ENZA [0.0071d^-1; dt 98d] appears superior to ABI [0.0091d^-1; dt 76d] (p < 0.01). However, preliminary analyses find g on 1^st line ABI remains constant in the majority and ABI continuation may be beneficial. Importantly g was independent of age, treatment location, and race, demonstrating comparable benefit in AA and non-AA Veterans. Conclusions: This is the largest real world assessment of ABI and ENZA efficacy in PC with a high percentage of AAs. The results underscore the value of determining g as an excellent measure of efficacy and argue for its use in outcomes research.