<p><b>INTRODUCTION</b>Osteosarcoma treatment has experienced a renaissance in the last 3 decades with the institution of multimodality treatment involving multiagent chemotherapy and surgery. Yet globally, treatment success has stagnated at about 70% survival at 5 years in most single institution series. We performed survival analyses on 2 national databases in 2 countries and compared these with corresponding institution specific survival.</p><p><b>MATERIALS AND METHODS</b>All patients with the diagnostic code of non-metastatic intramedullary osteosarcoma in the long bones of the upper and lower limbs less than 30 years of age were selected from the Surveillance Epidemiology and End Result (SEER) database to ensure uniformity with respect to disease and treatment. We studied the factors: ethnicity, gender, age, grade, histology, size, site, surgery, compartmentalisation, number of primaries and venue of treatment for their contribution to survival. In addition, the data were stratified into 3 decades (seventies, eighties and nineties) to account for variations due to the evolution of treatment paradigms and imaging modalities.</p><p><b>RESULTS</b>Institution-specific survival was predictably better than national survival in the 4 databases. One thousand patients were selected from the SEER database. Oriental descent, state-specific treatment, female gender, treatment in the nineties, low-grade disease, intra-compartmental disease, small size, wide resections as opposed to forequarter or hindquarter amputations, and single primaries were good prognostic factors on univariate analysis as well as multivariate analysis (P <0.05). Survival was better in the more affluent states (P <0.05). Males were affected at an older age than females (P = 0.004). Blacks tended to have larger tumours although their overall survival was similar to whites. Orientals were more likely to be treated in the nineties with wide resections for smaller tumours and were located around states associated with good treatment. Orientals in Singapore and the United States had the same survival (P = 0.45). Survival in Orientals in Singapore was not significantly different from other races. The standard of healthcare for osteosarcoma varies greatly across the United States but is uniform in Singapore. Hence the observed differences in the United States were likely due to socioeconomic factors.</p><p><b>CONCLUSION</b>This analysis confirms the importance of a number of prognostic variables in osteosarcoma and suggests the possibility of an ethnic and economic bias for good survival.</p>