Estimating the prevalence of disease is an essential component to understanding disease burden, unmet need, and the potential budget impact of new treatments. Economic and epidemiologic models in oncology often calculate disease prevalence using median overall survival (OS). In this study, we propose an alternative approach that provides more accurate estimates of prevalence. A formula was developed from base principles to provide a reasonable estimate of prevalence for diseases of long duration but high mortality, such as cancer. Given the assumption that incidence is stable, prevalence can be calculated as incidence times the area under the survival curve. This is often simplified to incidence times average disease duration. In oncology, median OS is routinely substituted as average disease duration. We derived a more accurate formula, assuming a constant hazard for OS and a nonzero cure rate. The derived formula is prevalence ≈ 0.12*I*mOS*(1-0.5ˆ(Cm/mOS)) where I is annual incidence, mOS is median OS in months, and Cm is time until a patient is considered cured in months. Assuming Cm = 60 months, the formula predicts that true prevalence is up to 44% higher than estimated using the simplified approach of incidence times median duration of survival for diseases with median OS less than 12 months. Conversely, the formula predicts that true prevalence is at least 25% lower than current estimation methods for diseases with long median survival (e.g., more than 5 years). This formula provides a simple alternative approach for more accurate estimation of estimating disease prevalence. Given the essential role that prevalence estimates play in oncology models, this has the important implications for future health care decision-making.