Since 2001, cases of myelodysplastic syndromes (MDSs) have been tracked by cancer registries. Examining registry data in the USA, the reported age-adjusted incidence of MDS per 100,000 was 3.3 per year for 2001-2003 and 4.9 per year for 2007-2011, with increases likely a result of growing awareness of reporting requirements. However, active case-finding methods repeatedly demonstrate that population-based registries have underestimated the incidence of MDS due to underreporting and underdiagnosis. Using keyword search strategies of electronic pathology reports or other novel case capture methods, the true incidence of MDS has been estimated between 5.3 and 13.1 per 100,000. Using Medicare billing claims data, the incidence of MDS per 100,000 in patients aged ≥65 years has been estimated between 75 and 162. MDS prevalence is estimated to be 60,000 and -170,000 in the USA and projected to grow. Epidemiologic data can help estimate the burden of MDS and expose unmet clinical needs. For example, patients with MDS receiving transfusions had significantly higher reported health care costs versus those that did not (3-year mean of $88,824 vs $29,519). Epidemiologic data also revealed that most MDS patients receiving transfusions do not receive active therapies, despite strong evidence that hypomethylating agents and lenalidomide significantly reduce transfusion burden. Other unmet needs identified by epidemiologic studies include high need for treatment options after failing first-line therapy and shared decision making by older MDS patients.