OBJECTIVES: To determine which clinical outcomes, both “soft” and “hard”, are currently used in phase III treatment studies for hematoncologic diseases, and what proportion of them use overall survival (OS) as the primary outcome. METHODS: Through a search in PubMed, we identified and analyzed all randomized clinical trials published in the last 10 years, for de novo treatments in adults. RESULTS: The initial search yielded 310 references, from which 90 trials were selected. The most studied disease was multiple myeloma, with 29 studies, followed by non-Hodgkin lymphoma with 26. The others were acute myeloid leukemia 12, chronic lymphocytic leukemia 10, chronic myeloid leukemia 8, myelodysplastic syndromes 3 and Hodgkin lymphoma 2. The 90 studies had a total of 108 “primary” and 252 “secondary” outcomes; 20 studies (22%) had OS as primary endpoint (though only 3 of them reached statistical significance), in over 37 (41%) OS was grouped with other outcomes to form a composite endpoint. In 55 studies (61%) overall survival was a secondary outcome. Quality of life was a “secondary” outcome in 10 studies. CONCLUSIONS: Although OS is the gold standard in cancer therapy, grouped outcomes, or others intermediate outcomes, such as progression-free survival or paraclinical indicators of disease activity are more frequently used and may be good predictors. Intermediate outcomes require smaller sample sizes and less follow-up. Their capacity to predict OS (or, indeed, quality of life) should be carefully assessed. Only in the most aggressive forms of cancer is routine use of OS survival justified as the primary endpoint.