In this issue of the Journal of Clinical Oncology, Prieto et al report the results of a prospective study of the relationship between depression and mortality following hematopoietic stem cell transplantion (HSCT) for a malignant disease. They found that a diagnosis of major depression during the first 3 to 4 weeks of hospitalization for HSCT was associated with poorer survival at 1 and 3 years post-HSCT. This elevated mortality risk for depressed HSCT recipients, compared with nondepressed HSCT recipients, was evident even after they adjusted for a variety of demographic and clinical variables potentially linked to prognosis after HSCT. Superficially, these results are neither surprising nor innovative. As the authors note, depression as an independent risk factor for early mortality has been the focus of quite a bit of prior research using a variety of community and clinic populations. While the relationship between depression and mortality has been studied most extensively in the context of cardiovascular disease, this relationship has undergone scrutiny in the oncology setting in general, as well as the HSCT setting in particular. As the authors also note, results in the oncology setting have been mixed, with the majority of published studies supporting the hypothesized link between depression and poorer survival. The present study is thus largely consistent with prior research in the oncology setting, including research linking depression to poorer survival after HSCT. However, the present study by Prieto et al is noteworthy for its relatively superior methodology. This superiority is evidenced in several significant aspects of the research design and data analyses, including (1) remarkably high rates of study accrual and retention; (2) the use of formal DSM-IV (Diagnostic and Statistical Manual of Mental Disorders) diagnostic criteria for indexing the presence or absence of major depression; (3) serial assessment of depressive disorder at specific points in time ranging across the first 3 to 4 weeks of HSCT hospitalization; and (4) control in the data analyses for multiple demographic, clinical, and behavioral variables that might also be associated with post-HSCT survival, thus enabling more precise examination of the potential independent link between depressive disorder and post-HSCT survival. Despite its methodological strengths, however, the conclusions drawn by the present study are not quite definitive. To their credit, the authors acknowledge a number of limitations of their study, some of them quite serious. For example, the failure to account for concurrent or subsequent treatment for depression in those diagnosed with depressive disorders and the focus on depressive disorders present only during the first 3 to 4 weeks of HSCT hospitalization represent significant obstacles to a reliable and comprehensive explication of the depression-survival relationship in HSCT recipients. Other limitations are less well acknowledged. For example, one can question the adequacy of the procedures used to classify HSCT recipients with regard to the primary study variable: the presence or absence of depressive disorder. Specifically, the appropriateness of the use of modified DSM-IV criteria for diagnosing depressive disorder and the procedures for considering inconsistent diagnoses across serial assessments in determining whether a depressive disorder was present or absent could be debated by reasonable individuals. Finally, the relatively small sample size (n 199) and low prevalence of major depression (9%) make interpretation of null results problematic. In particular, although the authors attempted to identify variables that might interact with a diagnosis of major depression in predicting post-HSCT survival, none of the interaction terms were statistically significant. JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 23 NUMBER 25 SEPTEMBER 1 2005