On 21 and 22 May 2001, the National Institute of Mental Health convened a workshop to explore imminent scientific opportunities and encourage new research on preventing relapse in major depression, as a part of a larger effort to find treatments capable of producing durable long-term recovery from major depression. Participants considered definitional and developmental perspectives on depression relapse, the prophylactic potential of current treatments and their cost effectiveness and the neurobiological and psychological risk factors for episode return. It was recommended that the definition of the relapse construct be expanded to capture salient features of incomplete recovery or partial response to treatment that are associated with significant functional impairment. This information is often overlooked by the categorical criteria currently in use. With respect to interventions, there was support for sequencing pharmacological remission with psychological prophylaxis. Provision of focal, short-term treatments that embed relapse prevention skills augment the routes to effective prevention available to patients, beyond that afforded by continuation pharmacotherapy. The challenge will be to identify those subgroups of patients for whom each treatment algorithm is indicated. Finally, the link between basic science findings of biological and psychological markers of relapse vulnerability and treatment design needs to be strengthened. This could be accomplished by assessing patients in clinical prevention trials for the presence of, and changes in, relapse vulnerability markers, thereby providing direct, outcome-based data to gauge the protective value of different treatments that modify these markers.