In evaluating the risk of recurrence in MDD, some generalizations can be made that apply to all patients, and some predictors of risk inform the decision for an individual patient. Although earlier data suggested that half of patients with unipolar depression may have only one episode in their lifetime, more recent prospective studies indicate much higher recurrence rates. For example, the National Institute of Mental Health’s Collaborative Depression Study found a 15-year recurrence rate of 87%. However, this estimate included all patients enrolled in the study—those with numerous prior episodes as well as those with relatively few. Predictors of recurrence from a variety of studies include early onset of illness (before age 21), later onset of illness (after age 60), multiple prior depressive episodes, long duration of individual episodes, the presence of interepisode dysthymia (double depression), and incomplete resolution of the previous depressive episode. Of these predictors, the most important is the number of prior episodes. With each subsequent depressive episode, the likelihood of another episode in the future increases.