Only ten percent of older adults with depression are estimated to receive treatment, in part due to low levels of detection by primary care clinicians. There is a growing body of evidence that depression among older adults may present in atypical ways, involving low levels of dysphoric mood, especially sadness. However, there has been little research to identify correlates or risk factors for these conditions of masked presentation. The compelling need for such research is suggested by recent findings that subclinical, intermediate-level symptoms of dysphoria which fail to meet DSMIV diagnostic criteria may be an important risk factor for future stroke episodes. The literature provides evidence that certain physiological conditions that occur with advancing age (vascular diseases, hypotension) are associated with three broad categories of """"""""depression with low sadness"""""""": vascular depression, depression related to systolic hypotension, and depression related to diastolic hypotension. Psychosocial factors (chronic life stress, social isolation) are believed to influence expression of depressive symptoms, although it appears that the proposed study would be the first to investigate this. The proposed study is based upon a conceptual model of co-occurrence, or comorbidity, of different forms of """"""""depression with low sadness"""""""" associated with underlying physiological conditions. The purpose of the proposed study is to investigate the etiology of these masked forms of depression with the aims of: 1) clarifying distinct patterns of items for depressed affect, low positive affect, and other depressive symptoms; and 2) identifying age-related and depression-specific correlates of these different patterns. A methodologically innovative use of latent trait analysis (MIMIC) will involve interactions between multiple covariates to model the endorsement of specific symptoms by well-focused subgroups, while simultaneously adjusting for the level of overall depression. The proposed study is based on secondary analysis of the first wave from the New Haven site of the EPESE community survey. The empirical model will be replicated on targeted sub-samples (i.e., blacks, white females, white males).
