Journal of Consulting and Clinical Psychology 2000, Vol. 68, No. 2, 226-232 Copyright 2000 by Ihc American Psychological Association, Inc. 0022-006X/00/S5.00 DOI: I0.1037//0022-006X.68.2.226 Long-Term Posttreatment Functioning Among Patients With Unipolar Depression: An Integrative Model Charles J. Holahan University of Texas at Austin Rudolf H. Moos Department of Veterans Affairs Health Care System, Palo Alto, and Stanford University Carole K. Holahan University of Texas at Austin Ruth C. Cronkite Department of Veterans Affairs Health Care System, Palo Alto, and Stanford University This study tested an integrative structural equation model of posttreatment functioning among 165 depressed patients followed for an average of 9 years after the end of an episode of treatment. The model examined (a) the link between life change and psychosocial resource change and (b) the role of resource change in mediating the relationship between life change and change in depression. An increase in the preponderance of negative over positive life events was associated with a decline in resources and an increase in depressive symptoms. A decline in resources was associated with an increase in depressive symptoms. The association between changes in events and depressive symptoms was completely mediated through resource change. These findings indicate that life stressors contribute to posttreatment depression through an erosion of personal and social resources. Most studies of the psychosocial aspects of depression have focused on the onset and treatment phases of depressive episodes. Much less is known about remission or the failure to recover after treatment. However, because 50% to 60% of depressed patients relapse after treatment (Nezu, Nezu, Trunzo, & McClure, 1998), long-term follow-up is essential for a full understanding of depres- sive disorders. The purpose of the present study was to test an integrative structural equation model of posttreatment functioning among 165 depressed patients followed for an average of 9 years after the end of an episode of treatment. Negative life events are related to poorer outcome of treatment for depression (Monroe, Kupfer, & Frank, 1992). For example, negative events during the year after hospitalization lengthened time to recovery among depressed patients (S. L. Johnson & Miller, 1997). Moreover, undesirable events after recovery pre- dicted more negative clinical outcomes up to 3 years postrecovery among individuals with recurrent depression (Monroe, Roberts, Kupfer, & Frank, 1996). Positive events during treatment have Charles J. Holahan, Department of Psychology, University of Texas at Austin; Rudolf H. Moos and Ruth C. Cronkite, Center for Health Care Evaluation, Department of Veterans Affairs Health Care System, Palo Alto, and Stanford University; Carole K. Holahan, Department of Kinesi- ology and Health Education, University of Texas at Austin. This work was supported by Department of Veterans Affairs Health Services Research and Development Service funds and by Grant AA06699 from the National Institute on Alcohol Abuse and Alcoholism. We gratefully acknowledge the assistance of Nancy Wonacott, Caryn Cohen, and Eric Berg in data analysis. Correspondence concerning this article should be addressed to Charles J. Holahan, Department of Psychology, University of Texas, Austin, Texas 78712. Electronic mail may be sent to holahan@psy.utexas.edu. been associated with better 1-year treatment outcome (J. G. John- son, Douglas, Han, & Russell, 1998). Psychosocial resources are related to better outcome for treat- ment of depression. More social support predicted fewer depres- sive symptoms among depressed outpatients at 6-month (Lara, Leader, & Klein, 1997) and 2-year (Sherbourne, Hays, & Wells, 1995) outcomes. Negative family relationships predicted poorer 1-year recovery among hospitalized depressed patients (Miller et al., 1992). Low personal resources (e.g., self-critical attitudes and low self-esteem) are also related to poorer remission among de- pressed individuals (Bothwell & Scott, 1997; Fairbrother & Mor- etti, 1998). A key limitation of existing research on posttreatment function- ing among depressed patients is a failure to consider more dynamic interrelationships between predictive factors. The few studies that have addressed this issue have been limited almost exclusively to nonclinical samples. For example, "support deterioration" (Kani- asty & Norris, 1993) and "wear and tear" (Park & Folkman, 1997) models have been proposed to explain how life stressors can deplete social resources. Life stressors may also be associated with a depletion of personal resources, such as hardiness (King, King, Fairbank, Keane, & Adams, 1998). An important reconceptualization of the stress process postu- lates that such resource loss operates as a mediator between life change and functioning (Kaniasty & Norris, 1993; Quittner, Glueckauf, & Jackson, 1990). King et al. (1998) found that the relationship between stressful events and posttraumatic stress dis- order among Vietnam veterans was partially mediated by a lack of hardiness and social support. Holahan, Moos, Holahan, and Cron- kite (1999) demonstrated that the association between changes in life events and functioning in a community sample was completely mediated through change in psychosocial resources.