Social problem solving and depressive symptoms over time: a randomized clinical trial of cognitive-behavioral analysis system of psychotherapy, brief supportive psychotherapy, and pharmacotherapy.

Numerous studies have documented associations between social problem solving and depressive disorders and symptoms. Depressed individuals often exhibit a negative orientation toward problems in living (e.g., appraising a problem as a threat, doubting one’s own problem-solving ability) and deficits in specific problem-solving skills on self-report inventories and performance-based measures (Dixon et al., 1993; Haaga et al., 1995; Kant et al., 1997; Marx et al., 1992; Nezu et al., 1986; Reinecke et al., 2001). Hypothesizing that ineffective problem solving plays a role in the etiology and maintenance of depression, several investigators have developed treatments for depression that specifically target social problem solving (D’Zurilla & Nezu, 2007; Mynors-Wallis et al., 1995). The efficacy of social problem-solving therapies has been examined in several dozen clinical trials, generally with favorable results (e.g., Arean, et al., 2010; Barrett et al., 2001; Mynors-Wallis et al., 1995, 2000; Nezu, 1986; Nezu & Perri, 1989; Williams et al., 2000). In two recent meta-analyses, social problem-solving therapies for depression had a moderate effect size (Bell & D’Zurilla, 2009; Cuipers et al., 2007). The presumptive mechanism through which problem-solving therapies alleviate depression involves their effects on social problem solving. Indeed, several studies have reported that problem-solving therapies produce greater change in social problem solving than comparator treatments and being on a wait-list (Alexopoulos et al., 2003; Nezu, 1986; Nezu et al., 1989; Nezu et al., 2003; Sahler et al., 2002; also see Arean et al. [1993], who found no significant treatment condition by time interaction, but reported that only problem-solving therapy was associated with significant gains in social problem solving ). Three of these studies also attempted to test mediation more explicitly. In a sample of distressed cancer patients, Nezu et al. (2003) reported that change in social problem solving correlated with change in symptoms. In geriatric patients with major depressive disorder (MDD) and deficits in executive functioning, Alexopoulos et al. (2003) found a significant interaction between the effects of treatment condition and improvement in social problem solving on change in depression, which they interpreted as evidence of mediation. Finally, using path models in a sample of mothers of children with recently diagnosed cancer, Sahler et al. (2002) reported that change in social problem solving partially mediated the association between problem-solving skills training and change in negative affect. Unfortunately, in all of these studies data on social problem solving and symptoms were only available at pre- and post-intervention. Hence, it was not possible to determine whether change in social problem solving preceded, followed, or occurred contemporaneously with change in symptoms. Thus, although the extant literature suggests that social problem solving may mediate the antidepressant effects of problem-solving therapies, the specialized nature of the samples and the limited numbers of assessment points do not permit firm conclusions. The present paper, from the Research Evaluating the Value of Augmenting Medication with Psychotherapy (REVAMP) trial (Kocsis et al., 2009), examines whether the Cognitive Behavioral Analysis System of Psychotherapy (CBASP; McCullough, 2000), an approach that targets interpersonal problem solving in chronic depression, influences depressive symptoms via its effects on social problem solving. This study extends previous investigations by assessing social problem solving and depressive symptoms biweekly throughout treatment, and using time-lagged analyses within a mixed models framework to test whether changes in problem solving predict subsequent depression over time. In addition, we examine whether the associations over time between social problem solving and subsequent depressive symptoms are greater for patients receiving CBASP plus medication than patients receiving Brief Supportive Psychotherapy (BSP) plus medication and patients receiving medication alone. The larger REVAMP trial tested the efficacy of a stepped treatment strategy in which CBASP or BSP was added to pharmacotherapy switch or augmentation in chronically depressed patients who failed to fully respond to an initial trial of medication. As chronic depression is characterized by poor coping and social adjustment (Klein & Leader, 1996; McCullough et al., 1990; Miller et al., 1998), a problem-solving intervention may have particular relevance for this population (McCullough, 2000). As described elsewhere (Kocsis et al., 2009), the three treatment conditions (CBASP plus pharmacotherapy, BSP plus pharmacotherapy, and pharmacotherapy alone) produced similar change in depressive symptoms. Hence, although our analyses explore whether the treatment conditions influence depressive symptoms through different mechanisms, they differ from traditional tests of mediation; rather than determining whether change in social problem solving accounts for differential treatment effects, we examine whether the magnitude of the associations over time between social problem solving and subsequent depressive symptoms differ between treatment conditions. We address three specific questions: (a) do CBASP plus pharmacotherapy produce greater change in social problem solving than BSP plus pharmacotherapy and pharmacotherapy alone; (b) is social problem solving associated with subsequent depression over time; and (c) is this association stronger for CBASP than for each of the two comparison treatments conditions?