We addressed three questions in this study: (a) does 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 reductions in depressive symptoms; and (c) is this association stronger for CBASP than for the two comparison treatments conditions? We found that patients who received CBASP plus pharmacotherapy exhibited significantly, greater gains in social problem solving than patients who received BSP plus pharmacotherapy. In addition, there was a trend for patients in the CBASP condition to exhibit greater improvement in social problem solving than patients receiving medication alone. Examining specific forms of social problem solving, patients receiving CBASP plus medication exhibited significantly greater increases in rational problem solving and positive problem orientation than patients in either of the other two conditions. In addition, patients receiving BSP plus medication exhibited significantly less change in avoidant problem solving than patients in the other two conditions, and significantly less change in impulsive problem solving than patients receiving CBASP plus medication.
The differences in change in social problem solving tended to be greatest between the two psychotherapy augmentation conditions, with the pharmacotherapy alone condition occupying an intermediate position. This raises the possibility that BSP may have hindered some aspects of problem solving, perhaps by reinforcing emotion-focused coping strategies. However, this should not be overinterpreted, as BSP plus pharmacotherapy did not differ from pharmacotherapy alone on change in the total SPSI-R.
The present findings are consistent with prior studies that found problem-solving therapy was associated with greater improvement in social problem solving than comparison conditions, such as supportive therapy and being on a waiting-list (Alexopoulos et al., 2003
; Nezu, 1986
; Nezu et al., 1989
; Nezu et al., 2003
; Sahler et al., 2002
). Importantly, this study extended past work by examining a more severe and chronic sample than most previous studies.
We also found that gains in social problem solving predicted subsequent reductions in depressive symptoms over time. Several other studies reported an association between change in social problem solving and change in depression (Alexopoulos et al., 2003
; Nezu et al., 2003
; Sahler et al., 2002
). However, unlike the present study, these studies assessed both variables only at baseline and follow-up, hence the direction of the association could not be determined.
As CBASP plus medication predicted greater change in social problem solving and gains in social problem solving predicted declines in depression over time, one might expect that patients in the CBASP condition would have better outcomes than patients in the comparison conditions. However, as reported in our previous article (Kocsis et al., 2009
) and confirmed in the analyses in the present paper, the three treatment conditions did not differ on depression outcomes. This pattern of findings suggests that different treatments influence depressive symptoms through different processes. If so, and social problem solving plays a greater role in alleviating depression in CBASP than other treatment approaches, then one might expect that change in social problem solving would be closely linked to change in depressive symptoms over time in the CBASP condition. Contrary to our hypotheses, however, the interaction of treatment condition with the association between social problem solving and lagged depression scores did not approach significance. Instead, the magnitude of the relation between change in social problem solving and subsequent depression was similar regardless of whether or not patients received CBASP.
Taken together, this pattern of findings suggests that while CBASP plus medication may produce greater gains in social problem solving than BSP plus medication and, at a trend level, pharmacotherapy alone, and that better social problem solving is associated with subsequent improvements in depressive symptoms, this mechanism may be common across treatments. That is, to the extent that any treatment positively influences social problem solving, a decline in depressive symptoms is likely to follow. However, it important to consider that possibility that problem solving does play a somewhat greater role in producing change in CBASP, but that despite our large sample, we did not have sufficient power to detect this effect. A more thorough dismantling of the specific aspects of CBASP that enhance problem solving and consideration of the ways in which the intervention might be augmented to strengthen its effect may be useful, given that problem solving appears to be an important, and teachable, aspect of coping. Alternatively, although there is considerable convergence between the processes targeted in CBASP and the constructs assessed by the SPSI-R, it is conceivable that a problem-solving measure that was specifically designed for CBASP might have greater sensitivity.
This study had a number of strengths, including a large, carefully characterized sample and multiple assessments of social problem solving and depressive symptoms over time. However, several limitations should be considered. First, CBASP is an integrative treatment. Although training in interpersonal problem solving is a central component, CBASP also addresses long-standing maladaptive interpersonal patterns, dysfunctional cognitions, social skills deficits, and problems in the therapeutic relationship (McCullough, 2000
). Hence, it is conceivable that the results would differ for treatments that focus more narrowly on social problem solving. Second, treatment was limited to 16–20 sessions over 12 weeks. Although this was more intensive than most problem-solving therapies (D’Zurilla & Nezu, 1999; Mynors-Wallis et al., 1995
), the patients in this sample were considerable more severe and chronic than in most previous problem-solving trials. Hence, a longer duration of treatment might be needed to isolate specific therapeutic mechanisms for this population. Third, all patients in the CBASP condition received concomitant pharmacotherapy and had previously experienced at least one unsuccessful medication trial. This may have adversely influenced some patients’ motivation to fully engage with CBASP and situational analysis, attenuating its effects on social problem solving. Fourth, assessments were conducted biweekly, and we examined only a two-week lag between social problem solving and depressive symptoms. It is possible that the effects of social problem solving on depressive symptoms unfold over much shorter or longer intervals, reducing the sensitivity of our analyses. Fifth, we used a self-report inventory to assess social problem solving. Although the SPSI-R is the most widely used and best-validated social problem-solving measure, other approaches, such as performance-based measures or experience sampling methods, might conceivably yield different results. Finally, patients were treated at academic centers, hence may not be entirely representative of chronically depressed patients in treated in community settings.
In conclusion, in a large, randomized clinical trial of chronically depressed patients, we found that CBASP in conjunction with medication produced significantly greater improvement in social problem solving than BSP plus medication, and a similar trend compared to pharmacotherapy alone. In addition, change in social problem solving predicted subsequent change in depressive symptoms over time. However, the magnitude of the associations between changes in social problem solving and subsequent depressive symptoms did not differ between treatment conditions. Hence, these findings fail to support the hypothesis that improved social problem solving, as measured by the SPSI-R, is a mechanism that uniquely distinguishes CBASP, despite its emphasis on addressing social problem-solving deficits, from other treatment approaches.