پیگیری بلندمدت آزمایش کنترل شده تصادفی در مقایسه با درمان مبتنی بر پذیرش و تعهد و استاندارد رفتار درمانی شناختی برای اضطراب و افسردگی

The present study represents one of the first comparisons of the long-term effectiveness of traditional cognitive behavior therapy (i.e., Beckian cognitive therapy; CT) and acceptance and commitment therapy (ACT). One hundred thirty-two anxious or depressed outpatients were randomly assigned to receive either CT or ACT, and were assessed at posttreatment (n = 90) and at 1.5-year (n = 91) follow-up. As previously reported, the two treatments were equivalently effective at posttreatment according to measures of depression, anxiety, overall (social/occupational/symptom-related) functioning, and quality of life. However, current results suggest that treatment gains were better maintained at follow-up in the CT condition. Clinical significance analyses revealed that, at follow-up, one-third more CT patients were in the clinically normative range in terms of depressive symptoms and more than twice as many CT patients were in the normative range in terms of functioning levels. The possible long-term advantage of CT relative to ACT in this population is discussed.

The term cognitive behavioral therapy (CBT) reflects a broad collection of evidence-based approaches that have become the most widely utilized and researched of all psychotherapeutic methods (Norcross, Hedges, & Castle, 2002), with Beckian cognitive therapy (CT; Beck, 1991) representing the most widely used and empirically supported form of CBT (Butler et al., 2006 and Hofmann and Smits, 2008). A newer subcategory of CBT, sometimes referred to as acceptance-based behavior therapies, has risen to prominence in recent years. Examples include mindfulness-based cognitive therapy (MBCT; Z. V. Segal, Williams, & Teasdale, 2002), mindfulness-based stress reduction (Kabat-Zinn, 1990), acceptance-based behavior therapy for generalized anxiety disorder (Roemer & Orsillo, 2005), dialectical behavior therapy (DBT; Linehan, 1993), and acceptance and commitment therapy (ACT; Hayes, Strosahl, & Wilson, 1999), among others. Of these, ACT has received the most attention in terms of empirical study (Hayes et al., in press and Hayes et al., 2006) and scientific debate (Arch and Craske, 2008, Corrigan, 2001, Corrigan, 2002, Gaudiano, 2009a, Gaudiano, 2009b, Hayes, 2002, Hayes, 2008, Hayes et al., in press, Herbert and Forman, in press, Herbert and Forman, 2011, Hofmann and Asmundson, 2008, Hofmann and Asmundson, in press, Öst, 2008 and Öst, 2009). At the level of technology, there are some important differences in how ACT and CT treat psychopathology (Forman & Herbert, 2009). CT makes use of cognitive disputation and other forms of reappraisal (including behavioral experiments) designed to correct systematic biases in information processing, with the goal of reducing symptom intensity (Beck, Rush, Shaw, & Emery, 1979). The goal of ACT is not symptom reduction per se, but helping patients to “accept” difficult internal experiences (thoughts, images, emotions, sensations) in the service of engaging in values-consistent behavior change. ACT has demonstrated preliminary effectiveness across a range of problem behaviors, including mood (Zettle & Hayes, 1986) and anxiety (Block, 2003; L. A. Brown et al., 2011, Dalrymple and Herbert, 2007, Roemer et al., 2006 and Twohig et al., 2006) disorders, psychosis (Bach and Hayes, 2002 and Gaudiano and Herbert, 2006), polysubstance abuse (Hayes, Wilson, et al., 2004), and smoking (R. A. Brown et al., 2005, Hernández-López et al., 2009 and Hernandez Lopez et al., 2005), among others. A meta-analysis (Hayes, et al., 2006) reported ACT to be superior to active treatments, including standard CBT. However, Öst (2008) has criticized the rigors of the trials on which the meta-analysis relied, and a subsequent meta-analysis (Powers, Zum Vörde Sive Vörding, & Emmelkamp, 2009) concluded that ACT was equally effective as established treatments (but see also Gaudiano, 2009a). The RCT upon which the current study was based similarly detected no differences in efficacy between ACT and CT at posttreatment in the treatment of depression and anxiety (Forman, Herbert, Moitra, Yeomans, & Geller, 2007). It is important to examine longer-term effectiveness of psychotherapies, as immediate effects can fade over time, patients might take time to master skills learned in treatment, and treatments that are equivalent at one time point can diverge later (Gifford et al., 2004 and Lappalainen et al., 2007). Generally speaking, standard CBT has demonstrated longer-term efficacy (Butler et al., 2006, Gloaguen et al., 1998 and Shapiro et al., 1994), though several recent reviews suggest effects weaken considerably after 1 year (e.g., Durham, Higgins, Chambers, Swan, & Dow, 2011). Open trials and trials comparing ACT to nonactive treatments or treatment-as-usual support ACT's lasting benefits for anxiety (Dalrymple and Herbert, 2007, Ossman et al., 2006, Twohig, 2008 and Zettle, 2003), depression (Blackledge & Hayes, 2006), trichotillomania (Woods, Wetterneck, & Flessner, 2006), psychosis (Bach and Hayes, 2002 and Gaudiano and Herbert, 2006), substance dependence (Hayes et al., 2004b, Stotts et al., 2009 and Twohig et al., 2007), smoking (Gifford et al., 2004), obesity (Forman et al., 2009, Lillis et al., 2009 and Tapper et al., 2009), and chronic pain (Dahl et al., 2004, McCracken et al., 2007 and Vowles and McCracken, 2008). However, the follow-up periods of these studies tended to be relatively short (i.e., 1 to 3 months), limiting the extent to which conclusions can be drawn about longer-term effects. Moreover, few of these trials compared ACT to another well-established, active intervention, and none compared ACT to traditional CBT. The current study compares long-term (18-month follow-up) outcomes from a group of outpatients randomly assigned to receive either CT or ACT. The study is an extension of an earlier report of posttreatment outcomes (Forman, Hebert, et al., 2007). As mentioned, the effects of standard CBT may attenuate in the longer-term (Durham et al., 2011), and some proponents have hypothesized that ACT might hold certain advantages over standard (i.e., Beckian) CBT because ACT is argued to be more tightly linked to basic research on its mechanisms and underlying theory (Hayes, 2008 and Hayes et al., in press). ACT proponents have also questioned the putative mechanisms of change of CT (i.e., modification of the content of dysfunctional cognitions; Hayes, 2008 and Hayes et al., 2011). On the other hand, standard CBT has proven long-term efficacy (Butler et al., 2006, Gloaguen et al., 1998 and Shapiro et al., 1994). Perhaps an overriding consideration is the accumulating evidence suggesting that it is the behavioral elements of treatment that represent the mechanisms of action, and that other components are superfluous (Dimidjian et al., 2006 and Longmore and Worrell, 2007). Thus, no specific hypotheses were made regarding differential long-term effectiveness of the two treatments. Given previous findings that mindfulness and acceptance variables moderate the impact of treatment, we tentatively hypothesized such a moderation effect at follow-up.

Key variables, including primary diagnostic category, χ2(1) = .69, p = .88., and days to follow-up (t = -.20, p = .84), were equivalent by assignment to condition. In addition, there were no significant differences between those who provided follow-up data (n = 91) and those who did not on key demographic or outcome variables, but those diagnosed with an anxiety disorder were more likely to have completed the follow-up assessment, χ2(3) = 8.11, p = .04. A roughly equal percentage of ACT (n = 67%) and CT (n = 71%) participants completed follow-up, χ²(1) = 0.35, p = .56. To probe yet further, a series of analyses were performed in order to determine if the follow-up status (i.e., whether or not a participant was available at follow-up), the treatment condition, or the interaction of follow-up status and condition was related to key demographic, comorbidity, and outcome variables. Logistic regression and ANOVAs indicated that neither follow-up status nor treatment condition nor their interaction was related to baseline characteristics of the sample ( Table 1 and Table 2).