CBT theories of mood disorders move beyond description to explain and predict depressive phenomena. We cannot do full justice to CBT theories of depression here and interested readers are referred to recent reviews (see: A. T. Beck, 1996; Ingram et al., 1998; Clark et al., 1999). In brief, CBT theories of mood disorders are based on several assumptions. First, a diathesis-stress biopsychosocial model is implicated in the development and maintenance of emotional disorders. That is to say, biological, psychological, and social factors can all be involved in both diathesis, predisposing someone to mood disorders, as well as acting as stressors that precipitate the onset/relapse of mood disorders. Second, maladaptive beliefs about the self, the external world and the future are shaped through formative developmental experiences. Third, these maladaptive beliefs lie dormant and are activated only when precipitated by resonant situations. Fourth, when precipitating situations occur, the beliefs interact with the situation through processes of selective attention and inference, and generate negative mood reactions. These negative beliefs and emotions lead to behavioral consequences that serve to maintain negative mood (A. T. Beck, 1976b; A. T. Beck et al., 1979). Figure 11.1 shows this original model, and provides an illustration by describing a typical situation–belief–emotion–behavioral cycle for Sheryl.

This basic cognitive model has been significantly elaborated and refined on the basis of over 30 years of empirical work (see: Clark et al., 1999). One significant refinement is the suggestion that maladaptive cognitive processing, including negative beliefs, becomes activated only after the onset of depressive mood problems when dysphoric states are present (Miranda and Persons, 1988; Teasdale and Cox, 2001). In the onset and maintenance of mood disorders, depression is fuelled by a stream of negative ruminative automatic thoughts (e.g., ‘My high functioning façade is breaking down,’ ‘My family think I am weak,’ ‘I will be unable to cope with a family Christmas’) that are congruent with underlying higher-order modes (e.g., ‘self-as-weak’) and dysfunctional assumptions (e.g., ‘If my high functioning façade breaks down people will think I am weak’).

A second significant refinement is the concept of core modes that become activated in depression. Core modes are interlocking information processing systems that draw on the parallel processing from cognitive, affective, and sensory processing modules (Teasdale and Barnard, 1993; A. T. Beck, 1996). Once instated in depression, these core modes have a self-maintaining property as mode-consistent biases of attention, overgeneralized memories, higher-order self-schemas, ruminative thinking, and sensory feedback loops from unpleasant bodily states ‘interlock’ in self-perpetuating cycles of processing. The more often a person has suffered depression, the more easily these core modes become automatic and easily activated (Segal et al., 1996). The content of depressive core modes tends to be organized around themes of loss, defeat, failure, worthlessness, and unloveability.

Several theoretical reformulations argue that core modes are directly linked to depressive affective and motivational symptoms (Teasdale et al., 1993; Power and Dalgleish, 1997), while lower order maladaptive beliefs are linked to depression only indirectly through the core modes. Maladaptive beliefs are secondary dysfunctional assumptions (e.g., ‘I have to put on a strong façade, or people will reject me’), rules for living (e.g., ‘don't show weakness’), and attitudes (e.g., ‘weak people are pathetic’) that are closely linked to core modes (e.g., ‘self-as-weak’). Various commentators have noted this distinction between higher-order self-schemas and lower-order maladaptive beliefs as reflecting the difference between emotional and intellectual belief, ‘hot’ and ‘cold’ cognition (Teasdale, 1993; J. S. Beck, 1995; Young et al., 2003), or as clients have described it to us ‘I know it in my gut rather than in my head.’

Fig. 11.1 Illustration of basic cognitive model of depression, with case example.

Related to core modes are cognitive and behavioral compensatory strategies that enable a person to cope with the negative consequences of core modes. The strategies are usually part of a spectrum of normal coping strategies but have become problematic because they have become inflexible and therefore inadvertently maintain core modes and maladaptive beliefs. In their most primitive form perceived threat triggers compensatory strategies to mobilize for action or inhibit into inaction. Examples of dimensions of compensatory strategies include:

• avoid intimacy—appropriate intimacy—overly intimate

• passive-aggressive—appropriate assertiveness—aggressiveness

• abdicate control to others—appropriate use of control—authoritarianism

A table and schematic diagram summarize the reformulation of depression (Table 11.1 and Figure 11.2).

Activating events (internal or external) activate orienting schemas, which in turn activate the patterns of cognitive processing (dynamic cognitive structures) that are core modes or interlocked ‘minds-in-place.’ The cognitive features that make up depression (e.g., ruminative thinking, negative appraisals, memory biases) are produced once a negative core mode is instated. The characteristics and relationship between core modes, maladaptive beliefs, and compensatory strategies are shown in Table 11.1 and Figure 11.3.

In parallel with the emphasis on compensatory strategies in cognitive accounts, recent behavioral conceptualizations of depression have highlighted the importance of avoidant behaviors in depression, particularly within the behavioral activation (BA) approach. BA was initially developed as part of a component analysis of the active components of CBT, and only reflected the behavioral components of CBT (Jacobson et al., 2001; Hopko et al., 2003). After BA was found to be as effective as BA plus thought challenging and as effective as full CBT in treating major depression (Jacobson et al., 1996), the treatment was further elaborated (see: Martell et al., 2001), drawing on behavioral approaches to depression (e.g., Ferster, 1973). Central to the BA conceptualization of depression is the concept of secondary avoidant behaviors in response to the symptoms of depression produced by negative events: BA proposes that unhelpful secondary coping responses lead to the maintenance of depression. Typically, unhelpful secondary coping responses involve attempts to escape from an aversive environment (e.g., arguments, confrontations, reminders of loss) or to avoid aversive situations or emotional states (risk of failure or embarrassment), that is, secondary coping responses are compensatory strategies. Avoidance behaviors include being passive, withdrawal, rumination, complaining, or avoiding new activities. Because these behaviors reduce exposure to aversive situations they are negatively reinforced and become more prevalent, reducing the frequency and narrowing the range of other behaviors, which in turn reduces contact with positive reinforcers and increases the risk for depression.

Fig. 11.2 Cognitive model of depression.

Fig. 11.3 Schematic diagram of cognitive reformulation of depression.

As in the original cognitive model for depression, cognitive approaches to bipolar disorder emphasize a diathesis-stress biopsychosocial model and focus on the importance of maladaptive beliefs and automatic thoughts. Although a comprehensive cognitive model of bipolar disorder is yet to be delineated, recent attempts to develop cognitive behavioral approaches for bipolar disorder have all focused on evidence suggesting that psychosocial stressors and adverse cognitive styles interact with an inherent biological vulnerability to produce manic and depressed episodes (Newman et al., 2002). The biological vulnerability appears in part to be hereditary, with bipolar disorder running in families, and genetic factors demonstrated in twin and adoption studies. Recent theories have suggested that biological vulnerability to bipolar disorder may result from dysregulation in the BA system, which is a putative neurobiological motivational system that regulates goal-directed approach to potential reward and is proposed to influence positive affect, energy, and attention.

Other evidence suggests that bipolar episodes occur in response to stressful life events, whether disruptions in daily routines (Malkoff-Schwartz et al., 1998), negative life events, or even goal attainment. Typically, it appears that negative life events predict bipolar depression, while goal attainment predicts mania; however, the relationship is not always straightforward, with negative events producing mania in the context of increased BA.

Cognitive-behavioral approaches to bipolar disorder emphasize that an individual's cognitive style and behavioral coping strategies in response to stressful life events mediates the extent to which the biological vulnerability is expressed in a full-blown bipolar episode. Consistent with this, Reilly-Harrington et al. (1999) report that negative attributional styles increase vulnerability to both manic and depressive symptoms following stressful life events.

Similarly, the response to prodromes of mania and depression is considered an important factor in the cognitive-behavioral model. Prodromes are the early signs and symptoms that can precede a full-blown episode, which the cognitive model assumes to be activated by the interaction between stressful life events and biological vulnerability. For example, disruptions in daily routine coupled with biological vulnerability may lead to reduced sleep. The cognitive-behavioral approach to bipolar disorder predicts that the particular thinking patterns and coping strategies instated in response to prodromal symptoms will determine whether a full bipolar episode will occur. As in models of unipolar depression, particular patterns of affect, cognition, and sensory input (e.g., depressed mood or hypomania) are hypothesized to activate associated schemas or core modes, which, in turn, will bias information processing towards information consistent with the schema, further fuelling the mood states. Thus, a patient in a hypomanic state will have positively valenced schemas activated, which will be characterized by processing that emphasizes goal attainment and potential rewards, while downplaying potential risks and problems. The particular schemas and modes that are activated in response to stress are hypothesized to determine which life events an individual will be more vulnerable to and to determine what form their response to stress will take. The activation of more adaptive, less extreme schemas and modes will lead to more stable mood, while less adaptive, more extreme schemas will produce further emotional dysregulation. For example, extreme beliefs about need for achievement and autonomy would be hypothesized to increase risk for depression and/or mania in patients with a biological vulnerability when exposed to potentially triggering events. In the case of perceived failure, such beliefs may lead to low self-worth and depressed mood, while in the context of perceived success such beliefs may lead to more grandiose thoughts about the self, feeding into hypomania.

Schemas and modes will also determine the strategies chosen to respond to stressful events and prodromes, e.g., achievement-related schemas would lead to overdriven behavior to compensate for lost time. Helpful coping strategies act against the prevailing prodrome, e.g., reduced arousal in hypomania, whereas unhelpful coping strategies further reinforce the initial stages of the bipolar episode, fuelling more extreme mood swings, e.g., rushing around doing many things at once (Lam et al., 2001). The specific schemas and modes available to be activated in any individual by stressful events or prodromes will depend upon his or her early learning history, as well as upon experiences in adolescence and adulthood, often linked to the onset and consequences of the bipolar disorder (e.g., beliefs such as ‘I am a difficult person’ and ‘I am defective’ following from the emotional fallout of mood swings).

Cognitive models of bipolar disorder also highlight the self-fulfilling nature of the disorder, with the consequences of a bipolar episode further contributing to the maintenance of the episode. For example, impulsive spending may lead to financial problems, irritability coupled with poor concentration may lead to problems at work or the loss of employment and promiscuous behavior may lead to problems in intimate relationships. All of these episode-related difficulties could then act as further stressors to interact with the underlying biological vulnerability to further generate bipolar symptoms. Furthermore, bipolar disorder is associated with a great deal of loss (e.g., lost potential, lost employment prospects, lost relationships), self-blame for impulsive acts committed during mania, and stigma, which can act as further stressors and/or further reinforce dysfunctional beliefs.

Thus, in summary, cognitive models of bipolar disorder emphasize: (1) underlying biological vulnerability (emotional dysregulation) and underlying cognitive vulnerability (dysfunctional schemas); (2) that these vulnerabilities interact with stressful life events to determine prodromes and patient's responses to prodromes; (3) less adaptive schemas will lead to less adaptive coping strategies and the exacerbation of prodromes into full-blown episodes; and (4) The consequences of episodes further exacerbate stressful life events and underlying cognitive vulnerabilities.

The advantages of these theoretical formulations to the CBT practitioner are that they introduce a more integrative model of mood disorders that provides clear rationales for why and how a broad range of CBT interventions might impact on cognition, behavior, and affect.

CBT has been demonstrated to be a generally effective treatment for depression in the large number of studies that have accumulated since the original study by Rush et al. (1977). CBT produces a greater improvement in symptoms than no treatment or waiting-list controls (Dobson, 1989) and demonstrates equivalent efficacy to pharmacotherapy for depression, although many studies did not employ a drug–placebo control condition or monitor plasma medication to check on the adequacy of pharmacotherapy (e.g., Blackburn et al., 1981; Hollon et al., 1992; Blackburn and Moore, 1997).

What about CBT compared with other psychotherapies? The large multisite National Institute of Mental Health (NIMH) Treatment of Depression Collaborative Research Program (TDCRP) trial (Elkin et al., 1989), which compared CBT, interpersonal psychotherapy, imipramine, and a placebo control, found that although there were few significant differences between treatments, for more people with more severe depression, pharmacotherapy and interpersonal therapy did better than CBT, with CBT only doing as well as placebo control on several outcome measures. This result has been much debated, with questions about differences in the skill in application of CBT across sites. Other studies comparing CBT with interpersonal or psychodynamic therapies for depression found CBT as effective as psychodynamic/interpersonal (PI) therapies (see: Leichsenring, 2001). However, the people in the NIMH-TDCRP study tended to have more severe depression than the other studies. More recent evidence attests to the efficacy of CBT for people diagnosed with depression, across a wide range of depression severity (DeRubeis et al., 1999).

One randomized controlled trial (RCT) compared CBT with BA (Jacobson et al., 1996; Gortner et al., 1998). The BA component of CBT focused on monitoring daily activities, assessment of pleasure and mastery, graded task assignment, cognitive rehearsal, problem solving, and social skills training. There were no significant differences between BA, BA plus modification of automatic thoughts (AT) and a full CBT treatment, at completion of treatment, 6-month follow-up (Jacobson et al., 1996), or 2-year follow-up (Gortner et al., 1998).

How well does CBT work for more chronic and severe depression? A recent trial examined CBT with and without nefazodone for chronic depression, operationalized as major depression lasting at least 2 years or a current major depression superimposed on preexisting dysthymia (Keller et al., 2000). This version of CBT, Cognitive-Behavioral Analysis System of Psychotherapy (CBASP) differs from classical CBT in its explicit focus on the consequences of client's interpersonal behavior through the use of a situational analysis protocol, which helps clients to identify whether their expectations and behaviors help or hinder movement towards their goals (see: McCullough, 2000). This trial found that CBASP and nefazodone in combination produced more remission in chronic depression (48%) than either nefazodone (29%) or CBASP alone (33%) (Keller et al., 2000). One limitation of this study was that treatment-resistant participants, that is, people who had not responded to previous antidepressants or psychotherapy, were excluded, i.e., the study lacked an important subgroup of chronic depression. Nonetheless, if replicated, this study would suggest that a combination of CBT and pharmacotherapy may be most appropriate for chronic depression.

One potential benefit of CBT for depression is that it reduces relapse/recurrence to a greater extent than antidepressant medication. Given that recurrence is a significant problem for people with major depression (Judd, 1997a,b), treatments that reduce relapse/recurrence are urgently needed. A number of studies report that after 1 or 2 years follow-up, relapse rates following treatment for depression with CBT were lower than for people treated with pharmacotherapy, when both treatments are stopped at termination (Kovacs et al., 1981; Simons et al., 1986; Evans et al., 1992; Shea et al., 1992; Gortner et al., 1998) (see also meta-analysis: Gloaguen et al., 1998). However, interpretation of these findings needs to be cautious because different studies used different criteria for relapse; Beck Depression Inventory scores greater than 16 or ‘treatment reentry for depression’ (Kovacs et al., 1981; Simons et al., 1986), compared with fulfilling criteria for major depression (Shea et al., 1992). Clearly, diagnosis of major depression is the most stringent criterion, while reentry into treatment is problematic as people in the CBT condition may still be symptomatic but attempting to deal with their symptoms themselves. Furthermore, an important comparison group is antidepressant continuation; people maintained on antidepressant appear to do as well as people who received a brief course of CBT (Evans et al., 1992). Blackburn and Moore (1997) in a randomized acute trial for recurrent major depression compared acute antidepressant treatment followed by maintenance antidepressants, acute CBT treatment followed by maintenance CBT and acute antidepressant treatment followed by maintenance CBT. All three groups showed clinical improvements during the acute and maintenance phases of treatment, with no significant differences between the three groups. CBT is therefore as effective in preventing the recurrence of depression as continued antidepressant medication.

More recently, several trials have specifically investigated the role of CBT treatments as relapse prevention for people whose depression was in remission rather than as an intervention for current depression. Fava et al. (1994, 1996, 1998) have developed a version of CBT to be used after successful treatment of an acute episode by pharmacotherapy. This therapy protocol involves a combination of CBT focused on residual symptoms of depression, life-style modification, and well-being therapy. Residual symptoms of depression are known to predict increased risk of relapse and therefore targeting such symptoms may well help reduce future episodes of depression. RCTs suggest that CBT for residual depression results in significantly less relapse/recurrence over 2 years (25%) than standard clinical management in the absence of antidepressant medication (Fava et al., 1998). Paykel et al. (1999) further demonstrated that compared with clinical management alone, clinical management plus CBT reduced relapse in 158 people with recent major depression that had partially remitted with antidepressant treatment.

An alternative approach to preventing relapse/recurrence has specifically targeted people with a history of recurrent depression who are currently in remission. Based on the hypothesis that these people tend to be caught up in ruminative depressive processing at times of potential relapse/recurrence, Teasdale et al. (1995) proposed that using mindfulness meditation, which fosters a relationship to thoughts and feelings antithetical to such rumination, might prevent future episodes of depression. Therefore, elements of a mindfulness-based stress reduction program (Kabat-Zinn, 1990) were incorporated into CBT to create mindfulness-based cognitive therapy (MBCT). MBCT is delivered in weekly group training sessions, in which participants practice and develop a moment-by-moment nonjudgmental awareness of sensations, thoughts, and feelings, through the use of formal and informal meditation exercises. These awareness exercises are further practiced during homework (see: Segal et al., 2002). For people with a history of three or more episodes of major depression, MBCT significantly reduced risk of relapse/recurrence over 1 year compared with treatment as usual (Teasdale et al., 2000). Without a further component trial, it is not possible to determine whether it was the mindfulness element or the CBT element or the combination thereof that was effective in this treatment.

In recent years, several RCTs have shown that compared with standard clinical management (including the prescription of mood stabilizers), standard clinical management plus CBT can reduce the recurrence of future bipolar episodes in people with bipolar disorder (e.g., Perry et al., 1999; Lam et al., 2000, 2003; Scott et al., 2001). These trials have focused on CBT as an adjunct to mood stabilizers.

Understanding the process and mechanisms of successful CBT for mood disorders is essential to developing more efficacious, more effective, and more appropriately targeted treatments for depression. The cognitive model (A. T. Beck, 1976a) predicts that CBT should produce specific changes on measures of cognitions, that these changes in cognitions are unique to CBT and that these changes in cognitions should predict symptomatic improvement.

One approach to testing this model is to examine changes on questionnaires designed to assess cognitive-specific changes, such as the Dysfunctional Attitude Scale (DAS: Weissman and Beck, 1978) and the Attributional Style Questionnaire (ASQ: Peterson et al., 1982). Several studies have found that people receiving pharmacotherapy for depression achieved similar changes in mood and cognitive processes as people receiving CBT, suggesting that cognitive changes were secondary to mood change (e.g., Imber et al., 1990). However, Seligman et al. (1988) found that CBT significantly improved explanatory style on the ASQ and that change in explanatory style correlated with change in depressive symptoms. However, without comparing CBT with other therapies, it was not possible to determine whether this change was unique to CBT or just secondary to symptomatic improvement. DeRubeis et al. (1990) found that change from pretreatment to mid-treatment on the ASQ and DAS predicted change in depression from mid-treatment to posttreatment for depressed patients in a CBT group but not in a pharmacotherapy group, suggesting that cognitive change is associated with improvement in CBT but is not alone sufficient to produce symptom relief.

There are general problems with the use of questionnaire measures to investigate cognitive change. Self-report responses are vulnerable to demand effects, response biases and the mood of the reporter may influence which items are endorsed, as many items differ in hedonic tone. Furthermore, it is not clear how well self-report questionnaires measure underlying cognitive structures and processes, such as schema, which are hypothesized to be important in the development of depression. There is also evidence to suggest that the specific cognitive biases associated with depression can only be observed when people are tested in a negative mood (Teasdale and Dent, 1987; Dent and Teasdale, 1988; Miranda et al., 1988).

More supportive of the cognitive change hypothesis, recent research found that a significant minority of people diagnosed with depression undergoing CBT showed ‘sudden gains’, where there was substantial symptom improvement in one between-session interval (Tang and DeRubeis, 1999). Such sudden gains are associated with better long-term outcomes, with people who experienced sudden gains significantly less depressed than those not experiencing sudden gains at 18-month follow-up. In CBT, sudden gains seemed to be preceded by critical sessions in which substantial cognitive changes occurred. However, as the sudden gains effect has recently been found in supportive-expressive psychotherapy for depression (Tang et al., 2002), the exact mechanism underpinning sudden gains (i.e., different mechanisms for different therapies versus nonspecific treatment effects) remains unresolved.

Recent studies have suggested that changes in the style of processing depression-related information, rather than just changes in thought content, might be important in the mechanism of CBT. Teasdale et al. (2001) found that in people with residual depression, CBT reduces an absolutist all-or-nothing thinking style, which, in turn, was found to mediate the effects of CBT on preventing relapse. Similarly, CBT successfully reduces relapse in people who report increased ‘metacognitive awareness’ at the end of treatment (defined as the ability to view thoughts as mental events in a wider context of awareness) (Teasdale et al., 2001). Thus, these studies suggest that CBT may prevent relapse by shifting the mode or style of processing. However, these studies have exclusively focused on residual depression with relapse as the outcome measure, leaving the generalizability of these findings to acute depression unresolved.

Interestingly, successful CBT for acute depression produces significantly greater reductions in ‘cognitive reactivity’ (operationalized as increases in dysfunctional attitudes following a negative mood induction) than successful pharmacotherapy for depression (Segal et al., 1999). Together with Teasdale et al.'s findings, this result is consistent with the notion that CBT helps people to acquire compensatory or metacognitive skills (Barber and DeRubeis, 1989) that regulate their cognitive responses to sad mood and stressful events.

An alternative approach to examining the process of change in CBT is to study the effects of specific techniques on outcome. Several therapy process-outcome studies suggest that homework is perceived as helpful and contributes significantly to change in cognitive therapy (Burns and Nolen-Hoeksema, 1991; Detweiler and Whisman, 1999; Burns and Spangler, 2000). Concrete symptom-focused methods of CT predict subsequent symptom reduction when assessed early in treatment (DeRubeis and Feeley, 1990). These concrete methods involved setting an agenda, asking for specific examples, labeling cognitive errors, examining evidence, and monitoring thoughts. However, less focused, more abstract approaches, such as exploring the meaning of thoughts and discussing the therapy, did not predict improvement.

What predicts whether someone will respond to CBT for depression (a prognostic indicator) and whether someone will respond better to CBT than to another treatment (a prescriptive indicator)? Various client variables predict poor outcome to CBT (see: Hamilton and Dobson, 2002), including increased severity and chronicity of the depression and perfectionistic beliefs (Shahar et al., 2003), although these variables predict poor outcome for all interventions. Married clients do better with CBT than single clients (Jarrett et al., 1991). People with avoidant personality disorder may respond better to CBT than to interpersonal therapies (Barber and Muenz, 1996), although higher levels of endorsement of avoidant beliefs predicts poorer outcome in CBT (Kuyken et al., 2001).

In summary, there is now a large and converging body of evidence to indicate that CBT is an effective acute treatment for unipolar depression and is an effective relapse prevention treatment for unipolar depression, and, potentially, a relapse prevention treatment for bipolar disorder. While early studies failed to show that changes in cognition precede symptom changes in CBT, more recent work suggests that sudden gains are preceded by important shifts in beliefs and that CBT effects changes in the process (rather than the content) of cognition. Process-outcome research suggests that CBT can be made more effective by explicitly and concretely teaching patients metacognitive skills in generating specific plans and evaluating their own thoughts.

Cognitive therapy for depression will follow the key practice principles of all CBT treatments (see Chapter 2 by Grant et al.). In the rest of this section, we will elaborate on how the key principles are applied to depression, using the case example of Sheryl as an illustration.

Cognitive therapy tends to be short to medium term

Cognitive therapy for depression typically involves 16–20 meetings, although brief versions have been developed for particular circumstances (e.g., Bond and Dryden, 2002) and more sessions are indicated for chronic and recurrent depression (e.g., Moore et al., 2003). Initial sessions tend to be frequent (either twice a week or weekly) to initiate the change process, manage suicide risk, and achieve symptom relief, and later sessions tend to be less frequent (monthly and perhaps even 3-monthly) to consolidate gains and prevent relapse.

Fig. 11.4 Cognitive-behavioral formulation diagram for depression.

Cognitive therapy draws on a wide range of cognitive and behavioral techniques to change thinking, beliefs, and behaviors

The first class of therapeutic approaches focus on the client's behavior. The rationale is that for some people behavior monitoring, BA, and behavioral change can lead to substantive gains. For example, people with more severe depression often become withdrawn and inactive, which can feed into and exacerbate depression. The person withdraws, and then labels him/herself as ‘ineffectual,’ thereby fuelling the depression. By focusing on this relationship and gradually increasing the person's sense of daily structure and participation in masterful and pleasurable activities the person can take the first steps in combating depression (A. T. Beck et al., 1979). Other behavioral strategies include scheduling pleasurable activities, breaking down large tasks (e.g., finding employment) into more manageable graded tasks (e.g., buying a newspaper with job advertisements, preparing a resume…), teaching relaxation skills, desensitizing a person to feared situations, role-playing, and assertiveness training. To maximize the likelihood of success, plans need to be operationalized at a very concrete, detailed level, including consideration of when, where, how, and with whom the plans will be implemented, as well as potential obstacles and how to overcome them. It is important to note that within CBT, these behavioral techniques are used with the ‘collaborative empiricism’ approach, such that before plans are implemented, thoughts and beliefs relevant to the activity (e.g., ‘It is pointless to try’, ‘I won't succeed’, ‘I am too tired’, ‘I am not interested’) can be set out as hypotheses to be tested. Recent adaptations to CBT suggest that the changes in behavioral contingencies may be particularly important in treating severe and recurrent depression (see: McCullough, 2000; Martell et al., 2001).

The second class of therapeutic approaches focus on the client's negative automatic thoughts and maladaptive beliefs. Cognitive techniques are designed to increase clients’ awareness of these thoughts, challenge them by evaluating their basis in reality, and providing more adaptive and realistic alternative thoughts. The Dysfunctional Thought Record is used as a primary tool for developing this skill (Figure 11.5). Repeated practice at dealing with negative thinking is required for thought challenging to become a robust skill. Useful approaches to challenging automatic thoughts include listing evidence from past experience that supports and refutes each hypothesis, generating alternative explanations, checking whether a thought may reflect a cognitive error, and reattributing negative events to factors other than the client's personal inadequacy.

Fig. 11.5 Dysfunctional thought record. From Beck, A. T., Rush, A. J., Shaw, B. F., and Emery, G. (1979). Cognitive therapy for depression. New York: Wiley. Copyright 1979 by Wiley. Reprinted with permission.

In cognitive theory maladaptive beliefs (e.g., ‘If I drop my façade, others will despise me’) and higher-order core mode beliefs (e.g., ‘self-as-weak’) underlie automatic thoughts and are the next focus of cognitive interventions. Careful questioning about and exploration of client's unrealistic and maladaptive beliefs is carried out to examine if beliefs are based in reality, and to correct the distortions and maladaptive beliefs that perpetuate emotional distress. The advantages and disadvantages of the assumptions are explored and the possibility of adopting more functional, alternative rules is discussed. Early, often childhood, events that may have led to the adoption of these rules are explored and can be challenged, for example by using imagery to relive the event coupled with questions to introduce new perspectives. For Sheryl growing up in her family of origin a family maxim was ‘stiff upper lip’ or ‘don't show weakness.’ Behavioral plans designed to act against assumptions are a powerful way to change beliefs by providing personal experience that counters the assumption. For Sheryl this involved a process of applying the same standard to herself that she applied to other people, i.e., ‘a ‘capable’ person can be both strong and vulnerable and it is OK to show both these sides of the coin.’

Core modes require a further set of therapeutic strategies (J. S. Beck, 1995; Young et al., 2003). For example, when core modes such as ‘self-as-weak’ are identified, more adaptive beliefs (e.g., ‘I am basically capable and likeable’) can be established through Socratic questioning, examining advantages and disadvantages of the old and new core beliefs, acting ‘as if’ the new core beliefs were true, using coping cards, developing metaphors, subjecting the beliefs to tests across the person's life history and reconstructing associated memories and images (J. S. Beck, 1995). For many clients, automatic images, rather than thoughts, are powerfully associated with emotions and behaviors. Images are central to the sequelae of trauma and to psychiatric disorders such as posttraumatic stress disorder (PTSD) and other anxiety disorders that are often comorbid with depression. Images are handled in similar ways, but instead of verbally evaluating and challenging images, more visual techniques are used (J. S. Beck, 1995).

The third range of approaches takes place between therapy sessions as homework assignments. Homework is an essential element of cognitive therapy, aimed at building understanding and coping skills throughout the week, increasing self-reliance and rehearsing adaptive cognitive and behavioral skills. Homework moves the discussions in session from abstract, subjective discussion of issues to real day-to-day experiences. The therapist acts as coach, guiding and debriefing the client from week to week. Homework assignments are tailored to the individual, are set up as no-lose propositions, and may range from the therapist suggesting a relevant book, to the person undertaking a long procrastinated assignment (e.g., telephoning a friend to resolve an area of unspoken conflict), while monitoring the thoughts and images that come to light in preparing for the assignment (e.g., ‘the friend will be angry towards me’). As therapy progresses, the client takes on more responsibility for setting and reviewing the homework.

Having outlined the principles that underpin cognitive therapy, we aim to convey a sense of how cognitive therapy works in practice. We will outline a typical therapy session, as well as the progression of therapy as a whole, illustrating this through the case of Sheryl.

CBT for bipolar disorder adapts the classic cognitive therapy approach for depression in several ways (see: Basco et al., 1996; Lam et al., 2000). First, there is psychoeducation explaining the diathesis-stress model, outlining the joint role of medication and psychological treatment in reducing acute episodes. Therapists help clients to recognize that bipolar disorder involves a biological vulnerability, perhaps in the form of heritable changes in brain chemistry, which interact with stress to produce episodes of mania or depression. Cognitive therapy is emphasized as a means of reducing stress by learning improved coping skills and through testing personal perceptions that can themselves be stressful (e.g., self-critical thoughts).

Second, clients are taught self-monitoring and self-regulation skills, with an emphasis on identification and early recognition of prodromes and development of good coping strategies in response to prodromes. An idiosyncratic evaluation of early, late, and middle warning signs of an impending episode are drawn up with each client and useful coping plans made for each symptom. Encouraging clients to keep daily activity schedules and mood charts can be very helpful in facilitating effective self-monitoring, and ensuring that potential episodes are caught early enough.

Typical prodromes for mania include reduced sleep/need for sleep, increased goal-directed activity, reduced anxiety, increased optimism, irritability, increased libido, increased sociability, racing thoughts, and distractibility. Typical prodromes for depression include reduced interest in people or activities, feeling sad or depressed, disturbed sleep, tiredness, low motivation, increased worry, and poor concentration. For mania prodromes, engaging in calming activities, increasing rest, reducing stimulation and decreasing activity would be useful strategies, whereas increasing levels of activity, enjoying the ‘high’, and ‘making up for lost time’ would be unhelpful strategies likely to increase the risk of a full-blown manic episode. Similarly, for depression prodromes, keeping busy and maintaining routines are associated with better outcomes, while cutting down on activities, withdrawing from other people and going to bed are associated with worse outcomes. For each client, an individual case formulation is required to determine the idiosyncratic prodromes and the most functional responses, as there is a great degree of individual variability. For example, some clients report changes in sensory experiences, such as colors becoming brighter or noises sharper, or increased pleasure at the sensation of moving at speed, when they are becoming hypomanic. For such clients, behavioral plans would need to modulate their experience of such sensations, e.g., pacing their exposure to stimulating environments such as art galleries, museums, shopping malls, and temporarily reducing travel by car, plane, or train.

Third, behavioral plans are made to promote good sleep and good daily routine, in recognition of the evidence that disruptions in sleep and working routine are implicated in the onset of bipolar episodes (Healy and Williams, 1989; Malkoff-Schwartz et al., 1998). Maintaining regular times to go to bed and get up, as well as meal times, can significantly help to stabilize mood. Clients learn to balance their activity schedules, not to do too much or too little, and to pace their own activities. The role of social activity needs to be carefully monitored and paced—social withdrawal is a warning sign for depression, while increased social contact can be overstimulating and feed into the development of mania. The roles of medication, substance, and alcohol use also need to be explored in detail with clients as potential risk factors for episodes. For a number of clients, discontinuation of their mood stabilizer is associated with the recurrence of a bipolar episode, as is very apparent from reviewing their life history—for these clients, explicit recognition of what their own experience tells them about the effectiveness of medication can be very productive. Given the high rates of alcohol and substance abuse in bipolar disorder, close monitoring of usage and explicit plans to keep use to a level that is not associated with the spiraling of mood is also important. As with all plans with bipolar clients, such plans need to be arrived at collaboratively and from an open exploration of the advantages/disadvantages of the options.

Fourth, as in CBT for unipolar depression, therapists challenge unhelpful automatic thoughts and clients use Daily Thought Records. However, as well as challenging negative thoughts associated with depression, therapists can also challenge excessively positive thoughts that may be involved in the development of mania. When focusing on hypomanic thoughts, it is important to be open, giving clients plenty of autonomy in their responses, as well as to review with clients the consequences of hypomanic thoughts, in order to overcome reluctance to dwell on positive thoughts. For example, clients can be taught to examine grandiose thoughts such as ‘I know better than everyone else’ and check whether these thoughts accurately reflect their past experience, and whether such thoughts are dependent upon their mood. Similarly, clients can be encouraged to examine the longer-term costs and benefits of their more-impulsive and grandiose thoughts. For example, questions such as ‘How many of your ideas still seem a good idea a week later? If this is genuinely a good idea, it should still be a good idea next week. Can you try and leave it for a week?’ can help to reduce impulsivity.

It is useful here to explore client's ambivalence about their manic episodes—many bipolar clients experience the initial stages of mania as positive, as they are no longer depressed, feel more confident, have more energy, and are more creative, but also report that more extreme mania is distressing because they feel out of control, act in self-destructive ways, and, in some cases, experience extreme anxiety and/or unpleasant psychotic experiences. As the cognitive model would expect, however, the information processing biases afforded by the hypomania mean that clients tend to focus on the positives of their mood state and forget the negative experience of the mania. Exploring both sides of the mania with clients in a Socratic way can be very helpful for facilitating rapport and for helping clients maintain the perspective necessary to motivate themselves to maintain therapy plans during an incipient hypomania.

Finally, as in standard CBT, therapy works to challenge dysfunctional assumptions that increase risk of relapse. Bipolar clients display the same dysfunctional assumptions as clients with unipolar major depression in the context of a depressed episode. However, bipolar clients also have more idiosyncratic assumptions centering on highly driven and extreme goal-attainment beliefs, e.g., ‘I should be happy all the time’, ‘If I put in enough effort, I should be able to achieve everything I want’. Bipolar clients with these attitudes are more likely to engage in extreme goal-pursuing behavior (‘trying to make up for lost time’), which is likely to disrupt their sleep and daily routines precipitating further episodes. Cognitive therapy can identify such beliefs collaboratively with clients and then explore how realistic and useful such beliefs are. Final sessions can also productively explore the losses and stigma that clients have incurred as a result of their illness and help clients to work through these issues, by grieving for these losses and developing more functional views. As the losses and stigma associated with bipolar disorder are genuine, Socratic questions more helpfully focus on people's approach to this reality (e.g., ‘What constructive lessons can I learn from my past?’ ‘How can I go forwards from here in a way that makes my life worthwhile?’), rather than by challenging the evidence for the losses. Clients’ own self-stigma can be challenged, particularly the relatively common beliefs that they are defective. Such beliefs often arose during adolescence when clients had difficult relationships with family and peers, as a consequence of mood swings that occur as the first manifestation of the illness. It is particularly helpful to refocus clients on their strengths and to encourage them to consider and pay attention to the multiple roles they occupy (e.g., parent, son, friend, worker) rather than exclusively focus on the label of bipolar disorder.

In this section, we will briefly consider what to do when faced with some of the most common difficulties and challenges that occur during CBT for depression, particularly when it is severe, chronic, and recurrent.

The last three decades have seen CBT for mood disorders develop as a treatment of choice for unipolar depression and a promising intervention for bipolar disorder. We would predict that the next 25 years will see a range of exciting developments in CBT research and practice. In the area of outcome research, the most obvious area for advancement is where promising initial research suggests that CBT may prove to be an evidence-based approach: depression that is comorbid with personality disorders, PTSD, and substance misuse, dysthymia and bipolar disorder. Similarly, psychotherapy outcome research is needed to examine how cognitive therapy fares when it is adapted to different populations (e.g., older adults) and to different service settings (e.g., primary care). As we increasingly recognize depression as a potentially chronic relapsing condition, efforts to address depression in young people are urgently required. Given the scale of depression as a public health problem (Murray and Lopez, 1997), alternative formats (e.g., Internet-facilitated group therapy) are required.

In a climate of managed health care, evidence-based practice, and practice guidelines, researchers, practitioners, and policy makers are increasingly asking the question ‘What works best for whom?’. Beyond the comparative outcome studies, this sets the stage for interesting psychotherapy process and psychotherapy process outcome research. The mechanisms by which cognitive therapy is effective are not well understood, and this research will inform practice and health care policy. The stepped care approach to planning services and interventions is likely to be important here, as we become increasingly knowledgeable about what works for whom and through what mechanism. Cognitive therapy for depression is amenable to contemporary stepped care approaches, whereby clients are assessed and offered increasingly specialized, intensive, and complex interventions based on an algorithm of clinical need and optimal cost-effectiveness. Using the range of established cognitive therapy approaches, steps might graduate from bibliotherapy (Jamison and Scogin, 1995), to computer-based approaches (Wright et al., 2002), to brief psychoeducational approaches in primary care, to brief group approaches in secondary care to more in depth and extended individual or group cognitive therapy in either secondary or tertiary care (DeRubeis and Crits-Christoph, 1998).

The recent focus on primary and secondary prevention of mood disorders is welcome and there is much mileage in building on initial successes (e.g., Jaycox et al., 1994; Segal et al., 2002). The acceptability of cognitive therapy to many children and adolescents with depression and to people with recurrent depression combined with an increasing acknowledgment that primary and secondary prevention are high priority healthcare areas suggests we are likely to see much innovative and important work in this area.

Cognitive therapy is established as a mainstream psychotherapy of choice and training, supervision and accreditation are areas that require further development that extends and builds on existing best practice. There is an increasing body of cognitive therapy practitioners and researchers who are well placed to continue this work.