Behavioral Activation Therapy for Depressed Cancer Patients: Factors Associated with Treatment Outcome and Attrition

Psychological interventions for cancer patients have included educational methods, supportive therapy, cognitive-behavioral therapy, relaxation training, problem-solving and social skills training, biofeedback, and hypnosis (Andersen, 1992; Baum & Andersen, 2001; Carlson & Butz, 2004; Evans et al., 2005). In studies assessing the efficacy of these interventions, there is substantial support that they effectively reduce symptoms of depression, anxiety, and pain (Antoni et al., 2001; Carlson & Butz, 2004; Goodwin et al., 2001; Hopko et al., in press; Moorey, Greer, Bliss, & Law, 1998; Trijsburg, van Knippenberg, & Rijpma, 1992). Conversely, there are a few studies in which psychosocial interventions have seemingly had minimal impact in reducing psychological distress (e.g., Cunningham et al., 1998; Edelman, Bell, & Kidman, 1999), and in a fairly recent review of the literature, it was concluded that although no intervention can be highly recommended for reducing depression in cancer patients, there is most empirical support for group therapy, education, structured counseling, cognitive behavioral therapy, communication skills training, and self-esteem building approaches (Newell, Sanson-Fisher, & Savolainen, 2002). Among the several limitations inherent to psychosocial treatment outcome research with cancer patients, very few studies have targeted patients with well-diagnosed clinical depression and little is known about the factors that most predict treatment outcome and attrition in this population. The present study addresses precisely these issues. In treatment outcome research focused on individuals with clinical depression, several factors have been associated with a negative (or limited) treatment response, including increased severity and chronicity of depression, earlier age of onset, family history of depression, presence of a personality disorder (but see Mulder, Joyce, & Luty, 2003), co-existent Axis I disorders, presence of psychotic symptoms, perceived social stigma, increased cognitive and/or social dysfunction, high levels of self-criticism, marital problems or being unmarried, decreased treatment expectations, and double depression (Duggan, Sham, Minne, Lee, & Murray, 1998; Earle, Smith, Harris & Longino, 1998; Enns, Cox & Pidlubny, 2001; Jarrett, 1995; Kung & Elkin, 2000; Mynors-Wallis & Gath, 1997; Robinson & Spiker, 1985; Sirey et al., 2001; Sotsky et al., 1991; Viinamaki et al., 2006). A recent study utilizing functional magnetic resonance imaging also revealed potential treatment outcome differences as a function of organic processes. Specifically, patients whose reactivity to emotional stimuli was low in the subgenual cingulate cortex and high in the amygdala demonstrated more attenuation of depressive symptoms following cognitive-behavioral therapy (Siegle, Carter, & Thase, 2006). Pertaining to treatment attrition, variables associated with an increased likelihood of prematurely terminating psychotherapy include lower socioeconomic status, reduced income and lower educational level, ethnic or racial minority status, and younger age (Grilo et al., 1998; Rabin, Kaslow, & Rehm, 1985; Roffman, Klepsch, Wertz, Simpson, & Stephens, 1993; Self, Oates, PinnockHamilton & Leach, 2005; Stratterfield, 1998). Patients presenting with co-existent psychiatric diagnoses also tend to exhibit higher attrition rates (Arnow et al., 2007; Frank et al., 2000; Issakidis & Andrews, 2004). More specific to treatment process, treatment expectations may impact attrition, with negative attitude towards treatment and a discontinuity between patients' personal beliefs and proposed treatment rationale increasing attrition rates (Rabin et al, 1985). Arnow et al. (2007) also reported that lower therapeutic alliance scores predicted earlier attrition in treatment. Finally, patients enrolled in cognitive-behavioral group therapy also appear more likely to discontinue treatment if they have higher pretreatment hopelessness scores and are more pessimistic about symptom control (Westra, Dozois & Boardman, 2002). …