DiscussionThe findings indicated that both symptoms of anxiety and depression decrease significantly when treated with both CBT focused on depression and CBT targeting both depression and anxiety. Approximately two-thirds of patients showed reliable improvement and recovery from their symptoms of depression, and three-quarters showed reliable improvement on symptoms of anxiety although only 56% could be considered ‘recovered’ on this measure. There were relatively few significant differences between the two groups. There were other indications of the possible superiority of CBT-D compared to CBT-DA in that the mean depression and anxiety scores for participants receiving CBT-DA but not CBT-D remained above the clinical thresholds post-treatment but these differences were not significant with the current sample size. In addition, the mean post treatment depression and anxiety scores forparticipants receiving CBT-DA were above the clinical thresholds whereas the mean for partic-ipants receiving CBT-D were below threshold.Overall, there was no evidence to suggest that CBT-DA is superior to CBT-D.The study goes some way to answering the research question posed with regard to optimalapproaches to the treatment of comorbid anxiety and depression although it is noteworthythat the question of clinical outcome when CBT focuses on anxiety only could not be answered since there were too few cases. There are several possibilities for why clinicians would prefer to choose treatment that includes depression in cases of comorbidity. First and foremost, clini- cians may be reluctant to treat patients with depression with an anxiety protocol due to con- cern about risk. Second, clinicians may feel more competent and comfortable with CBT for depression rather than CBT for anxiety. Most clinicians in routine psychological services are first trained in CBT for depression and then trained in specific anxiety disorder protocols. Per- haps the order in which clinicians are trained influences the order in which the treatments are provided in cases of ambiguity and comorbidity. It is possible that clinician experience influ- ences decision making, with more experienced clinicians viewing CBT-D as a simpler and equally effective approach, leading to more effective treatment in this group. Alternatively, itmay be the case that some of the clinicians observe that anxiety improves when depression is successfully addressed and therefore see no reason to include CBT for anxiety as an additional component. One reason for this observed improvement may be due to the overlap in symp- toms between anxiety and depression (e.g. [35], [36]); it would be interesting for future work to investigate whether CBT-DA does have an added benefit for specific symptoms of anxiety. Another explanation is that CBT protocols for anxiety are more specified (and arguablymore prescriptive) than Beck’s protocol for depression which allows other difficulties to be addressed within it e.g. managing anxiety, sleep. It is therefore possible that treatment for depression is more likely to involve additional treatment for anxiety, whereas treatment for GAD is less likely to involve additional treatment for depression. It may also be that ‘therapist drift’ [17] is more likely to occur in CBT-D as our review of 135 case notes indicated that 28 drifted from the evidence-based protocol and a further 10 incorporated techniques from CBT for Post Traumatic Stress Disorder. Further research is needed to explore these issues and to understand therapist decision making and their appraisals of their patients. It is likely that therapists will use clinical judgement to identify patients for whom they think a combinationis likely to be more effective. For example, they may believe that more complex patients require CBT-DA in order to address the range of symptoms, or alternatively they may make formula- tion based judgements as to why a particular patient may not respond to a combination treat- ment. Understanding the decision making process is an important area for future researchas it would both capture clinician wisdom and provide data on how clinicians incorporate research evidence into everyday practice.