The ideas outlined in this paper developed from three separate strands of my work as a therapist treating older people. Firstly, I was concerned with reducing the number of patients not benefiting from therapy. In keeping with Pinquart et al.'s (Am J Psychiatry 163:1493–1501, 2006) recent meta-analytic survey concerning the use of psychotherapy and medication, 50% of the depressed patients treated by my team failed to respond adequately. Secondly, the ideas grew out of my attempts to determine why different therapies with older people have often been shown to have equivocal beneficial impacts (Davies and Collerton, J Ment Health 6:335–344, 1997). Thirdly, I was interested in integrating the relevant neuropsychological features, particularly memory, into my therapeutic work. These three strands have led me to propose a pan-theoretical model of distress based on memory functioning and existing information processing perspectives (Teasdale and Barnard, Affect, cognition and change: Remodelling depressive thought, 1993; Power and Dalgleish, Behav Cogn Psychother, 27:129–142, 1996).