Coping has attracted much attention in research as a possible mediator of the psychological impact of cancer. Yet, conceptual ambiguity and methodological limitations have resulted in weak and contradictory findings. A major shortcoming has been the use of designs which do not represent adequately the complexity or the diversity of demands which arise from the diagnosis and treatment of cancer. The neglect of appraisal in the assessment of the relationship between coping and mental health is of particular concern, given the role it has been found to play in the onset and maintenance of affective disorders. In a prospective study 673 newly diagnosed cancer patients were interviewed 4 to 8 weeks and 1 year later to assess the effects of their appraisals, coping responses and resolution of any concerns on subsequent mental health. Logistic regression analyses, adjusted for possible confounding variables, were used to investigate the relationships between coping variables and affective disorders. When examined separately the degree of threat appraised, the reporting of a helpless response and perceived success of primary responses in resolving concerns all predicted subsequent effective disorder. However, in a multivariate model only appraisal and success of the response in resolving the concern were significant. No response was found to prevent affective disorders, however certain types of responses were associated with the resolution of specific concerns. The results highlight the importance of assessing the whole coping process and the need to address the complex and multifaceted characteristics of cancer demands. On the basis of our findings we describe a maladaptive cycle of coping, which we believe contributes to the later onset of affective disorders in cancer patients.