Self-Reported Stress and Risk of Breast Cancer
Stress and the Development of Breast Cancer: A Persistent and Popular Link Despite Contrary Evidence T he March 15, 1996 issue of Cancer published an article and editorial that raise questions regarding the role of psychosocial factors in the etiology of breast cancer. Using a retrospective case-control design, Roberts et al. 1 found no relationship between life stressor frequency and the diagnosis of breast cancer. They concluded therein: ''for clinicians, there is presently no evidence that commonly experienced difficult life events in any way constitute an increased risk of breast cancer.'' 1 In the editorial response to that article, Cassileth concludes more strongly that there is ''powerful information. . . [that] provides evidence that cancer is not caused by stress'' and that the role of ''stress or other emotional factors in the etiology of cancer... would be relatively minor. '' 2 Cassileth bases this conclusion on the study by Roberts et al. as well as three longitudinal studies in which people who experienced traumatic life stressors -war survivors, prisoners of war, and concentration camp incarceration victims -reported no higher rates of cancer than control groups. Before concluding that there is no link between life stressors and the etiology of breast cancer and cancer of other types, it is important to examine carefully what needs to be addressed to dismiss this line of research.
A fair test to examine life stressors in cancer would require a number of elements. First, it is important to distinguish the terms ''stressor'' and ''stress''. A ''stressor'' is an environmental stimulus that has the potential to cause distress. Stress is the response, both psychologic and physiologic, to a stressor. Life stressor research has been confusing because of the lack of precision using such terms. 3 Moreover, there is a complex interaction between a life stressor and the individual it affects. The individual response to the stressor is affected by the perception of the level of threat that the stressor poses. 4 The perception of threat in both animal and human studies relates to a stressor's duration and severity and its level of controllability and predictability. Furthermore, the stressor response can be mediated or buffered by both the individual coping responses to the stressor and the amount of social support that the individual can obtain. Therefore, any test of the role of life stressors in the etiology of cancer should examine this environment-individual interaction (i.e., ''psychosocial context'').
Measuring this interaction can then provide data quantifying the internal distress that an individual experienced subsequent to exposure to an external life stressor. Distress, the response that an individ-Supported in part by NIMH R01 MH48628.