The relationship between stress and illness: a historical and theoretical review of some conceptual and methodological problems in research

Abstract

Research into stress and illness is fraught with methodological and conceptual problems. These problems have slowed progress in research. Life stress variables are still conceptualized at a crude, simplistic and naive level. Research findings in life stress, either in terms of increasing the predictive power of life stress variables, or enhancing our understanding of the stress-disorder relationship, have advanced little in the last ten to twenty years. A possible approach to this problem is adopted in this thesis. By looking at how the ways in which the term stress has been used and developed in different areas of research, the diverse uses of this concept can be distinguished. The background to stress and illness research can now be approached with a clear conception of these different uses. Although there is general evidence for the link between stress and illness, knowledge about the processes and mechanisms involved is sparse. Many of the insights made by early researchers in psychosomatic medicine, that disease causation is mullticausal, appear to have been forgotten by many researchers who use only a few variables in their research designs. The idea of 'mediators' of stress presupposes a certain model of stress, loosely based on a engineering analogy, where stress is pictured as an external force, which the individual will resist, and moderating factors will reduce the impact of the force. This analogy is influential in life stress research, but little evidence exists to suggest it may be correct. Recent moves towards assessing daily stress and coping have been criticised as such variables are contaminated by others. An unresolvable difference exists between those who see stress varibles as objectively measurable, and those who view stress and health as part of a much larger ongoing interaction between the person and their environment, and coping and social support variables as part of a more general effort to adapt.