It's long-term stressors that take a toll: Comment on Cohen et al. (1998).

In an extremely well-controlled study, Cohen et al. (1998) add to prior knowledge of stress-illness relationships by showing that self-reports of stress occurrence and duration of 1 month or more, rather than estimates of stressor severity, predict susceptibility to experimentally induced colds (i.e., viral replication and cold symptoms). Although ruling out obvious behavioral and personality factors as causes of the association of stressors to colds, they were unable to identify mediational immune factors, a deficit attributable to the difficulty of assessing the multi-layered, dynamic physiological processes within the bidirectional connections of the nervous (stress) and immune systems. The findings provide an interesting complement to data, showing that people use stressor duration in evaluating the illness implications of somatic symptoms (Cameron et al., 1995), and suggest caution with regard to overestimating the prevalence of stress-induced colds in natural settings.