Clinical observations that began centuries ago and extend to the present describe an association between depressive symptoms and cancer. This discussion focuses on such symptoms, not clinical depression. One could not conclude with confidence from studies done during the last 40 years whether depression would predict an excess risk of later cancer. Most of the earlier studies suffered from faulty design, small samples, failure to adjust for confounders, and a number of biases.
The first large prospective study on a normal cohort was that of Shekelle et al,1 appearing in 1981. They gave 2020 men the Minnesota Multiphasic Personality Inventory and determined the group's cancer mortality over 17 years. Men who scored highest, among that measure's scales, on the Depression scale showed a relative risk of 2.3 after controlling for several relevant variables; that is, a risk of death from cancer 2.3 times that of men whose depression score