Premorbid psychological factors as related to cancer incidence.

In planning for studies relating psychological factors and/or stress (PF&/oS) to cancer, one should be aware of epidemiological findings that might contribute to or even account wholly for any found relationships. Most studies have not examined the known biological causes of cancer, nor have they described a rationale for relationships sought. The two broad mechanisms leading to cancer, carcinogens and lowered resistance to it, include physical and chemical causes, viruses and chronic infection, medication, genetic predisposition, hormonal stimuli, and aging. Interfering variables may bias or dilute a real relationship. Validity and reliability of instruments measuring PF&/oS are so variable as to warrant considerable care in their use. The latent periods of different cancers are measured in years, not months, with consequent potent impact on possible inferences drawn from prebiopsy and short prospective studies. In these and in retrospective studies, cancer can have strong and biasing effects on apparently straightforward PF&/oS measurements, as can iatrogenic effects. Some theoretical issues are discussed. The known prospective studies are discussed and reasons are given for the view that they are less convincing then many seem to think. A sketch of a model relating PF&/oS to cancer appearance is outlined, with some theoretical implications, and issues in research design are addressed.