Family attitudes in youth as a possible precursor of cancer among physicians: a search for explanatory mechanisms.

A measure of youthful family attitudes, the Closeness to Parents Scale, has continued to be predictive of cancer among physicians in a prospective study of medical students. Nonetheless, questions have remained concerning the meaning and reliability of this measure and whether its predictive value is diminishing over time. Perhaps more important, it is necessary to ascertain whether the relationship is the result of some methodological artifact or whether it is mediated by an association with known risk factors, such as smoking, drinking, and radiation exposure. Each of these issues was examined in turn, using a variety of statistical techniques to refine the scale and to equate cancer and control groups with respect to risk factors as well as possible artifacts. In a group of 913 men, it was found that the scale is primarily a function of good father-son relationships and that its association with later cancer persists even after the influence of possible mediating and artifactual variables is statistically controlled. Several possible explanations for these findings are discussed.