Changes in self-reported family history of breast cancer with change in case–control status

In order to study recall bias in self-reported family history of breast cancer, we selected a sample of cases and controls from a case–control study nested in a cohort, and compared family history reported in a questionnaire at enrollment, and after the development of the disease. We assessed whether changes occurred in self-reported familial breast cancer due to a change in health status in women who developed breast cancer. The κ of agreement for maternal history of breast cancer was 0.92 in cases, and 1.00 in controls; the κ for history of breast cancer in sisters was 0.65 in cases, 0.88 in controls. By comparing two questionnaires collected before the diagnosis of the cases (index date), and one questionnaire administered after the index date, we were able to assess that the changes in answer observed among the cases were recorded in the second questionnaire before the index date, and therefore were independent from the diagnosis of cancer. The study seems to suggest that change in recall is of limited importance when collecting family history of first degree relatives, and that women recall the health history of their mother better than the health history of their sisters.