Artifacts in "limited prospective" designs? Testing confounding effects on response behaviour of women prior to breast surgery

Abstract

Data from a 'limited prospective' study on life events and breast cancer were tested whether they were biased by knowledge of the prospective diagnosis or by related mood states. Data were gathered by semi-structured and tape-recorded interviews. Event classifications were made according to Brown and Harris' Life Events and Difficulties Scale, an interviewer-based rating method. Ratings were performed before the interviewer learned the final diagnosis. Among the women coming to hospital with a suspicious breast lump, 33 had a malignant and 59 had a benign diagnosis. For comparison for mood effects two more control groups (women prior to gall stone surgery and healthy women interviewed at home) were included. Although artifacts may in general play a role for reporting events in 'limited prospective' designs, the present study could not document a sizeable effect of presumed confounding variables on response behaviour.