A population sample was obtained from the British Columbia (BC) Cancer Registry of all women diagnosed with a first breast cancer in 2002 who were resident in Greater Vancouver or Greater Victoria, BC. Information on treatment and prognostic factors were obtained from source records. The study group was linked to the records of the Screening Mammography Program of BC to identify screening histories on women prior to diagnosis. Logistic regression was used to determine the relationship between screening participation and treatment and to predict treatment use from prognostic factors. Fifteen hundred and eighty-nine women with breast cancer were included in the study and 1,071 (67%) had participated in screening prior to diagnosis: 786 (49%) had been screened within the 30 months prior to their diagnosis (regular participants). Breast conserving surgery (BCS) rates were higher (OR = 2.3, p < 0.001) and chemotherapy use lower (OR = 0.53, p < 0.001) among regular participants compared with nonparticipants after adjustment for age. A predictive model based on the distribution of prognostic factors between participants provided estimates of OR = 1.47 and OR = 0.54 for BCS and chemotherapy, respectively, and adjustment for self-selection changed the predicted values to OR = 1.16 and OR = 0.67, respectively. Participation in screening produced a considerable change in the use of chemotherapy but less on BCS use.
(c) 2007 Wiley-Liss, Inc.