Breast cancer treatment is complex, involving several modalities, such as surgery, radiation therapy, and chemothempy. Little is known about determinants of patterns of treating breast cancer patients. The long term goal of this work is to describe treatment patterns and their association with outcomes for breast cancer patients. This proposal focuses specifically on the choice of surgical treatment for early breast cancer in the elderly. Brent-conserving surgely has become a widely accepted modality of treatment for breast cancer patients in the past 5-10 years. However, little is kwown about who is receiving this treatment. This research is designed to elucidate factors associated with the choice of surgical therapy, concentrating on the choice of breast-conserving surgery vs mastectomy.
The specific aims of this proposal are as follows: 1. We will confirm our pilot studies of the patterns of patient demographic factors (age and race) associated with breast-conserving surgical treatment for early breast cancer in the Medicare population, and include adjustment for comorbidities. 2. We will confirm and extend our pilot studies of geographic and hospital factors (such as state, region, size of the metropolitan statistical area, size of hospital, facilities available, medical staff characteristics) associated with breast-conserving surgical treatment for early breast cancer. 3. We will determine some early outcomes associated with these patterns by observing the short-term mortality, length of stay, and early charges associated with the different treatment strategies, controlling for patient characteristics. The proposed methodology uses data already available from the Medicare claims records of 1986 and 1987, and from the 1986 American Hospital Association survey of hospitals. For each year, cohorts will be constructed of elderly women undergoing surgical procedures for early breast cancer, and the most invasive procedure undergone will be determined. Univariate analyses will be run of patient and hospital characteristics associated with choice of breast-conserving as opposed to mastectomy surgery. A logistic regression model will be constructed to explain the choice of therapy. Early outcomes will then be determined, and stratified by treatment type. This proposal addresses an important and relatively unexplored area of research. In addition, the use of the Medicare claims information offers a method of studying a large number of relatively unselected elderly women, cared for by a large number of relatively unselected physicians.
