Background: Few studies of practice variation in the management of early breast cancer for elderly women have examined the process of care in depth. This study evaluated the effects of age and other factors on surgical staging techniques and treatment.
Methods: Virginia cancer registry data were linked with Medicare claims and 1990 census data. The sample included all newly diagnosed patients with pathologic confirmed local and regional breast cancer in 1985-1989 (n = 3,361). Analyses included descriptive univariate statistics and multiple logistic regression analysis for staging and treatment alternatives. Process of care variables included tumor size determination, axillary lymph node dissection, use of adjuvant therapy, and radiation if breast conserving surgery (BCS) was performed.
Results: About 75 percent of women had tumor size and axillary node dissection. Increasing comorbidity was associated with a lower likelihood of axillary node dissection. Nine percent of local compared to 44 percent of regional disease patients received adjuvant therapy. Hormonal therapy increased from 13 percent of women in 1985-1988 to 24 percent in 1989. Hormonal therapy did not vary with patient age. One-third of the patients with positive lymph nodes compared to 8 percent of node negative women received hormonal therapy. Blacks were more likely to present with advanced disease. A logistic regression model evaluated the multiple effects of patients and clinical characteristics: older women were more likely to present with larger tumors, were less likely to have axillary node dissections, and were less likely to receive chemotherapy or radiation.
Conclusions: Younger age was most consistently associated with staging and the use of chemotherapy in this cohort of elderly breast cancer patients. Based on the reported initial treatment plan, hormonal therapy was infrequently used and information from axillary lymph node assessment was used to stratify treatment. Although the low use of adjuvant hormonal therapy in elderly women may compromise survival, neither comorbid nor socioeconomic factors as measured in this study explained this practice pattern.