CHICAGO — A substantial and independent association between individual provider, the treatment center, and type of breast surgery among elderly women with breast cancer suggests the presence of physician and institutional biases.1
That is the conclusion of a Medicare claims database review presented at the 2016 American Society of Clinical Oncology (ASCO) Annual Meeting.
“Mastectomy or breast conserving surgery is one of the key decisions facing women with invasive breast cancer,” said Anthony Paravati, MD, MBA, University of California San Diego, and Moores Cancer Center, San Diego, California. “Ideally, patients should receive an unbiased presentation of all appropriate treatment options with the ultimate decision coming from patient preference,” he added.
The study evaluated the impact of individual surgeons on the likelihood of mastectomy vs breast conservation surgery for the definitive treatment of breast cancer.
Dr. Paravati and colleagues identified female patients age 66 years and older with nonmetastatic, histologically confirmed breast cancer from the Surveillance, Epidemiology, and End Results (SEER) – Medicare linked database. Of the initial 135 162 patients identified, 29 358 were included in the final study cohort. The Medicare files were also used to distinguish care at a teaching hospital, the Charlson comorbidity score, chemotherapy, and use of breast MRI prior to surgery.
Using ICD and HCPCS codes, the type of surgery was identified with Medicare claims data. The individual surgeon was identified from the Unique Physician Identification Number or National Provider Identified, and surgeon characteristics were obtained from a tertiary linkage with the American Medical Association Masterfile.
The study used hierarchical multivariable logistic models clustered by surgeon and geographic region, controlling for a patient’s clinical and demographic covariates and a physician’s demographic and practice information.
A total of 6594 women comprised the mastectomy cohort.
Unadjusted rates of mastectomy ranged from 0.0% in the bottom quintile of surgeons to 46.1% in the top quintile.
Patient level predictors of mastectomy were found to be older age, primary tumor size greater than 2 cm, 4 or more positive lymph nodes, poor or undifferentiated grade, chemotherapy, breast MRI, and teaching hospital. Fewer mastectomies were associated with higher income and unknown nodal status. Median odds ratios were 1.71 for type of institution and 1.97 for provider and 2.38 for the combination of institution and provider.
“Efforts to reduce provider bias could potentially include the increased use of multidisciplinary clinics, improved patient or physician education, or implementation of shared decision making tools,” he concluded.
1. Boero I, Paravati AJ, Gillespie E, Hou J, Murphy JD. The impact of individual surgeons on the likelihood of mastectomy in breast cancer. Oral presentation at: 2016 ASCO Annual Meeting; June 3-7, 2016; Chicago, IL.