Decision making about surgery for early-stage breast cancer.

BACKGROUND Practice variation in breast cancer surgery has raised concerns about the quality of treatment decisions. We sought to evaluate the quality of decisions about surgery for early-stage breast cancer by measuring patient knowledge, concordance between goals and treatments, and involvement in decisions. STUDY DESIGN A mailed survey of stage I/II breast cancer survivors was conducted at 4 sites. The Decision Quality Instrument measured knowledge, goals, and involvement in decisions. A multivariable logistic regression model of treatment was developed. The model-predicted probability of mastectomy was compared with treatment received for each patient. Concordance was defined as having mastectomy and predicted probability >0.5 or partial mastectomy and predicted probability <0.5. Frequency of discussion about partial mastectomy was compared with discussion about mastectomy using chi-square tests. RESULTS Four hundred and forty patients participated (59% response rate). Mean overall knowledge was 52.7%; 45.9% knew that local recurrence risk is higher after breast conservation and 55.7% knew that survival is equivalent for the 2 options. Most participants (89.0%) had treatment concordant with their goals. Participants preferring mastectomy had lower concordance (80.5%) than those preferring partial mastectomy (92.6%; p = 0.001). Participants reported more frequent discussion of partial mastectomy and its advantages than of mastectomy, and 48.6% reported being asked their preference. CONCLUSIONS Breast cancer survivors had major knowledge deficits, and those preferring mastectomy were less likely to have treatment concordant with goals. Patients perceived that discussions focused on partial mastectomy, and many were not asked their preference. Improvements in the quality of decisions about breast cancer surgery are needed.