OBJECTIVES: To examine whether receipt of chemo-, radiation, and hormonal therapy regimens that are jointly guideline concordant improve survival outcomes among women diagnosed with breast cancer in a rural region of the United States. METHODS: All women identified by the state cancer registry residing in rural southwest Georgia diagnosed with early stage breast cancer during 2001-2003 were included.  Medical chart abstraction and state registry data were used to determine treatment concordance with guidelines established by the 2000 NIH consensus development conference on breast cancer treatment.  Patients were Concordant versus Non-Concordant according to whether their receipt (or non-receipt) of each adjuvant therapy type was according to guidelines.  To examine the effects of concordance on all-cause and breast cancer-specific survival, Cox models were developed that used both propensity score (PS) weighting and 2-stage residual inclusion (2SRI) instrumental variable techniques to adjust for patient selection effects.  RESULTS: In all-cause analyses, Concordance versus Non-Concordance was associated with significantly better survival (hazard ratios (HRs) 0.41 (95% CI: 0.24-0.72) to 0.54 (95% CI: 0.33-0.87).  Similar findings emerged in breast cancer-specific survival analyses, with HRs significantly less than 1.0 in most cases.  Diagnosis at older age or later disease stage strongly predicted poorer survival outcome; being not married was significant in all-cause but not breast cancer-specific models.  Survival was not generally associated with surgical treatment delay, insurance status, socioeconomic status, rural/urban status, comorbidities, tumor grade, or hormonal status.  HR for black women versus white was greater than 1.0 across models but never significant (p=0.05). CONCLUSIONS: Breast cancer patients in rural Georgia who received guideline-concordant adjuvant therapy had significantly better all-cause and breast cancer-specific survival, based on Cox model analyses that attempted to control for multiple clinical and demographic factors, as well as selection effects.  These findings extend the evidence that guideline bundles of care improve outcomes.