Individual, area, and provider characteristics associated with care received for stages I to III breast cancer in a multistate region of Appalachia

Purpose: We describe individual, area, and provider characteristics associated with care patterns for early-stage breast cancer in Appalachian counties of Kentucky, North Carolina, Ohio, and Pennsylvania. Methods: Cases of stages I to III breast cancer from 2006 to 2008 were linked to Medicare claims occurring within 1 year of diagnosis. Rates of guideline-concordant endocrine therapy (n = 1,429), chemotherapy (n = 1,480), and radiation therapy (RT) after breast-conserving surgery were studied; RT was studied in women age ≥ 70 years with stage I estrogen receptor (ER)-positive/progesterone receptor (PR)-positive cancer, for whom RT was optional (n = 1,108), and in all others, for whom RT was guideline concordant (n = 1,422). Univariable and multivariable analyses were performed. Independent variables included age, race, county-level economic status, state, surgeon graduation year and volume, comorbidity, diagnosis year, Medicaid/Medicare dual status, histology, tumor size, tumor sequence, positive lymph nodes, ER/PR status, stage, trastuzumab use, and surgery type. Results: Population mean age was 74 years; 97% were white. For endocrine therapy, chemotherapy, and RT, guideline concordance was 76%, 48%, and 83%, respectively. Where it was optional, 77% received RT. Guideline-concordant endocrine therapy was lower in North Carolina versus Pennsylvania (odds ratio [OR], 0.60; 95% CI, 0.41 to 0.88) and higher if surgeon graduated between 1984 and 1988 versus ≥ 1989 (OR, 1.58; 95% CI, 1.06 to 2.34). Guideline-concordant chemotherapy varied significantly by state, county-level economic status, and surgeon volume. In guideline-concordant RT, lower surgeon volume (v highest) predicted RT use (OR, 1.63; 95% CI, 1.61 to 2.36). In optional RT, North Carolina residence (v Pennsylvania; OR, 0.29; 95% CI, 0.17 to 0.48) and counties with higher economic status (OR, 0.61; 95% CI, 0.40 to 0.94) predicated RT omission. Conclusion: Notable variation in care by geographic and surgical provider characteristics provides targets for further research in underserved areas.