Table 2.

Adjusted Odds for Treatment Delay by Race/Ethnicity in Generalized Estimating Equations (GEE) GEE and Fixed Effects Models

Characteristic	Model 1 (GEE, Patient Characteristics)*	Fixed Effects (Patient Characteristics + Hospitals)†
Race/ethnicity
White	1.00	1.00
Black	1.25 (1.10, 1.42)	1.17 (1.02, 1.33)
Hispanic	1.14 (0.98, 1.33)	1.07 (0.92, 1.25)

^*Using logistic regression with GEE to account for clustering at hospital level with inclusion of variables above and adjusted for age, income, percent without high school diploma, marital status, Charlson score, year of diagnosis, tumor size, number of nodes positive, stage, tumor grade, hormone receptor status, surgery, SEER registry, and county.

^†Using fixed effects model with the variables above in addition to hospitals categorized by number of patients eligible for radiation (<10, 10–19, 20–29, and individual hospitals for those with ≥30 patients).