Transfusion is associated with multiple risks and morbidities. Little is known, however, about preoperative predictors of transfusion in gastrointestinal surgery patients. To identify factors that influence transfusion practices, we analyzed hospital discharge data from colorectal cancer surgery patients in Maryland between 1994 and 2000 (n = 14,052). The primary outcome variable was whether or not patients received a blood product ("Any Transfusion"). Characteristics independently associated with an increased risk of receiving Any Transfusion included: advanced age (>80 yr: OR 2.3; 95% CI 1.9-2.9; 70-79 yr: OR 1.6; 95% CI 1.4-2.0 vs. <60 yr), moderate to severe liver disease (OR 2.5; 95% CI 1.5-4.2), mild liver disease (OR 2.1; 95% CI 1.5-2.9), diabetes with complications (OR 2.1; 95% CI 1.6-2.6), chronic renal disease (OR 2.1; 95% CI 1.4-3.0), female gender (OR 1.3; 95% CI 1.2-1.5), chronic pulmonary disease (COPD) (OR 1.3; 95% CI 1.1-1.4), and metastatic disease (OR 1.2; 95% CI 1.1-1.4). Patients at hospitals with an annual case volume in the highest quartile were at an increased risk for receiving Any Transfusion (OR 2.1; 95% CI 1.3-3.4) and those with surgeons in the highest volume quartile (>12 cases/yr) were at a decreased risk (OR 0.8; 95% CI 0.6-0.99). The association between greater surgeon case volume and low transfusion rates was seen in all but the very high volume hospitals (>74 cases/yr). Blood product transfusion was associated with a 2.5-fold (95% CI 2.1-3.1) increased mortality, 3.7 day (95% CI 2.1-3.1) increase in hospital length of stay, and a 7120 dollars (95% CI 6472 dollars-7769 dollars) increase in total charges compared to patients that did not receive Any Transfusion. This data can be used by providers in discussions with patients regarding the risks for transfusion and in identifying patients in whom strategies to reduce transfusions should be evaluated.