OBJECTIVES: To evaluate patient preferences and their willingness to accept risks of specific adverse events (AEs) associated with metastatic colorectal cancer (mCRC) treatments for expected survival benefits associated with these treatments. METHODS: We administered an online questionnaire to 100 patients with CRC and 150 with mCRC, which asked each respondent nine discrete choice experiment (DCE) questions. Each question was designed to elicit relative preferences for clinically relevant levels of four attributes of treatments for mCRC—the chance of living at least 3 years after starting therapy, the risk of severe skin rash, the risk of any gastrointestinal (GI) bleeding, and the risk of any other bleeding problems. Preference weights for the levels of these attributes were elicited using a random-parameters logit model. These preference weights were used to calculate the maximum acceptable risk (MAR) of each AE that patients were willing to accept for a 10% increase in 3-year survival. RESULTS: Results from a chi-square test identified no statistically significant differences in the preferences of patients with metastatic and nonmetastatic disease. Thus, a single preference model was used with both groups for the analysis. For an increase from 24% to 34% in the chance of 3-year survival with treatment, the MARs (95% confidence interval—CI) were as follows—severe skin rash, 77.3% (62.0%–92.6%); any GI bleeding, 49.4% (39.5%–59.4%); and any other bleeding problems, 40.9% (27.8%–54.0%). CONCLUSIONS: Patients were willing to accept a higher risk of developing a severe skin rash than suffering from any GI bleeding or any other bleeding problems for a given increase in survival benefit. This research shows substantial differences in patient preferences for treatment-specific AEs, making patient preference an important factor in the choice of mCRC treatment.