Treatment options for mesalamine-refractory ulcerative colitis (UC) include chronic immunosuppressive medications or colectomy surgery. Current treatment paradigms presume the patients' foremost desire is to avoid surgery and therefore view surgery as a consequence of medication failure. However, immunosuppressive therapy may not be ideal for all patients due to unclear durable efficacy and potential lethal serious adverse events (SAEs). We sought to quantify UC patients' risk tolerance of chronic immunosuppression to avoid colectomy. We first conducted a meta-analysis of all-cause and cause-specific mortality in both Crohn's disease (CD) and UC, and examined the effect of study design on this outcome. We found elevated all-cause and cause-specific mortality in both UC and CD including colorectal-, pulmonary- and non-alcoholic liver disease-related relative mortality. We further found little evidence that study design impacted all-cause relative mortality summary estimates. We next conducted a study examining the reliability of the 6-Point Mayo score, a simple two-item non-invasive non-physician driven index, for measuring UC disease activity. We found the 6-Point Mayo to strongly correlate with more extensive disease assessment tools, with a similar sensitivity, specificity and ROC area under the curve for patient-defined clinical remission. With these insights, we conducted a discrete choice experiment to quantify the UC patients' mean maximum acceptable risk for life-threatening SAEs associated with immunosuppressant therapy to avoid colectomy surgery with various outcomes. We found that UC patient tolerance for medical and surgical risks do not conform to conventional preference-elicitation methodology assumptions. UC patients were willing to accept very high levels of fatal SAEs to avoid an ostomy. However, if a durable medication-induced remission could not be achieved, patients were equally satisfied with J-pouch surgery. Several important clinical phenotypes impacted patient risk tolerances. This is the first empirical demonstration that UC patients view a well-functioning J-pouch as equivalent to mild clinical disease. It further demonstrates that patients value medication efficacy and suggests that clinical remission, rather than response, be the preferred outcome for therapy trials and treatment algorithms. Our findings underline the need for rigorous methodologies to accurately measure patient-preferences; and suggest potential avenues to enhance UC patient autonomy and facilitate shared decision-making.