BACKGROUND: Increasing patient age independently predicts nonreferral and nonreceipt of adjuvant therapy among patients with colorectal cancer. OBJECTIVE: This study aimed to identify factors affecting surgeons' decisions to refer older patients for adjuvant therapy. DESIGN/SETTINGS/PARTICIPANTS: A self-administered survey was sent to all Australian and New Zealand colorectal surgeons (n = 146). MAIN OUTCOME MEASURES: The survey consisted of 3 sections: 1) knowledge of research evidence, 2) opinions on evidence and adjuvant therapy in older patients, and 3) self-reported practice, or likelihood of patient referral in different scenarios. Demographic information was also obtained. RESULTS: Seventy percent of surgeons responded. Surgeons were significantly less likely to refer older patients than younger patients for adjuvant therapy in all scenarios (P < .001). The difference in referral recommendations was greatest when patients lived a long way from treatment, had a comorbid condition, or had little social support. There was greater variation in referral recommendations for older patients, and marked disagreement between surgeons in knowledge and opinion questions. Surgeon age was the only significant predictor of survey responses. Greater knowledge and more positive opinions predicted similar referral recommendations for older and younger patients (P = .02, P = .01). LIMITATIONS: Although decreased referral and receipt of adjuvant therapy among older patients is most likely multifactorial, this survey focused on the views of one physician group and a number of specific scenarios. CONCLUSIONS: Chronological age alone appears to impact colorectal surgeons' decisions to refer patients for adjuvant therapy. Sociodemographic and physiological factors further decrease the likelihood of referral of older patients. A lack of consensus among surgeons suggests that more research is needed both to predict how older patients with cancer will react to treatment, and to determine how information from emerging evidence can be best used to assist physicians' treatment decisions.