OBJECTIVES: Long waiting times for hip and knee total joint replacement (TJR) are a major concern in health care systems with universal health care coverage. Giving patients the option of choosing the next available surgeon, as opposed to waiting for a specific surgeon, may improve access to care. We measured patient preferences about referral choices and waiting times using a discrete choice experiment (DCE) survey. METHODS: The attributes and levels informing DCE choice tasks were derived through a literature review, clinical and policy experts, and seven focus groups in four Canadian centers. 176 consecutive patients referred to an orthopaedic surgeon for TJR at two sites completed the survey of 14 choice tasks, each with 5 attributes (reputation, process of referral from primary care to specialist, waiting time to consult, waiting time to surgery, travel time to hospital) and 3 to 6 attribute levels. We tested different specifications for the surgeon reputation attribute and two alternative specifications of status quo/opt-out to inform the full study. We estimated preferences using conditional logit regression. RESULTS: Poor surgeon reputation dominated in one of four survey versions. Using the remaining respondents (n=135), surgeon reputation was the most important attribute in each version. ‘Not knowing’ the surgeon reputation was as important as an ‘excellent’ reputation. The waiting time to surgery was the next most important attribute, and was more important than waiting time to consult and the remaining attributes. There was no statistically significant difference between ‘waiting for a specific surgeon’ or the ‘next available surgeon’. Preferences in alternative specifications of status quo/opt-out were statistically different. CONCLUSIONS: The surgeon reputation is a key element of surgeon selection, and patients are willing to choose the next available surgeon in order to reduce waiting times providing that they know the surgeon has at least a satisfactory reputation.