Statistical Analyses

Each center was represented only once in the analysis. For centers with multiple respondents, we selected a single participant to represent the center using an a priori selection algorithm, as described in previous studies (6,7,15); similar to the US survey, for this survey the algorithm prioritized responses from transplant surgeons, followed by transplant physicians, and other providers. For centers with more than one response from a respondent in the same role at the same transplant center, after the above two steps, we retained the earliest submitted survey. Responses to each survey question were described using percentages and frequencies. To derive response rates, we divided the number of center responses by the total number of centers that responded to the question, such that percentages reflect proportions of respondents, as per previously described methods (5–7,15). For questions where participants were asked to “select all that apply,” the denominator for calculating percentages was the number of participants who submitted a response to the question. For the “select all that apply” questions, column total percentage may exceed 100% because each respondent was permitted to select more than one option. Free-text comments were assessed qualitatively and grouped into thematic categories by the investigators. All analyses were performed using SAS for Windows v14.0 (SAS Institute, Cary, NC).