PURPOSE: Considerable variation among surgeons exists in the current practice of patient surveillance after colon cancer treatment. We evaluated whether geographic factors are responsible for this observed variation. METHODS: Profiles of hypothetical patients and a detailed questionnaire based on the profiles were mailed to 2,733 members of two national surgical societies. The influence of the geographic location of the respondents on practice patterns were assessed in two ways. Repeated-measures analysis of variance was used to compare the practice patterns among 19 large metropolitan statistical areas (MSAs) and chi 2 analysis was used to determine whether these patterns differed by MSA population size. RESULTS: Seven of nine commonly used surveillance modalities were ordered significantly more frequently with increasing tumor-node-metastasis (TNM) stage and significantly less frequently with year postsurgery among the 995 respondents with assessable responses, but MSA population size and geographic location of physicians generally had no effect on documented practice variability. The remaining two modalities (bone scan and computed tomography [CT]) were used so infrequently as to preclude meaningful analysis. CONCLUSION: Surveillance after potentially curative colon cancer surgery for otherwise healthy patients is not significantly affected by the geographic location of the surgeon who performs the surveillance testing and only modestly affected by the population size of the MSA in which he/she practices. These data should help in the design of prospective trials of this topic.
PURPOSE: Considerable variation among surgeons exists in the current practice of patient surveillance after colon cancer treatment. We evaluated whether geographic factors are responsible for this observed variation. METHODS: Profiles of hypothetical patients and a detailed questionnaire based on the profiles were mailed to 2,733 members of two national surgical societies. The influence of the geographic location of the respondents on practice patterns were assessed in two ways. Repeated-measures analysis of variance was used to compare the practice patterns among 19 large metropolitan statistical areas (MSAs) and chi 2 analysis was used to determine whether these patterns differed by MSA population size. RESULTS: Seven of nine commonly used surveillance modalities were ordered significantly more frequently with increasing tumor-node-metastasis (TNM) stage and significantly less frequently with year postsurgery among the 995 respondents with assessable responses, but MSA population size and geographic location of physicians generally had no effect on documented practice variability. The remaining two modalities (bone scan and computed tomography [CT]) were used so infrequently as to preclude meaningful analysis. CONCLUSION: Surveillance after potentially curative colon cancer surgery for otherwise healthy patients is not significantly affected by the geographic location of the surgeon who performs the surveillance testing and only modestly affected by the population size of the MSA in which he/she practices. These data should help in the design of prospective trials of this topic.