Geographic variation in the conduct of patient surveillance after lung cancer surgery.

Abstract

PURPOSE Considerable variation among surgeons exists in the current practice of patient surveillance after lung cancer treatment. We evaluated whether geographic factors are responsible for this observed variation. METHODS Profiles of hypothetical patients suitable for postoperative surveillance and a detailed questionnaire based on the profiles were mailed to the 3,700 members of the Society of Thoracic Surgery (STS). The influence of the geographic location of the respondents on practice patterns was assessed among eight large metropolitan statistical areas (MSAs) with sufficient numbers of respondents, among nine broad geographic areas (United States census regions), and by the population size of the MSA from which the respondents reported. RESULTS There were 2,009 responses (54% return rate); 768 of those respondents both operate on and provide long-term follow-up care for lung cancer patients. There were sizeable effects of tumor-node-metastasis (TNM) stage and year postsurgery on practice patterns. Respondents from the Los Angeles/Long Beach MSA consistently had the highest frequency of follow-up test usage and those from the Tampa/St Petersburg MSA usually had the lowest. This held true for most testing modalities and was consistent across TNM stages I to III and years 1 to 5 postsurgery. Follow-up strategy was generally most intensive in the largest MSAs (population size, 2.5 to 10 million). The STS respondents from the Pacific US census region generally used the most intensive follow-up strategies and those from the East North Central and Mountain regions often used the least intensive. The differences disclosed in all three analyses were small. CONCLUSION There is marked variation among STS members in surveillance strategy, and the determinants of testing intensity are complex and interrelated. TNM stage and year postsurgery clearly affect follow-up practice; this analysis provides the first evidence that geographic setting has rather little effect on the surveillance strategies of clinicians.