A novel approach to measuring costs and efficiency: Lung nodules as a case study

每年有超过150万退伍军人接受胸部x光或CT扫描，以帮助评估常见症状

Over 1.5 million Veterans receive a chest x-ray or CT scan each year to help evaluate common symptoms such as shortness of breath. Improvements in imaging quality have led to nearly half of these routine scans finding small incidental lung nodules that are potentially worrisome for lung cancer. Although clinical guidelines exist for how to manage these indeterminate nodules, we do not know if providers follow recommended guidelines, or if these patients receive too little or too much follow-up imaging and testing. In this project we propose to identify all Veterans in FY2011 (estimated to be 200,0000 patients) who were identified with a lung nodule and determine whether they received guideline recommended care over the subsequent 24 months. A prior study of chest imaging reported that a suspicious finding generates over $1300 in annual follow-up costs. This would translate to over $500 million dollars a year in repeat imaging and testing costs for VHA due to incidental findings. Because this cost is potentially enormous, we will measure both whether a patient received guideline recommended care, as well as the cost of all the lung nodule care activities he or she received over this 24 month period. Preliminary evidence suggests there may be wide variation in the care Veterans receive when an incidental lung nodule is found. The goal of this project is to assess the quality of care Veterans receive and identify facilities that appear to be delivering the best quality care at the lowest cost. Our long term objective is to identify the approaches used at those facilities and disseminate those practices across VHA. In order to identify Veterans with an incidental lung nodule and measure adherence with recommended guidelines, we will access radiology text reports, which are now available nationally across VHA. This work will be accomplished using natural language processing to extract detailed information about the characteristics of each lung nodule from the text reports, as well as any changes reported in nodule characteristics as patients undergo repeated imaging during the 24 month period. We propose two study aims: Aim 1: Use natural language processing to operationalize a measure of adherence to Fleishner guidelines for managing lung nodules, and assess variation in guideline adherence across VHA. Aim 2: Measure utilization and costs of managing lung nodules, and assess variation in utilization and costs across VHA.