发表时间： 2014-09-11 作者： 分类： 医学大数据文献 来源：
We read with interest the article by Hershmann et al entitled Erythropoiesis Stimulating Agent [ESA] Use After Changes in Medicare Reimbursement Policy that reported the rapid decline in ESA usage by Medicare beneficiaries after changes to reimbursement policies. We note that the data were derived from an analysis of 121,169 patients in the SEER-Medicare linked database that used the SEER data to identify breast cancer cases in the population and the Medicare data to identify claims for ESA use by those patients. Claims data, however, are a surrogate for actual medical care and may not accurately reflect either the extent or reasons for the use of medical services.
ASCO has embarked on building a rapid learning system for oncology called CancerLinQ. To test the feasibility of this project, ASCO built a prototype with funding from Susan G. Komen.2 More than 170,000 de-identified records of patients with breast cancer, containing all structured data elements but not clinical notes, were put into the prototype from multiple different electronic health records (EHR) systems at 29 different oncology practices around the country. The prototype successfully achieved the five main goals set out: (1) capture and aggregate complete longitudinal patient records from any source, in any format, and make use of the data; (2) provide real-time clinical decision support based on clinical guidelines and integrate that into a demonstration EHR system; (3) measure clinical performance on a subset of Quality Oncology Practice Initiative performance measures; (4) explore and generate hypotheses from clinical data; and (5) provide lessons learned about technical and logistical challenges of CancerLinQ implementation.