Development of Data Infrastructure to Monitor Lung Cancer Screening Use and Outcomes in Four Cancer Research Network Sites
lung cancer, screening
Background/Aims: With the National Lung Screening Trial finding a 20% reduction in lung cancer mortality among individuals screened by low-dose computed tomography (LDCT) versus chest radiography, and the subsequent U.S. Preventive Services Task Force (USPSTF) grade B recommendation, lung cancer screening is reemerging as a priority area in cancer prevention research. Questions remain, however, as to whether the ratio of benefits to harm observed in the National Lung Screening Trial will differ in routine clinical practice. Furthermore, in its National Coverage Determination, the Centers for Medicare & Medicaid Services (CMS) required that all facilities performing LDCT submit screening data to an approved registry. To facilitate research, quality improvement and compliance with CMS requirements, there is a need to create data systems to monitor LDCT screening.
Methods: The Cancer Research Network (CRN) provides infrastructure support to promote cancer research within integrated health care systems. Four CRN sites received pilot funding from the National Cancer Institute starting in December 2014 to develop capacity to collect data regarding several aspects of LDCT screening: eligibility (e.g. smoking status), exam indication (e.g. screening or surveillance), exam results (Lung-RADS category, nodule characteristics), and early outcomes (lung cancer detection, follow-up diagnostic testing).
Results: The four sites include the Henry Ford Health System and three Kaiser Permanente regions (Colorado, Northern California and Southern California). All sites use USPSTF recommendations (or slight variations) for screening eligibility; collectively, approximately 150,000 health plan members within these systems are eligible for screening. Implementation of screening ranged from a small pilot program starting at two medical centers to more broad availability at the other sites, with varying degrees of outreach to promote screening. Strategies to collect information on LDCT include referral templates that ascertain eligibility and exam indication, use of hashtags for extraction of Lung-RADS category from radiology reports, retrieval of structured data from electronic health records and natural language processing to ascertain nodule characteristics.
Conclusion: Screening registry data collection is often labor-intensive, requiring manual abstraction of medical record information. The development of strategies and tools to efficiently extract LDCT screening data from medical record systems will facilitate research to assess the utilization and outcomes of lung cancer screening as it is implemented in community settings.
Doria-Rose VP, Sadoka LC, Neslund-Dudas CM, Ritzwoller DP, Feigelson HS, Simoff MJ, Kushi LH, Gould MK. Development of data infrastructure to monitor lung cancer screening use and outcomes in four Cancer Research Network sites. J Patient Cent Res Rev. 2016;3:172.