Novel Integration of Multiple Data Sources in the United States Department of the Air Force Missile Community Cancer Study: A Methodological Analysis.
Concerns regarding cancer incidence, particularly non-Hodgkin lymphoma, among U.S. Air Force missile community members prompted the initiation of the Missile Community Cancer Study. This phased, retrospective process study was designed around the unique needs of the Missile Community and adapted to these requirements.
A cohort of all active duty Department of the Air Force (DAF) service members from 1976 to 2010 (N∼1.82 million), including approximately 65,000 personnel in missile-related career fields, was created via multiple data sources through a phased approach. Phase 1A examined Department of Defense (DoD) electronic health records (EHR) (2001-2020); Phase 1B added DoD and Department of Veterans Affairs (VA) cancer registry data (1976-2020) and VA EHR (1991-2020); and Phase 1C analyzed cancer mortality using National Death Index data (1979-2020). Armed Forces Health Surveillance Division and Surveillance, Epidemiology, and End Results cancer incidence and mortality case definitions were applied.
In Phase 1A, 4,865 incident cancers were identified from the DoD EHR across the cohort, including 198 cases in the missile community. After adding DoD and VA registry data and VA EHR data in Phase 1B, an additional 50,359 cases were identified, including 1,641 missile community cases. Mortality analysis (Phase 1C) identified 37,100 cancer deaths across the cohort, including 1,145 in the missile community. Across Phases 1A, 1B, and 1C, no statistically significant excess incidence or mortality of non-Hodgkin lymphoma or the 13 other study cancers was observed in the missile community compared to other DAF members or the U.S. population. The overlap between EHRs and cancer registries within both the VA and DoD systems found over half of EHR-identified cases were also included in their respective cancer registries and 76.1% of all Phase 1 cancer cases were captured in a registry.
This study's multi-phase approach with both internal and external comparison groups enhanced the comprehensiveness and accuracy of cancer case identification. The phased data linkage and validation approach provided a replicable framework for epidemiological cancer studies in military populations, illustrating methods for integrating and prioritizing data sources to address health concerns. The findings may ultimately guide future epidemiologic studies of cancer in the military to select optimal datasets, optimizing resource utilization and streamlining study timelines.
A cohort of all active duty Department of the Air Force (DAF) service members from 1976 to 2010 (N∼1.82 million), including approximately 65,000 personnel in missile-related career fields, was created via multiple data sources through a phased approach. Phase 1A examined Department of Defense (DoD) electronic health records (EHR) (2001-2020); Phase 1B added DoD and Department of Veterans Affairs (VA) cancer registry data (1976-2020) and VA EHR (1991-2020); and Phase 1C analyzed cancer mortality using National Death Index data (1979-2020). Armed Forces Health Surveillance Division and Surveillance, Epidemiology, and End Results cancer incidence and mortality case definitions were applied.
In Phase 1A, 4,865 incident cancers were identified from the DoD EHR across the cohort, including 198 cases in the missile community. After adding DoD and VA registry data and VA EHR data in Phase 1B, an additional 50,359 cases were identified, including 1,641 missile community cases. Mortality analysis (Phase 1C) identified 37,100 cancer deaths across the cohort, including 1,145 in the missile community. Across Phases 1A, 1B, and 1C, no statistically significant excess incidence or mortality of non-Hodgkin lymphoma or the 13 other study cancers was observed in the missile community compared to other DAF members or the U.S. population. The overlap between EHRs and cancer registries within both the VA and DoD systems found over half of EHR-identified cases were also included in their respective cancer registries and 76.1% of all Phase 1 cancer cases were captured in a registry.
This study's multi-phase approach with both internal and external comparison groups enhanced the comprehensiveness and accuracy of cancer case identification. The phased data linkage and validation approach provided a replicable framework for epidemiological cancer studies in military populations, illustrating methods for integrating and prioritizing data sources to address health concerns. The findings may ultimately guide future epidemiologic studies of cancer in the military to select optimal datasets, optimizing resource utilization and streamlining study timelines.
Authors
Beam Beam, Wolff Wolff, Fryman-Wynkoop Fryman-Wynkoop, Ruiz Ruiz, Escobar Escobar, Simon Simon, Guinn Guinn, Adkins Adkins, Crawford Crawford, Khalil Khalil, Kelly Kelly, Ramsey Ramsey, Britt Britt
View on Pubmed