Trends in lung cancer survival: superior outcomes in US veterans compared to general population
Menée aux Etats-Unis à partir de données 2007-2019 portant sur 1 463 996 patients atteints d'un cancer du poumon non à petites cellules, cette étude examine les tendances en matière d'évolution de la survie globale en fonction de deux systèmes de santé
Background : Non-small cell lung cancer (NSCLC) remains the leading cause of cancer-related mortality in the US, yet survival outcomes are shifting with evolving therapies and screening. Recent findings suggest that survival improvements vary across healthcare systems, with favorable trends observed in the Veterans Health Administration (VHA). We sought to compare NSCLC survival trends across US health systems.
Methods : We conducted a retrospective cohort study of patients with NSCLC (2007 to 2019) using data from the VHA and the National Cancer Database (NCDB), representing the non-VHA US general population. The primary exposure was health system (VHA vs non-VHA), and the primary outcome was 3-year all-cause mortality. Multivariable Cox models estimated annual adjusted overall survival (OS), with interaction terms for health system and stage.
Results : Among 1,463,996 patients (VHA n = 79,027; NCDB n = 1,384,96), adjusted 3-year OS increased from 24% to 51% (VHA) and from 24% to 41% (non-VHA) between 2007 and 2019. The VHA survival advantage persisted in analyses stratified by stage (all p-values <0.0001), and when analyses were limited to NCDB patients with Medicare or private insurance (51% vs 42% in 2019; p < 0.0001)
Conclusions : In this nationally representative study, 3-year OS among patients with NSCLC improved between 2007 and 2019, with larger and more rapid gains observed within the VHA compared to non-VHA settings. These findings suggest that lung cancer care delivered within an integrated, publicly funded system is associated with greater and more rapidly improving survival than care delivered across a heterogeneous mix of U.S. healthcare delivery settings represented in the NCDB.
JNCI Cancer Spectrum , article en libre accès, 2026