Centralization of Lung Cancer Screening and Adherence to First Follow-up Assessment: The Veterans Health Administration Experience
Lawrence N Benjamin, Eduardo R Núñez, Lillian Chen, Yash Motwani, Anita Yuan, Sitaram Vangala, Christopher G Slatore, Renda Soylemez Wiener, David A Elashoff, Elizabeth M Yano, Donna L Washington, Carol M MangioneAbstract
Rationale
Lung cancer screening (LCS) decreases lung cancer mortality, but patients must receive timely follow-up to realize these benefits. Unfortunately, real-world LCS follow-up adherence remains unacceptably low. LCS centralization into dedicated care teams and tracking resources (compared to decentralized, individual provider-led screening) may be a promising strategy for improving LCS adherence. However, it remains unclear if centralization’s benefits are similar across patient demographics, and whether hybrid or fully-centralized programs’ performance differ.
Objectives
To investigate if LCS centralization is associated with increased adherence to follow-up by comparing decentralized, hybrid, and fully-centralized programs, and assess if adherence varies among patient subgroups.
Methods
We performed a retrospective, nationwide cohort study using Veterans Health Administration data via a difference-in-differences analysis of quarterly facility-level LCS adherence rates. We included Veterans aged 55-80 years old who entered screening from October 1, 2015 through September 30, 2021. The primary exposure was facility LCS program centralization status at time of Veteran entry into screening. The primary outcome was adherence to first follow-up recommendations for the initial screening study based on Lung CT Screening Reporting & Data System (Lung-RADS) score.
Results
146,321 Veterans were analyzed. Average LCS adherence overall was 58.5%. Compared to decentralized programs, both hybrid and fully-centralized programs were associated with improved adherence, (hybrid programs OR 1.16 [95% CI 1.11, 1.22]; fully-centralized programs OR 1.13 [1.07, 1.20]) in models that additionally adjusted for race & ethnicity, sex, age, out-of-pocket cost, residential rurality, Area Deprivation Index, and medical comorbidity. There was no statistically significant difference between hybrid and fully-centralized programs. In subgroup analyses, improved adherence at centralized programs was primarily associated with Lung-RADS 1 & 2 scans. Compared with decentralized programs, Fully-centralized programs were associated with improved adherence for Black Veterans, hybrid programs for rural Veterans, and both for white Veterans and urban Veterans. However, certain sociodemographic groups had lower odds of adherence regardless of program centralization.
Conclusions
Our large, retrospective nationwide cohort study suggests centralized LCS programs are associated with comparably higher odds of LCS adherence compared to decentralized programs, increasing the likelihood patients benefit from screening. However, adherence rates, even at centralized facilities, remain suboptimal.