Language Preference Associated with Disparities in Annual Lung Cancer Screening Adherence in an Integrated Healthcare System

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Abstract Introduction Lung cancer screening (LCS) with annual low-dose CT reduces mortality, but benefit depends on timely adherence to annual screening. Language barriers may introduce additional challenges to screening adherence. We examined the association between non-English language preference and annual adherence to LCS. Methods We conducted a retrospective cohort study of LCS-eligible adults who underwent initial screening between January 1, 2015, and March 10, 2025, within a four-hospital regional health system serving a diverse urban and rural population in Massachusetts. The primary outcome was receipt of timely follow-up screening within the recommended 10–15-month interval across up to three subsequent annual screenings. The primary exposure was preferred language (non-English vs. English). We employed generalized estimating equation models with logit link, adjusting for age, sex, comorbidities, neighborhood deprivation, and screening round. Patients with Lung-RADS 3 or 4 results, death, or who reached age 80 between screening rounds were censored. Results Among 7,218 patients (mean age 62.6 ± 5.8 years; 48.8% female), 92.9% preferred English and 7.1% non-English. Non-English speakers were more frequently Hispanic (77.7% vs. 4.5%) and had higher prevalence of comorbidities and neighborhood disadvantage. After the initial screen, 59.6% completed follow-up within 10–15 months and 21.1% had no follow-up. Overall, 26.5% of the cohort completed all three follow-up scans within the recommended interval. Non-English speakers were less likely to complete screening within the recommended interval at every screening round (T1: 51.4% vs. 60.2%; T2: 57.5% vs. 64.8%; T3: 52.5% vs. 60.6%) and more likely to have no follow-up at each interval. Non-English language preference was independently associated with lower odds of completing timely annual screens (adjusted OR 0.79; 95% CI 0.66–0.95), with findings consistent in sensitivity analyses using inverse probability of censoring weighting. Conclusions In this large community-based LCS cohort, non-English language preference was independently associated with 21% lower odds of timely annual screening adherence after adjustment for comorbidities and socioeconomic factors. Linguistically tailored outreach, patient education, and navigation support are critical to advancing equitable lung cancer screening.