ESTRO 2025 - Abstract Book

S2337

Interdisciplinary – Health economics & health services research

ESTRO 2025

4518

Digital Poster Impacts of healthcare insurance sector on NSCLC diagnosis and survival: a population-based study in Brazil Fernanda Malucelli Favorito 1 , Brooke E Wilson 2,3 , Fabio Y Moraes 2,3 1 Medicina, Faculdade de Ciências Médicas da Santa Casa de São Paulo, São Paulo, Brazil. 2 Department of Oncology, Queen’s University, Kingston, Canada. 3 Division of Cancer Care and Epidemiology, Sinclair Cancer Research Institute, Kingston, Canada Purpose/Objective: The Brazilian healthcare system operates through two main sectors: public and private. In the public sector, universal coverage is provided by the State to all citizens at municipal, state, and federal levels. The private sector includes private insurance plans, with or without out-of-pocket expenses, as well as uninsured, entirely out-of pocket services. Due to the distinct nature of these systems and their users, differences in care and outcomes are anticipated 1,2 . This study evaluated NSCLC population characteristics, treatment patterns, and overall survival (OS) across healthcare systems. Material/Methods: This retrospective cohort study of pathology-confirmed NSCLC in patients over 18 years-old with available health insurance data from 2000-2020 in the state of São Paulo, Brazil. Data were sourced from Fundação Oncocentro de São Paulo (FOSP) publicly available database. Chi-square (χ²) tests and Kaplan-Meier survival analysis were performed using GraphPad Prism v10.3.1 (GraphPad Software, San Diego, CA, USA). Survival comparisons and trends across systems and stages were evaluated by log-rank/Mantel-Cox tests. Results: The analysis included 21,618 patients, distributed between the public (86.65%) and private (11.49% insured vs 1.86% uninsured) systems. Surgery and immunotherapy were the treatments with higher discrepancy in usage between systems, while radiotherapy and chemotherapy had more proportional ranges, as shown in Table 1 (cohort characteristics). Histology was balanced across systems, though early-stage diagnoses were more prevalent within the private system. Median OS was 0.83 years for public system and 2.00 years for private system users (p<0.0001). Median stage-specific OS was higher for private system users, particularly for stages I (6.75 vs 4.00 years, p<0.0001) and IV (1.08 vs 0.58 years, p<0.0001), which held true for stages II (5.25 vs 1.91 years, p<0.0001) and III (2.00 vs 1.08 years, p<0.0001). Apart from stages III and IV, where median OS for the public system was equal to that of the overall NSCLC population (combined public and private systems), the public system consistently exhibited lower survival compared to the overall NSCLC population. For the latter, OS was 0.91 years, with stage-specific OS of 4.58 (stage I), 2.16 (stage II), 1.08 (stage III) and 0.58 (stage IV) years. Notably for all stages, total sample OS was greater/equal than public and lower than private OS. Survival trends deemed statistically significant (p<0.005) are shown in Figure 1.