Background
Pancreatic cancer is among the most lethal malignancies worldwide, with five-year survival below 10%. In Low- and Middle-Income Countries (LMICs), social and institutional inequalities strongly shape access to diagnosis and treatment, potentially widening survival gaps. Brazil’s universal health system coexists with marked racial and access-related inequities, offering a critical setting to examine how these factors operate along the cancer care pathway.
 
Objectives
To evaluate healthcare-related and social determinants of survival among patients with pancreatic cancer in Brazil, emphasizing how inequalities act at different stages of the therapeutic trajectory.
 
Methods
We conducted a retrospective cohort study of 1,426 adults with pancreatic adenocarcinoma treated at Brazil’s National Cancer Institute (INCA) between 2000 and 2023, using Hospital Cancer Registry data. Survival was analyzed across three sequential intervals: (1) consultation–death, (2) diagnosis–death, and (3) treatment–death. Kaplan–Meier estimates and log-rank tests assessed unadjusted differences. Multivariable Cox proportional hazards models estimated hazard ratios (HR) with 95% confidence intervals (CI), adjusting for race, referral source (public health system—SUS vs. private), prior diagnosis/treatment, and institutional follow-up.
 
Results
Median survival was extremely short before treatment initiation (diagnosis–death: 81 days; consultation–death: 75 days) and substantially longer after treatment start (170 days). In pre-treatment models, pronounced inequalities were observed. Black patients had shorter survival than white patients (consultation–death: 94.3 vs. 131.5 days; p<0.001), and patients referred through SUS survived less than those entering care outside SUS (101.4 vs. 123.2 days; p=0.05). Prior trajectory mattered: patients without prior diagnosis or treatment survived longer than those diagnosed but untreated (140.0 vs. 91.9 days). Institutional follow-up emerged as the strongest protective factor. In the consultation–death model, followed patients survived on average 368.8 days compared with 100.5 days among non-followed patients (difference: 268 days). Similar magnitudes persisted in the diagnosis–death model (372.7 vs. 109.2 days; difference: 263.5 days). In adjusted analyses, lack of prior treatment (HR=2.28; 95% CI: 1.34–3.89), Black race (HR=1.29; 95% CI: 1.08–1.54), and SUS referral (HR=1.17; 95% CI: 1.01–1.43) independently increased mortality risk before treatment. After treatment initiation, most social and access-related disparities attenuated and lost statistical significance, while institutional follow-up remained highly predictive (451.3 vs. 121.3 days), indicating that specialized care acts as a partial equalizer once access barriers are overcome.
 
Conclusions / Implications
Survival inequalities in pancreatic cancer in Brazil are concentrated before treatment initiation and are driven by race, referral pathways, and fragmented care trajectories. Ensuring timely diagnosis, equitable referral, and sustained institutional follow-up can substantially improve outcomes in highly lethal cancers. The study was fully led by Brazilian public health institutions, demonstrating LMIC leadership in analytical decision-making and generating policy-relevant evidence to strengthen equitable cancer care.