Background: The National Cancer Database (NCDB) is commonly used for analyzing survival outcomes in pancreatic cancer and has a robust number of patient-specific socioeconomic variables. Although these aggregated variables are commonly collected, their utility and impact on survival has infrequently been assessed. We aim to assess the distribution of income and education variables on their impact on survival models for patients who underwent resection for pancreatic ductal adenocarcinoma (PDAC). Methods: Patients undergoing definitive surgery for PDAC between 2004-2020 were identified from the NCDB. Variables of interest were income and education according to the patient’s ZIP code of residence. Pearson and Spearman correlation was assessed. Overall survival (OS) was calculated using Kaplan Meier method. The effects of income and education was calculated using Cox proportional Hazard (PH) models. Results: 69,362 patients met inclusion criteria. Median age was 66 (IQR 59-73), with 48.9% female. The correlation coefficient between income and education based on the 2020 American Community Survey data was 0.48 (p<0.01). Median OS from the time of diagnosis was 21 months. The 5-year OS was 19.4 and 22.7 months for groups with lowest and highest education (p<0.01). Group with missing education has a 5-year OS of 22.4 month, comparable to the group with highest education (p=0.82). The 5-year OS based on income quartiles was 18.7 and 22.5 months for the Q1 and Q4 (p<0.01). Group with missing income has a 5-year OS of 22.4 month, similar to the highest income group (p=0.85). When including combinations of income, education and their interaction, the concordance of Cox models was changed by less than 0.002. Conclusions: For patients who underwent definitive surgery for PDAC, the income and education variables are only mildly correlated, and their inclusion do not have significant impact on OS models. Investigators should reconsider the utility of the aggregate NCDB measures, as they many not fully elucidate income and education disparities and impact on survival.

*Concordance formodel without education & income = 0.648.