Racial and socioeconomic factors associated with palliative care utilization in pancreatic cancer: An analysis of National Inpatient Sample.

e16296 Background: Palliative care (PC) has been validated as a crucial component in end-of-life quality. Despite the higher PC utilization in cancer patients than other conditions, some studies have raised concern of systemic inequality in PC utilization. PC has a significant role in the care of pancreatic cancer, especially due to the poor prognosis. Here, we investigate the sociodemographic factors associated with the recent trend of PC utilization in pancreatic cancer. Methods: We used the Healthcare Cost and Utilization Project National Inpatient Sample database, all-payer inpatient care database in the United States, with data years 2016-2019. The end-of-life hospitalizations were defined as adult patients (age at least 18 years old) who were hospitalized at least three days and passed away. The cases were identified with International Classification of Diseases (ICD) 10th edition for pancreatic cancer (C25.0-25.9) within the top three diagnoses. Prevalence of PC consultation could be estimated with the billable code Z51.5. The factors that were statistically associated with PC utilization in univariable analysis would be included in the final multivariable logistic regression model. Charlson Comorbidity Index was used to adjust for comorbidities. Hospital factors such as size and base (rural, urban, urban teaching) were included in the analysis. Results: Among 339,318 adult patients who were hospitalized at least three days and passed away, 2,011 cases had pancreatic cancer within top three diagnosis codes. 1576 (78.4%) patients were at least age 60 or above. 1,275 (63.4%) of them had PC consult at their end-of-life. There was a significant yearly trend of increasing PC utilization (Cochran-Armitage test p-value < 0.05). Male, African American, lower income, and rural area were significantly associated with less PC consult in multivariable regression model. The adjusted odds ratio (aOR) of PC consult was 0.66 (p < 0.005) in African American. There was significantly more PC utilization in the top 25% income group (vs the bottom 25% counterpart, aOR 1.35; p < 0.05) and in urban hospitals (vs rural hospitals, aOR 1.80 in nonteaching; aOR 1.88 in teaching; p < 0.001). Conclusions: There was a trend of increasing PC utilization in pancreatic cancer. Nevertheless, racial and socioeconomic factors were still significantly associated with the PC utilization. Further study and systemic policy approach are required to investigate and address such disparities and promote the PC utilization.