Predictors of antidepressant use in older adults after pancreatic cancer diagnosis: A SEER-Medicare analysis
Presenter: Hussein Khalil, B Pharm;MS;PhD Session: Psychosocial and Behavioral Epidemiology, Health Services Research, Implementation Science, Pharmacoepidemiology, and Other Topics Time: 4/22/2026 9:00:00 AM → 4/22/2026 12:00:00 PM
Authors
Hussein Khalil 1 , Kourtney A. D. Byrd 1 , Yi Guo 2 , Shuang Yang 3 , Lisa Scarton 3 , Xiwei Lou 3 , Diana J. Wilkie 4 , Sherise C. Rodgers 5 , John M. Allen 1 1 Purdue University, West Lafayette, IN, 2 University of Florida College of Medicine, Gainesville, FL, 3 Department of Health Outcomes & Biomedical Informatics, University of Florida, Gainesville, FL, 4 Biobehavioral Nursing Science, University of Florida, Gainesville, FL, 5 Brown University Warren Alpert Medical School, Providence, RI
Abstract
Depression is highly prevalent among patients with pancreatic cancer (PC), driven by both the biological effects and the profound psychosocial burden of the diagnosis. Untreated depression contributes to poorer quality of life, reduced treatment adherence, increased symptom burden, and worse survival. Prior work suggests sociodemographic inequities may shape access to supportive care medications, yet little is known about how these factors operate specifically within older adults with PC. We aimed to characterize predictors of antidepressant use following pancreatic cancer diagnosis, with the goal of identifying inequities and informing patient-centered approaches to supportive care delivery in this high-risk population. We conducted a retrospective cohort study using SEER-Medicare data among patients aged ≥65 years diagnosed with primary PC between 2007 and 2019. Antidepressant use was identified from Medicare Part D claims up to 24 months post-diagnosis. We compared patient characteristics by antidepressant use status. Multivariable logistic regression models estimated adjusted odds ratios (ORs) and 95% confidence intervals (CIs). Among 81,535 patients with PC, 23.5% (N=19,181) received antidepressants post-diagnosis. Compared with those living in the highest-SES neighborhoods, patients in the lowest SES quintile had 19% lower odds of receiving antidepressants (aOR: 0.81; 95% CI 0.74-0.88). Increasing age was strongly associated with reduced use (≥80 vs. 65-69 years: aOR:0.60; 95% CI 0.57-0.63), while female sex was associated with higher odds (aOR:1.19; 95% CI 1.15-1.23). Marked racial/ethnic disparities were observed: non-Hispanic Black (aOR:0.63; 95% CI 0.59-0.67) and Asian (aOR:0.60; 95% CI 0.56-0.65) patients were substantially less likely to receive antidepressants compared with non-Hispanic White patients. Higher comorbidity modestly increased use (CCI per point: aOR:1.07; 95% CI 1.06-1.08). Strong associations were observed for prior supportive care medication use (aOR: 2.02; 95% CI 1.94-2.09) and hospice enrollment (aOR: 1.73; 95% CI 1.67-1.80). In our study, nearly a quarter of older adults with PC received antidepressants, with substantial variation by socioeconomic status, race/ethnicity, age. The markedly lower use among racial/ethnic minority and lower-SES patients indicates potential inequities in mental health treatment access and potential cultural influences. Increased use among those with higher comorbidity burden and hospice enrollment suggests that antidepressants are more often introduced in the context of advanced symptom burden or end-of-life care rather than early in the cancer course. These findings underscore the need for systematic depression screening across all demographic groups and proactive integration of mental health services into routine pancreatic cancer care.
Disclosure
H. Khalil, None.. K. A. D. Byrd, None.. S. Yang, None.. L. Scarton, None.. X. Lou, None.. D. J. Wilkie, None.. S. C. Rodgers, None.. J. M. Allen, None.
Cited in
Control: 4827 · Presentation Id: 5801 · Meeting 21436