Pre-Diagnostic and Cancer-Associated Depression and Anxiety Among Patients with Hematologic Malignancies in the United States

Background: Little is known about mental health disorders among patients with hematologic malignancies. As depression and anxiety substantially impair the quality of life of patients with cancer, the Institute of Medicine recommends diagnosis and management of these disorders as an integral part of cancer care (Cancer Care for the Whole Patient, IOM, 2008). Moreover, it has been suggested that non-White patients are less likely to disclose depression and anxiety to physicians due to concerns about stigma, and that physicians also inadequately screen this population (Bell, Annals FM, 2011). We aimed to characterize the prevalence and sociodemographic predictors of pre-diagnostic and cancer-associated (CA) depression and anxiety among patients with hematologic malignancies in the United States. Methods: We conducted a retrospective analysis using the Surveillance Epidemiology and End Results (SEER)-Medicare database. Patients ≥ 67 years old diagnosed with a hematologic malignancy (lymphoma, myeloma, leukemia, or myelodysplastic syndromes) between 2000 and 2015 who died prior to December 31, 2016 were eligible for inclusion. We examined prevalence of pre-diagnostic depression or anxiety, defined as having at least one inpatient or two outpatient Medicare claims for depression or anxiety starting from 24 months to 1 month prior to their blood cancer diagnosis. We also examined prevalence of CA-depression or anxiety, defined as at least one inpatient or two outpatient claims for these conditions between 1 month prior to and 3 months after their blood cancer diagnosis. CA-depression or anxiety were mutually exclusive with pre-diagnostic depression or anxiety respectively. We performed univariable analysis to determine sociodemographic and clinical covariates of CA-depression or CA-anxiety. Next, we fit multivariable logistic regression models to characterize factors independently associated with these two outcomes, adjusting for potential confounders: blood cancer type, age, sex, race, marital status, income, education, comorbidity regardless of univariable significance, and additional variables with p