Background: Approximately 40% of persons will develop cancer in their lifetime and a considerable proportion of these individuals will develop a subsequent mental health condition. Building from the work of Forbes et al. (2024), we sought to quantify the real-world and contemporary incidence of mental health conditions post cancer diagnosis, as well as to identify populations and time periods where risk is more pronounced in the United States (US).

Methods: Data were derived from the US based Optum® Claims Extended Date of Death dataset (January 2016 through December, 2024). Individuals with a leukemia or non-Hodgkin lymphoma (NHL) diagnosis were matched to individuals without a cancer diagnosis (1:1 nearest neighbor without replacement). With a standardized mean difference threshold of 0.25, this approach yielded adequate balance for most covariates. Primary outcomes of interest included any depressive disorder (defined as at least one diagnostic code for depressive disorders, using a broad list of MedDRA, SNOMED and OMOP Extension codes including mild depression, postpartum depression, chronic depression, recurrent moderate major depressive disorder co-occurrent with anxiety, severe recurrent major depression, dysthymia and other related disorders), and any anxiety disorder (defined as at least one diagnosis code for anxiety disorders using a broad list of LONIC, MedDRA, SNOMED and OMOP Extension codes including social phobia, generalized anxiety disorder, obsessive-compulsive disorder, paruresis, posttraumatic stress disorder, anxiety disorder caused by stimulant and other related disorders). A washout period of 365-days was used to help ensure that any observed outcomes were new-onset depressive and anxiety disorders. Incidence rates (IRs) of depressive and anxiety disorders were calculated, and adjusted hazard ratios (aHRs) with 95% confidence intervals (CIs) were derived from multivariate Cox models to examine risk conferred by cancer diagnosis at 30-days, 180-days, 1-year, and 3-years post cancer diagnosis.

Results: For the broader outcome ofdepressive disorders,each matched cohort was comprised of 80,224 individuals. While a stepwise increase in incidence of depressive disorders was present across the four time points in each matched cohort, incidence of depressive disorders was consistently higher among those with hematologic cancer (IR per 100,000 persons at 30-days [2,180], 180-days [12,643], 1-year [21,157], and 3-years [42,108]) as compared to those with no cancer diagnosis (IR per 100,000 persons at 30-days [543], 180-days [5,212], 1-year [12,246], and 3-years [33,147]). For depressive disorders, there was a stepwise decrease in effect size, with effect size appearing most pronounced at 30-days post cancer diagnosis in the hematologic cancer cohort (aHR, 95% CI at 30-days [2.73, 2.39-3.11], 180-days [1.8, 1.69-1.92], 1-year [1.51, 1.44-1.59], and 3-years [1.29, 1.24-1.34]). A comparable pattern was observed when examining leukemia and NHL individually.

For the outcome of anxiety disorders, 79,480 individuals remained in each cohort post-match Among those with hematologic cancer, incidence of anxiety disorders was also consistently elevated at each time point (IR per 100,000 persons at 30-days [3,674], 180-days [17,235], 1-year [27,006], and 3-years [48,032]) compared with the matched no cancer diagnosis cohort (30-days [757], 180-days [6,475], 1-year [14,461], and 3-years [37,021]). The effect size was greater for anxiety disorders than was observed for depressive disorders at each time point (aHR, 95% CI at 30-days [3.84, 3.45-4.28], 180-days [2.14, 2.03-2.26], 1-year [1.76, 1.68-1.83], and 3-years [1.47, 1.42-1.52]).

Conclusion: With an effect size most pronounced at 30-days for the occurrence of any depressive and any anxiety disorder in hematologic malignancy patients, and with consistently increased incidence rates at 3-years post hematologic cancer diagnosis, the importance of early and long-term mental health prevention and support post cancer diagnosis is one aspect underscored by these findings. There remains a growing unmet need for further exploration of these observations and risk mitigation strategies.

This content is only available as a PDF.
2025
Sign in via your Institution