A marginal structural model analysis for the effect modification by education on the association between cancer diagnosis history and major depressive symptoms: Findings from Midlife Development in the U.S. (MIDUS)

Abstract

BackgroundLimited research has employed a longitudinal approach to investigate the role of education level as an effect modifier on the relationship between cancer diagnosis history and the experience of major depressive disorder (MDD) with a nationally representative sample. MethodsWe harnessed data from three installments of the MIDUS Longitudinal study (n = 7108). A Marginal Structural Model facilitated the investigation of associations between a history of cancer diagnosis, MDD, and potential modifying effects of education level. Inverse probability weighting helped manage confounding factors. ResultsFindings indicated that a cancer diagnosis made one year prior was linked with 3.741 times greater odds of experiencing MDD (95 % CI: 1.411–9.918, p < 0.01). This connection was absent for diagnoses made two years earlier. Among individuals with education up to high school, a recent cancer diagnosis significantly increased the likelihood of MDD in the subsequent wave by 3.45 times (95 % CI: 1.31–9.08, p < 0.05). This pattern was not apparent among better-educated individuals. LimitationsAs the exposure variable was dependent on self-reported questionnaires, recall bias could be a potential limitation. Moreover, unaccounted variables like genetic factors could introduce confounding. ConclusionsA recent cancer diagnosis, particularly among less educated individuals, correlated with an increased probability of MDD, while the impact was not observed for older diagnoses. These findings emphasize that the timing of a cancer diagnosis and education level need consideration in the mental health assessment of cancer survivors.