Association of Depression and Anxiety With the Accumulation of Chronic Conditions

Key Points Question Are depression and/or anxiety associated with higher risk and higher rates of accumulating chronic conditions than having neither condition? Findings In this cohort study including 40 360 individuals, the risk of accumulating chronic conditions was significantly higher in women with depression and comorbid depression and anxiety in each of 3 age cohorts (anchored at the 20th, 40th, and 60th birthdays) compared with individuals without depression or anxiety, with similar observations for men in the cohort aged 20 years. Rates of accumulation of chronic conditions were highest for women and men with combined depression and anxiety. Meaning These findings suggest that depression and anxiety may be associated with higher rates of acquiring chronic conditions and that these associations may be magnified when depression and anxiety cooccur.

b Adjusted mortality analyses are weighted using weights developed from generalized boosting models (GBM) to balance the four exposure groups at index birthday on potential confounders including calendar year of birthday, race, ethnicity, education, smoking status, BMI, and prevalence of each of the 15 chronic conditions (hypertension, congestive heart failure, coronary artery disease, cardiac arrhythmias, hyperlipidemia, stroke, arthritis, asthma, cancer, chronic kidney disease, chronic obstructive pulmonary disease, diabetes, hepatitis, HIV infection, and osteoporosis).
c HR was non-estimable because zero deaths were observed in the stratum. eFigure 1. Inclusion of Individuals in Birthday Age Cohorts and Accumulation of Chronic Conditions eFigure 1. The complete populations identified by the REP Census are included in the yellow boxes. Persons were subsequently excluded for either 1) insufficient data to define study variables or for 2) presence of conditions at index included in exclusion criteria (pink boxes). The numbers of persons included in final analyses within each age cohort and separately by exposure group are summarized in the green boxes. The numbers of accumulated chronic conditions that were observed during follow-up are summarized in the blue boxes separately by exposure group and separately by sex. Abbreviations: Anx. alone = diagnosed anxiety disorder alone (without co-occurring depressive disorder diagnoses); Bipolar Dis. = bipolar spectrum disorders; Dep. alone = diagnosed unipolar depressive disorder alone (without co-occurring anxiety disorder diagnoses); Dep.-Anx. = co-occurring diagnosed unipolar depressive and anxiety disorders; Eating Dis. = eating disorders; Subs. Abuse = substance use disorders. Footnotes: a Insufficient historical diagnoses is defined as having less than two years of diagnosis codes in the five years before the index birthdate. b Exclusions based on ICD-9 diagnosis code groupings. c Exclusions based on DHHS-defined chronic conditions code groupings. eFigure 2. To account for potential confounders associated with the accumulation of chronic conditions, adjusted hazards ratios (HRs) were calculated using inverse probability weights (IPWs) derived from generalized boosting models (GBMs). The GBMs were constructed separately for men and women and separately within each birthday age cohort (ages 20, 40, and 60 years). The covariates included in the balancing were calendar year at index date (as an integer from 2005 through 2014), race (in categories as White, Black, Asian, Other/unknown), ethnicity (Hispanic, non-Hispanic), level of education (in categories as high school or less, some college, 4 years of college or more, and unknown), smoking status (never smoker, former smoker, current smoker, and unknown), body mass index (BMI; in categories as <25, 25 to <30, ≥ 30, and unknown), and prevalence of each of the 15 chronic conditions (hypertension, congestive heart failure, coronary artery disease, cardiac arrhythmias, hyperlipidemia, stroke, arthritis, asthma, cancer, chronic kidney disease, chronic obstructive pulmonary disease, diabetes, hepatitis, HIV infection, and osteoporosis). The figure shows the pre-weighting balance of covariates at index birthday (red circles), and the post-weighting balance of covariates at index birthday (green triangles). Note that the balance of almost all covariates improves with weighting, and that none of the covariates have an imbalance of greater than 0.25 absolute standardized differences after weighting.

Standardized differences: D-only
Absolute standardized differences, in SDs Absolute standardized differences, in SDs Absolute standardized differences, in SDs Unw eighted comparison Balanced (w eighted) comparison 2B. Balance of characteristics in the 40-year-old birthday cohort. eFigure 2. To account for potential confounders associated with the accumulation of chronic conditions, adjusted hazards ratios (HRs) were calculated using inverse probability weights (IPWs) derived from generalized boosting models (GBMs). The GBMs were constructed separately for men and women and separately within each birthday age cohort (ages 20, 40, and 60 years). The covariates included in the balancing were calendar year at index date (as an integer from 2005 through 2014), race (in categories as White, Black, Asian, Other/unknown), ethnicity (Hispanic, non-Hispanic), level of education (in categories as high school or less, some college, 4 years of college or more, and unknown), smoking status (never smoker, former smoker, current smoker, and unknown), body mass index (BMI; in categories as <25, 25 to <30, ≥ 30, and unknown), and prevalence of each of the 15 chronic conditions (hypertension, congestive heart failure, coronary artery disease, cardiac arrhythmias, hyperlipidemia, stroke, arthritis, asthma, cancer, chronic kidney disease, chronic obstructive pulmonary disease, diabetes, hepatitis, HIV infection, and osteoporosis). The figure shows the pre-weighting balance of covariates at index birthday (red circles), and the post-weighting balance of covariates at index birthday (green triangles). Note that the balance of almost all covariates improves with weighting, and that none of the covariates have an imbalance of greater than 0.25 absolute standardized differences after weighting.

Standardized differences: A and D
Absolute standardized differences, in SDs Absolute standardized differences, in SDs Absolute standardized differences, in SDs Depression alone vs. Referent 2C. Balance of characteristics in the 60-year-old birthday cohort. eFigure 2. To account for potential confounders associated with the accumulation of chronic conditions, adjusted hazards ratios (HRs) were calculated using inverse probability weights (IPWs) derived from generalized boosting models (GBMs). The GBMs were constructed separately for men and women and separately within each birthday age cohort (ages 20, 40, and 60 years). The covariates included in the balancing were calendar year at index date (as an integer from 2005 through 2014), race (in categories as White, Black, Asian, Other/unknown), ethnicity (Hispanic, non-Hispanic), level of education (in categories as high school or less, some college, 4 years of college or more, and unknown), smoking status (never smoker, former smoker, current smoker, and unknown), body mass index (BMI; in categories as <25, 25 to <30, ≥ 30, and unknown), and prevalence of each of the 15 chronic conditions (hypertension, congestive heart failure, coronary artery disease, cardiac arrhythmias, hyperlipidemia, stroke, arthritis, asthma, cancer, chronic kidney disease, chronic obstructive pulmonary disease, diabetes, hepatitis, HIV infection, and osteoporosis). The figure shows the pre-weighting balance of covariates at index birthday (red circles), and the post-weighting balance of covariates at index birthday (green triangles). Note that the balance of almost all covariates improves with weighting, and that none of the covariates have an imbalance of greater than 0.25 absolute standardized differences after weighting.

Standardized differences: A and D
Absolute standardized differences, in SDs Absolute standardized differences, in SDs Absolute standardized differences, in SDs