Co-Occurring Diseases and Mortality in Patients With Chronic Heart Disease, Modeling Their Dynamically Expanding Disease Portfolios: Nationwide Register Study

Figures

• 

  Figure 1. Example of how the statistical model works. (A) Illustration of an event sequence in which a hypothetical patient with heart disease (HD) receives the diagnosis of HD at time 0 and, subsequently, the hypertension, stroke, and cancer diagnoses at different times (measured in years following HD diagnosis) before death. (B) The corresponding longitudinal development of the hazard ratio of the patient relative to a theoretical patient who only has HD and is not multimorbid.

• 

  Figure 2. Diagnosis prevalence according to sex. Prevalence is reported at the time of heart disease (HD) diagnosis and for the entire span of the observed disease trajectories (Lifetime). COPD: chronic obstructive pulmonary disease.

• 

  Figure 3. Graphical representation of disease-disease interactions in the all interactions model. A ribbon connects chronic diseases that have any significant interaction (P<.001) between them. The connection’s width corresponds to the number of individuals diagnosed with HD developing both diseases throughout the observation period. The ribbon’s color represents the highest-order interaction relationship between 2 diseases. The ribbon chart is ordered by number of connections between diseases, starting from allergies (AL) with 5 connections all the way to cancer (CAN), which interacts with all the additional diseases. BP: back pain; COPD: chronic obstructive pulmonary disease; DEM: dementia; DEP: depression; DIA: diabetes; HC: high cholesterol; HT: hypertension; JD: joint disease; OA: osteoarthritis; OP: osteoporosis; SCH: schizophrenia; ST: stroke.

• 

  Figure 4. Difference in effect estimates for disease portfolios of increasing size for female and male individuals. Each bar represents a weighted average of the differences in effects between the additive only main effects (OME) model and the all interactions (ALL) model on the hazard scale exp(Inline Graphic 3). Thus, the bars indicate the average multiplier required to convert the hazard ratio (HR) from the OME model into the HR from the ALL model. The weights were determined based on the occurrence of each specific disease portfolio across the different educational attainment levels for each sex.

• 

  Figure 5. Effects of the 10 most frequent disease portfolio dyads (A), triads (B), tetrads (C), and pentads (D). Effects are shown for male individuals of varying educational attainment levels at the log-hazard rate scale. Comparisons are made to a male individual of the corresponding educational attainment level who only has heart disease (HD). Effects are presented for the all interactions model (different shades of blue) and the only main effects model (red). All comparisons are made at mean age and calendar time. HD is present in all disease portfolios. AL: allergies; BP: back pain; CAN: cancer; COPD: chronic obstructive pulmonary disease; DEP: depression; DIA: diabetes; HC: high cholesterol; HT: hypertension; OA: osteoarthritis; OP: osteoporosis; ST: stroke.

• 

  Figure 6. Effects of the 10 most frequent disease portfolio dyads (A), triads (B), tetrads (C), and pentads (D). Effects are shown for female individuals of varying educational attainment levels at the log-hazard rate scale. Comparisons are made to a female individual of the corresponding educational attainment level who only has heart disease (HD). Effects are presented for the all interactions model (different shades of blue) and the only main effects model (red). All comparisons are made at mean age and calendar time. HD is present in all disease portfolios. AL: allergies; BP: back pain; CAN: cancer; COPD: chronic obstructive pulmonary disease; DEP: depression; DIA: diabetes; HC: high cholesterol; HT: hypertension; OA: osteoarthritis; OP: osteoporosis; ST: stroke.

• 

  Figure 7. Effects of disease portfolio dyads (A), triads (B), tetrads (C), and pentads (D) involving stroke (ST), osteoporosis (OP), chronic obstructive pulmonary disease (COPD), dementia (DEM), and depression (DEP). Effects are shown for male individuals of varying educational attainment levels at the log-hazard rate scale. Comparisons are made to a male individual of the corresponding educational attainment level who only has heart disease (HD). Effects are presented for the all interactions model (different shades of blue) and the only main effects model (red). All comparisons are made at mean age and calendar time. HD is present in all disease portfolios.

• 

  Figure 8. Hazard ratios (HRs) of female (vs male) sex by disease portfolio and educational attainment level. Estimates for the 50 most common disease portfolios are shown with 99.9% CIs. The estimates are presented for each of the educational attainment levels: none, short, medium, and long, indicated by different shapes and always in ascending order from none to long. The reference group comprises male individuals with the same disease portfolio and educational attainment level. The disease portfolios are ordered by prevalence from left to right, with [hypertension (HT)] being the most frequent disease portfolio. All portfolios contain the heart disease (HD) condition, so it is not labeled in the plot. Therefore, the disease portfolio without a label in the plot (the second from the left) corresponds to the disease portfolio with only HD. AL: allergies; BP: back pain; CAN: cancer; COPD: chronic obstructive pulmonary disease; DEM: dementia; DEP: depression; DIA: diabetes; HC: high cholesterol; OA: osteoarthritis; OP: osteoporosis; ST: stroke.

• 

  Figure 9. Disease progression scenarios representing the mortality risk over time of a hypothetical (A) male individual with no education at mean age and calendar time who develops schizophrenia (SCH), cancer (CAN), and dementia (DEM) at 2.6, 5.6, and 7.5 years, respectively, following heart disease (HD) diagnosis; (B) male individual with no education who has hypertension (HT), high cholesterol (HC) at time of HD diagnosis and diabetes (DIA), chronic obstructive pulmonary disease (COPD), and CAN at 2.3, 4.8, and 6.1 years, respectively, following HD diagnosis; (C) male individual of varying educational attainment levels who develops depression (DEP), osteoporosis (OP), and DEM at 2.5, 5.9, and 7.5 years, respectively, following HD diagnosis under the all interactions (ALL) model (solid lines) and the additive only main effects (OME) model (dashed lines); and (D) male (green color) and female (red color) individual with no education who develops COPD, CAN, and DEM at 2.4, 5.8, and 8.0 years, respectively, following HD diagnosis under the ALL model (solid lines) and the additive OME model (dashed lines).