Abstract
Introduction: There is little information describing the trajectories of depression and anxiety symptomatology in the context of coronary heart disease (CHD), and their comparison according to sociodemographic and disability measures, cardiac risk factors, and health care costs.
Methods: Using a primary care cohort of 803 patients with a diagnosis of CHD, a latent class growth curve model was developed to study the distinct trajectories of depression and anxiety symptoms (using the hospital anxiety and depression scale) over a 3-year period comprised of 7 distinct follow-up points. Multinomial regression analysis was then conducted to study the association between latent classes, baseline risk factors, and total health care costs across time.
Results: The 5-class model yielded the best combination of statistical best-fit analysis and clinical correlation. These classes were as follows: “stable asymptomatic” (n = 558), “worsening” (n = 64), “improving” (n = 15), “chronic high” (n = 55), and “fluctuating symptomatology” (n = 111). The comparison group was the “stable asymptomatic” class. The symptomatic classes were younger and had higher proportion of women, and were also associated with non-white ethnicity, being a current smoker, and having chest pain. Other measures of disease severity, such as a history of myocardial infarction and co-morbidities, were not associated with class membership. The highest mean total health care costs across the 3 years were the “chronic high” and “worsening” class, with the lowest being the “improving” and “stable low” classes. The total societal costs for patients in the “chronic high” and “worsening” class were significantly higher, as compared to the “stable low” class.
Methods: Using a primary care cohort of 803 patients with a diagnosis of CHD, a latent class growth curve model was developed to study the distinct trajectories of depression and anxiety symptoms (using the hospital anxiety and depression scale) over a 3-year period comprised of 7 distinct follow-up points. Multinomial regression analysis was then conducted to study the association between latent classes, baseline risk factors, and total health care costs across time.
Results: The 5-class model yielded the best combination of statistical best-fit analysis and clinical correlation. These classes were as follows: “stable asymptomatic” (n = 558), “worsening” (n = 64), “improving” (n = 15), “chronic high” (n = 55), and “fluctuating symptomatology” (n = 111). The comparison group was the “stable asymptomatic” class. The symptomatic classes were younger and had higher proportion of women, and were also associated with non-white ethnicity, being a current smoker, and having chest pain. Other measures of disease severity, such as a history of myocardial infarction and co-morbidities, were not associated with class membership. The highest mean total health care costs across the 3 years were the “chronic high” and “worsening” class, with the lowest being the “improving” and “stable low” classes. The total societal costs for patients in the “chronic high” and “worsening” class were significantly higher, as compared to the “stable low” class.
| Original language | English |
|---|---|
| Pages | S312-S313 |
| Number of pages | 2 |
| DOIs | |
| Publication status | Published - 1 Apr 2017 |