Start Date

20-5-2011 5:00 PM

End Date

20-5-2011 7:00 PM

Document Type

Event

Description

Background

Heart failure (HF) carries a high burden of comorbidity with approximately one half of patients with HF having at least one additional comorbid condition present. Rates of comorbidity in patients with HF have steadily increased over the past 2 decades.

Objective

To examine patterns of comorbidity among older patients with HF in the Cardiovascular Research Network PRESERVE cohort.

Methods

PRESERVE Cohort

Data are from the CVRN PRESERVE cohort which is a multicenter cohort of 37,054 patients [mean age = 74 years (SD = 12.4 yrs); 46% female] with HF diagnosed between 2005 and 2008 currently being conducted at 4 CVRN sites: KPNC, KPCO, KPNW, and FCHP. The primary data source for the PRESERVE cohort was the HMO Research Network Virtual Data Warehouse.

Identification of Coexisting Diseases

Coexisiting illnesses at the time of HF diagnosis were based on diagnoses and procedures mapped to relevant International Classification of Diseases, Ninth Edition (ICD-9) codes. For the purposes of characterizing clusters of comorbidities, we focused on coexisting conditions with a prevalence rate of ≥3%.

Statistical Analysis

We used the Agglomerative Clustering technique to characterize patterns of comorbidity. Over multiple iterations, each condition is clustered with the condition with which it has the highest squared correlation. This process is repeated to determine whether assigning a condition to a different cluster increases the amount of explained variance [ranging from 1.0 (all variance explained) to 0.0 (no variance explained)]. The conditions in each cluster are as correlated as possible among themselves and as uncorrelated as possible with conditions in other clusters.

Results

Burden of Comorbidity

There was a high degree of comorbidity and multi-morbidity among patients with HF. (Table 1) Hypertension and arrhythmias were the comorbidities of HF that occurred most often in the absence of other chronic conditions (4.8% and 4.7%, respectively). The average number of comorbid conditions varied from 3.5 to 5.2. Patients with HF and unstable angina or other thromboembolic disorders had the highest multi-morbidity (mean = 5.2 conditions), whereas those with HF and hypertension had the lowest (mean = 3.5).

Clustering of Comorbiditites

A five-cluster structure was derived.

  • Cluster 1: Dyslipidemia, Hypertension, Diabetes Mellitus, Visual Impairment
  • Cluster 2: Acute Myocardial Infarction, Unstable Angina, Thromboembolic Disorder, Dementia
  • Cluster 3: Aortic Valvular Disease, Cancer, Hearing Impairment, Arrthythmia
  • Cluster 4: Peripheral Arterial Disease, Stroke
  • Cluster 5: Lung Disease, Liver Disease, Depression

Discussion and Conclusions

  • Cluster analysis is an innovative approach to examining the co-occurrence of diseases and allows for identification of broad patterns of multi-morbidity beyond the pairings of diseases or disease counts.
  • Patients with HF have a high rate of multi-morbidity, with an average of 4 co-occurring conditions. Intuitive and unintuitive patterns of clustering were identified.
  • Randomized clinical trials in HF will need to include more diverse patient populations in order to adapt to the increasingly complex patient population.
  • A cluster analysis approach to characterizing patterns of comorbidity may help indentify important patient subgroups.

Creative Commons License

Creative Commons Attribution-Noncommercial-Share Alike 3.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 License.

Share

COinS
 
May 20th, 5:00 PM May 20th, 7:00 PM

Patterns of Complex Comorbidity in Older Patients with Heart Failure

Background

Heart failure (HF) carries a high burden of comorbidity with approximately one half of patients with HF having at least one additional comorbid condition present. Rates of comorbidity in patients with HF have steadily increased over the past 2 decades.

Objective

To examine patterns of comorbidity among older patients with HF in the Cardiovascular Research Network PRESERVE cohort.

Methods

PRESERVE Cohort

Data are from the CVRN PRESERVE cohort which is a multicenter cohort of 37,054 patients [mean age = 74 years (SD = 12.4 yrs); 46% female] with HF diagnosed between 2005 and 2008 currently being conducted at 4 CVRN sites: KPNC, KPCO, KPNW, and FCHP. The primary data source for the PRESERVE cohort was the HMO Research Network Virtual Data Warehouse.

Identification of Coexisting Diseases

Coexisiting illnesses at the time of HF diagnosis were based on diagnoses and procedures mapped to relevant International Classification of Diseases, Ninth Edition (ICD-9) codes. For the purposes of characterizing clusters of comorbidities, we focused on coexisting conditions with a prevalence rate of ≥3%.

Statistical Analysis

We used the Agglomerative Clustering technique to characterize patterns of comorbidity. Over multiple iterations, each condition is clustered with the condition with which it has the highest squared correlation. This process is repeated to determine whether assigning a condition to a different cluster increases the amount of explained variance [ranging from 1.0 (all variance explained) to 0.0 (no variance explained)]. The conditions in each cluster are as correlated as possible among themselves and as uncorrelated as possible with conditions in other clusters.

Results

Burden of Comorbidity

There was a high degree of comorbidity and multi-morbidity among patients with HF. (Table 1) Hypertension and arrhythmias were the comorbidities of HF that occurred most often in the absence of other chronic conditions (4.8% and 4.7%, respectively). The average number of comorbid conditions varied from 3.5 to 5.2. Patients with HF and unstable angina or other thromboembolic disorders had the highest multi-morbidity (mean = 5.2 conditions), whereas those with HF and hypertension had the lowest (mean = 3.5).

Clustering of Comorbiditites

A five-cluster structure was derived.

  • Cluster 1: Dyslipidemia, Hypertension, Diabetes Mellitus, Visual Impairment
  • Cluster 2: Acute Myocardial Infarction, Unstable Angina, Thromboembolic Disorder, Dementia
  • Cluster 3: Aortic Valvular Disease, Cancer, Hearing Impairment, Arrthythmia
  • Cluster 4: Peripheral Arterial Disease, Stroke
  • Cluster 5: Lung Disease, Liver Disease, Depression

Discussion and Conclusions

  • Cluster analysis is an innovative approach to examining the co-occurrence of diseases and allows for identification of broad patterns of multi-morbidity beyond the pairings of diseases or disease counts.
  • Patients with HF have a high rate of multi-morbidity, with an average of 4 co-occurring conditions. Intuitive and unintuitive patterns of clustering were identified.
  • Randomized clinical trials in HF will need to include more diverse patient populations in order to adapt to the increasingly complex patient population.
  • A cluster analysis approach to characterizing patterns of comorbidity may help indentify important patient subgroups.

 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.