Article

Identifying subgroups of complex patients with cluster analysis

Institute for Health Research, Kaiser Permanente Colorado, Denver, USA.
The American journal of managed care (Impact Factor: 2.17). 08/2011; 17(8):e324-32.
Source: PubMed

ABSTRACT To illustrate the use of cluster analysis for identifying sub-populations of complex patients who may benefit from targeted care management strategies.
Retrospective cohort analysis.
We identified a cohort of adult members of an integrated health maintenance organization who had 2 or more of 17 common chronic medical conditions and were categorized in the top 20% of total cost of care for 2 consecutive years (n = 15,480). We used agglomerative hierarchical clustering methods to identify clinically relevant subgroups based on groupings of coexisting conditions. Ward's minimum variance algorithm provided the most parsimonious solution.
Ward's algorithm identified 10 clinically relevant clusters grouped around single or multiple "anchoring conditions." The clusters revealed distinct groups of patients including: coexisting chronic pain and mental illness, obesity and mental illness, frail elderly, cancer, specific surgical procedures, cardiac disease, chronic lung disease, gastrointestinal bleeding, diabetes, and renal disease. These conditions co-occurred with multiple other chronic conditions. Mental health diagnoses were prevalent (range 28% to 100%) in all clusters.
Data mining procedures such as cluster analysis can be used to identify discrete groups of patients with specific combinations of comorbid conditions. These clusters suggest the need for a range of care management strategies. Although several of our clusters lend themselves to existing care and disease management protocols, care management for other subgroups is less well-defined. Cluster analysis methods can be leveraged to develop targeted care management interventions designed to improve health outcomes.

3 Followers
 · 
180 Views
  • Source
    • "For example, apart from the technical difficulties implicit to the extraordinary numbers of theoretically possible combinations of diseases, observedto-expected ratios depend on the prevalence of the individual diseases and do not adequately adjust for chance multimorbidity when nonrandom multimorbidity exists, as is the case for other measures of association such as odds and risk ratios [35]. Cluster analysis does not allow for health problems to simultaneously belong to more than one cluster unless patients are grouped instead of diseases, which occurred in two studies included in this review [21] [29]. Recently, Ng et al. [36] developed a unified clustering algorithm that enabled the overlapping of clusters. "
  • Source
    • "For example, apart from the technical difficulties implicit to the extraordinary numbers of theoretically possible combinations of diseases, observedto-expected ratios depend on the prevalence of the individual diseases and do not adequately adjust for chance multimorbidity when nonrandom multimorbidity exists, as is the case for other measures of association such as odds and risk ratios [35]. Cluster analysis does not allow for health problems to simultaneously belong to more than one cluster unless patients are grouped instead of diseases, which occurred in two studies included in this review [21] [29]. Recently, Ng et al. [36] developed a unified clustering algorithm that enabled the overlapping of clusters. "
  • Source
    • "For example, apart from the technical difficulties implicit to the extraordinary numbers of theoretically possible combinations of diseases, observedto-expected ratios depend on the prevalence of the individual diseases and do not adequately adjust for chance multimorbidity when nonrandom multimorbidity exists, as is the case for other measures of association such as odds and risk ratios [35]. Cluster analysis does not allow for health problems to simultaneously belong to more than one cluster unless patients are grouped instead of diseases, which occurred in two studies included in this review [21] [29]. Recently, Ng et al. [36] developed a unified clustering algorithm that enabled the overlapping of clusters. "
Show more