Chaos to complexity: leveling the playing field for measuring value in primary care

Area of Science:

• Health Services Research
• Clinical Informatics
• Population Health Management

Background:

• Primary care patient populations are increasingly complex, presenting challenges for resource allocation and care coordination.
• Effective risk stratification is crucial for identifying patients who may benefit most from targeted interventions.
• Existing models may not adequately capture the interplay of co-morbidities and social determinants of health.

Purpose of the Study:

• To develop and validate a risk-stratification model for primary care patients.
• To cluster patients based on shared co-morbidities and social determinants of health.
• To rank patient clusters by their likelihood of hospital and emergency department (ED) utilization within a practice.

Main Methods:

• Retrospective cohort analysis of 10,408 adult primary care patients.
• Utilized a two-part generalized linear regression model for predictive modeling of ED and hospital utilization.
• Employed agglomerative hierarchical clustering to identify patient subgroups with similar co-morbidities.

Main Results:

• Factors like specific disease clusters (e.g., renal disease), low clinic adherence, and high poverty rates were associated with increased utilization.
• A stable set of four patient clusters emerged from the model.
• While the 'multiple chronic condition' cluster had the most high-utilization patients, the 'renal disease' cluster had the highest proportion of high-utilization patients (67%).

Conclusions:

• Risk stratification, enhanced by disease clustering, effectively organizes primary care populations.
• This approach facilitates focused care coordination efforts.
• Maximizing the value of care by targeting high-risk patient clusters is achievable.