Using Self-Reported Data to Segment Older Adult Populations with Complex Care Needs.

Area of Science:

• Geriatric Medicine
• Health Services Research
• Data Science in Healthcare

Background:

• Effective care management necessitates segmenting diverse patient populations.
• Patient-reported data offers insights beyond traditional clinical information.

Observation:

• Retrospective analyses of 9,617 older adults (age 65+) at high risk for utilization were performed using Medicare Health Risk Assessment (HRA) data.
• Clustering and latent class analyses (LCA) were applied to HRA variables, including self-reported quality of life, mood, activities of daily living (ADL), and advance directives.
• Demographic, utilization, and clinical characteristics were used to describe the identified patient subgroups.

Findings:

• Cluster analysis yielded 14 distinct subgroups, while LCA identified 8 distinct classes, each representing unique care needs.
• Identified groups included frail individuals with cognitive impairment, those with combined physical and mental health challenges, and individuals with functional limitations but good well-being.
• Subgroups were differentiated by factors such as age, advance directive status, and tobacco use.

Implications:

• Population segmentation using patient-reported data can lead to more personalized and effective care management strategies.
• Both clustering and LCA methods provide actionable insights, with clustering offering more intuitive subgroup definitions.
• Utilizing patient-reported data in segmentation can enhance care efficiency and promote a more patient-centered approach to managing complex health needs.