1147 — Typologies of VA and Medicare Utilization among Dually Enrolled Veterans with Type 2 Diabetes: A Latent Class Analysis

Radomski TR, Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System; Zhao X, Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System; Thorpe CT, Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System; Good CB, Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System; Mor MK, Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System; Fine MJ, Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System; Gellad WF, Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System;

Objectives:
To determine the degree to which Veterans were dual users for specific types of diabetes care and to identify typologies of dual use using latent class analysis.

Methods:
We performed a cross-sectional retrospective cohort study using linked VA and Medicare data from fiscal year 2009. We identified a cohort of community-dwelling Veterans age 65 or older who were diagnosed with type 2 diabetes. Dichotomous indicators were generated to represent any VA or Medicare utilization for the following diabetes-specific healthcare services: outpatient visits, laboratory tests, test strips, medications, inpatient admissions, and emergency department visits. We determined the number and percent of Veterans who were dual users for each service. Using latent class analysis (LCA), we established distinct classes of Veterans based upon their probability of utilizing VA or Medicare for their diabetes care.

Results:
There were 405,286 Veterans in the final cohort. Overall, 202,305 (49.9%) Veterans used both VA and Medicare benefits for at least one diabetes-specific service (dual use). LCA identified 7 distinct classes of VA and Medicare utilization. Veterans in classes 1 (51.0%), 2 (9.3%), and 3 (4.1%) exhibited high probabilities of VA use and low probabilities of Medicare use. Veterans in classes 4 (18.8%), 5 (10.6%), and 6 (3.8%) exhibited high probability of utilizing both VA and Medicare services, although they differed in the specific services used. For example, Veterans in class 4 received their test strips exclusively through Medicare while class 5 received them primarily through VA. Veterans in class 7 (2.4%) were defined by their Medicare reliance.

Implications:
We identified 7 overarching latent classes that account for the diverse ways in which dually enrolled veterans receive diabetes care. These differing patterns of dual use make it challenging to study this population as a single entity, as has typically been done. Each of these typologies may represent differing risks of care fragmentation that may impact cost, safety, and clinical outcomes.

Impacts:
There are 7 overarching typologies that account for the diverse ways in which dually enrolled veterans receive diabetes care. Veterans within each typology may be at differential risk of care fragmentation that may impact cost, safety, and clinical outcomes.