4096 — Relative Utilization Shifts in VA Services for Depression, Post-Medicaid Expansion in Arizona and New York
Lead/Presenter: Daniel Liaou,
COIN - Houston
All Authors: Liaou DW (Center for Innovations in Quality, Effectiveness & Safety, Michael E. DeBakey VA Medical Center), O'Mahen PN (Center for Innovations in Quality, Effectiveness & Safety, Michael E. DeBakey VA Medical Center), Petersen LA (Center for Innovations in Quality, Effectiveness & Safety, Michael E. DeBakey VA Medical Center)
Examine relative utilization changes in VA vs. Medicaid-reimbursed services for depression after New York (NY) and Arizona (AZ) expanded Medicaid in 2001.
Our difference-in-difference analysis paired Veterans in NY and AZ with demographically-comparable states Pennsylvania (PA) and New Mexico/Nevada (NM/NV), respectively. Inpatient and outpatient visits documenting depression as the primary diagnosis in VA and Medicaid administrative data from 1999 to 2006 were extracted using ICD-9-CM codes. The proportion of visits occurring at VA facilities was modeled using fractional logit with covariates of gender, age, race, and Deyo (comorbidity) scores. Relative utilization changes from pre- to post-2001 were calculated, comparing Veterans in Medicaid expansion vs. non-expansion states and those with priority-5 (indicating financial hardship) vs. non-priority 5 status.
From pre- to post-expansion periods, the change in proportion of inpatient visits at the VA was -0.088 [99% CI: -0.154, -0.022]* (an 8.8 percentage point decrease) among non-priority-5 enrollees in the expansion states compared to non-expansion states. Among priority-5 enrollees, this was -0.137 [-0.239, -0.036]*. The change in proportion of outpatient visits at the VA from pre- to post-2001 was -0.006 [-0.016, 0.004] in expansion states for non-priority-5 enrollees, compared to non-expansion states. Among priority-5 enrollees, this was -0.025 [-0.042, -0.009]*.
We observed trends away from utilizing VA services for depression among Veterans after states expanded Medicaid. This effect was particularly distinct for Veterans with priority-5 status, as well as for hospitalizations, regardless of priority-5 status. Non-priority-5 Veterans appeared to trend towards decreased utilization of outpatient services, but this did not reach statistical significance.
Our analysis indicates Medicaid expansion significantly shifts utilization of mental health services for depression away from the VA. As the VA is an integrated healthcare system with tailored and coordinated mental health services for Veterans, these findings engender economic and healthcare questions concerning the consequences of shifting utilization towards Medicaid-reimbursed care. As data becomes available in states that expanded Medicaid under the Affordable Care Act, this study provides a preliminary understanding and framework to further exploring these questions.