2019 HSR&D/QUERI National Conference

1157 — Medicare-eligible Veterans voting with their feet: Events associated with changes in a Veteran's reliance on the VA vs. Medicare for healthcare

Lead/Presenter: Paul Hebert,  COIN - Seattle/Denver
All Authors: Hebert PL (Seattle-Denver COIN), Wong, ES (Seattle-Denver COIN and Primary Care Analytic Team), Gunnink, (Primary Care Analytic Team) Ashok, R (Seattle-Denver COIN and Primary Care Analytic Team) Nelson, KE (Seattle-Denver COIN and Primary Care Analytic Team)

Objectives:
We can learn something about how Veterans value the VA versus community providers by observing how Veterans vote with their feet to use VA over Medicare providers when they turn 65. We addressed two questions: How soon after becoming Medicare eligible does a Veteran decide to become VA- or Medicare-reliant, and what events are associated with a change in VA reliance over time.

Methods:
This was a longitudinal cohort study of Veterans' responses to a 1999 health survey merged with VA and Medicare data from 1998-2016. We defined VA reliance as VA outpatient visits as a fraction of VA plus Medicare visits. We estimated the change in reliance associated with three events: an incident life-threatening diagnosis (cancer, heart failure, renal failure or dementia); a Medicare-paid hospitalization; and a change in residence that increased the distance to the VA. We estimated a logistic regression where the dependent variable was as indicator for whether a visit occurred in the VA vs Medicare. Independent variables included indicators of occurrences of the events in the 48 months surrounding the visit.

Results:
More Veterans chose to rely on the VA (n = 4361) than Medicare (n = 1571). Among the Veterans who chose Medicare, half made this choice in their first year of Medicare eligibility. For a VA-reliant veteran, a Medicare-paid hospital stay was associated with a -7.8 percentage points (pps) (p < 0.001) change in VA reliance in the subsequent 12 months but by 36 months reliance had recovered to near pre-hospitalization levels (-1.5 pps; p = 0.138). Diagnoses of cancer, heart- ore renal failure had no significant association with subsequent reliance; however, a diagnosis of dementia was associated with a decrease in VA reliance (-8.6 pps; p = 0.026). Moving further from the VA had no effect on reliance.

Implications:
By a 2:1 majority, newly Medicare-eligible VA users voted with their feet in favor of maintaining the VA as a provider of comprehensive medical care.

Impacts:
Observing how veterans vote with their feet might inform how they value VA-provided care at the facility and service-line level.