Evaluating the VA Make-or-Buy Decision in Emergency Care

Background: In response to concerns about access to and quality of care at VA facilities, the VA has begun to
redirect resources toward financing care for Veterans outside of the VA. However, the quality and cost of care
for Veterans that will result from this “make-or-buy” decision remain a scientific unknown, with significant policy
implications for how the VA can optimize Veteran health. Quality and cost of care are particularly influenced by
emergency services, with 13% of care and more than $800 million in yearly costs directed outside the VA.
Objective: Our objective is to identify how quality and cost of care for Veterans are affected by its provision
inside or outside the VA. Studying the quality and costs of VA vs. non-VA care faces the fundamental concern
that patients who receive care at VA facilities may have different levels of underlying health than do those who
receive care outside the VA. Without understanding the consequences of VA vs. non-VA emergency care on
quality and cost, VA policymakers will be unable to direct Veterans to the best care location, nor will they
understand mechanisms behind quality and cost differences between VA vs. non-VA care. Finally, a lack of
knowledge about the effects of VA vs. non-VA care, and about how Veterans access care, prevents
policymakers from predicting quality and cost outcomes for Veterans from policy interventions redirecting
Veterans to non-VA care.
Methods: In Aim 1, we will adopt a quasi-experimental approach, based on instrumental variables (IVs), to
study the effect of the care source on health, utilization, and spending. In preliminary results, we find that VA
EDs reduce elderly Veteran mortality by 50%, or five percentage points. In Aim 2, we will assess the factors
altering the effect of VA vs. non-VA care, evaluating mechanisms behind the effect. For example, we will
investigate whether the VA performs better for Veterans with certain conditions or demographics, such as a
mental health diagnosis or low socioeconomic status. We will assess whether the capacity of local VA options
plays a key role in determining outcomes, and whether coordination of care mediates improved VA outcomes.
In Aim 3, we will evaluate how Veteran use VA and non-VA ED alternatives, and we will use these results and
those in Aims 1 and 2 to simulate quality and cost outcomes under policies for expanding VA capacity,
contracting with non-VA options, and redirecting Veterans to care depending on their characteristics and on
local options.
Expected Outcome: Based on our preliminary results, we expect there to be important differences in health
outcomes and spending between VA and non-VA emergency care. We therefore expect that policymakers can
save thousands of Veteran lives and can improve health outcomes by making decisions based on this
research. The effect of VA care likely differs across Veterans and local conditions. Accordingly, we expect
there to be large additional gains in health and spending outcomes that result from tailoring policies related to
this important make-or-buy decision. Finally, by understanding how outcomes differ across settings and
Veteran types, we expect to contribute to a general understanding about which components of health care
delivery are most critical to achieving excellence. In partnership with the Offices of Policy and Planning,
Community Care, and Emergency Medicine, we anticipate that our findings will be disseminated widely and will
be applied directly to VA decisions and guidelines.

Grant Number: 5I01HX002631-07
NIH Institute/Center: VA
Principal Investigator: David Chan