Amal Trivedi is a Professor of Health Services, Policy and Practice at Brown University and a Research Investigator at the Providence VA Medical Center. His research evaluates the effects of payment, financing and other health system interventions on the quality, equity and outcomes of health care.

More than 9 million veterans receive health care through the Department of Veteran Affairs (more commonly known as the VA) in the nation’s largest integrated health care delivery system. Reports of excessive waiting times for VA outpatient care and outdated facilities have led some policymakers to propose shifting veterans’ care from the VA to non-VA systems. Even though empirical evidence has found the VA performs performs “similarly or better than the non-VA system on most of the nationally recognized measures of inpatient and outpatient care quality,” this debate continues.

A recent study published in The BMJ by David Chan, Kaveh Danesh, Sydney Costantini, David Card, Lowell Taylor and David Studdert offers new insights on a key question: Are veterans who get emergency care in the VA less likely to die than those who get this care in non-VA hospitals? 

It can be difficult to compare patient outcomes for VA and non-VA hospitals because many veterans are more medically high risk and more likely to have low incomes and disabilities than non-veteran populations. To get around this, the authors used data for veterans with VA and Medicare coverage who could therefore get care inside or outside the VA. The authors also limited the study to veterans living within 20 miles of a VA who went to the hospital in an ambulance because of an acute emergency. 

The researchers had three key findings:

• Veterans getting emergency care at a VA hospital had a 20% lower risk of death within 30 days than those at a non-VA hospital. 

• This lower mortality rate was much lower for Black and Hispanic veterans and veterans who had previously received care in the same VA. 

• Veterans getting care in the VA were sicker — they had more chronic conditions and a higher risk of death. 

Despite its careful design, the study may not fully account for all differences between patients who use VA and non-VA care. For example, ambulances may play a role in selecting hospitals based on the emergency, as could patient preferences. Also, the findings may not extend to non-emergency care or younger populations. 

Despite these limitations, the study’s findings have crucial policy implications, especially as the VA prepares for a potentially massive overhaul of its network of hospitals and clinics. They suggest that it may be harmful to shift veterans’ care from VA to non-VA hospitals. At least for emergency visits, going to a non-VA hospital seems to have lethal consequences for veterans.