BACKGROUND/RATIONALE:
Most Veterans enrolled in VA also have healthcare access through other health programs. Dual use of VA and non-VA care is potentially problematic for reasons such as duplication of services, fragmentation or discontinuity of care and challenges in communication across health systems that limit providers' ability to make fully informed treatment decisions. Despite the high rate of dual use among Veterans, current understanding of its underlying causes and economic implications is incomplete. The extent to which policy changes, such as the Affordable Care Act (ACA), affects Veterans' dual use of VA and non-VA care has not been examined. Analysis of Veterans' health choices following Massachusetts' healthcare reform in 2006 may provide valuable insights for the impact ACA on use of VA care because many components of the two policy changes are similar.

OBJECTIVE(S):
This research investigated economic and policy factors impacting Veterans' interaction with VA and other health systems. Aims 1 and 2 examined the potential impact of ACA on Veterans' enrollment in VA, utilization of VA outpatient services, costs and dual use. Aim 3 identified patient-reported predictors of dual VA and non-VA health service use.

METHODS:
All aims employed retrospective, observational analyses using administrative and survey data. Aim 1 examined changes in VA, private insurance and Medicaid enrollment among a representative sample of New England Veterans from the Current Population Survey during 2003-2013. Aims 2 and 3 used VA and Medicare administrative data from fiscal year (FY) 2004 through FY2013 and examine VA outpatient utilization, costs and dual use (measured as reliance on VA care) among all VA enrolled veterans. Analyses for Aims 1 and 2 employed a quasi-experimental study design comparing changes in outcomes before and after healthcare reform between respective groups of Massachusetts and other New England Veterans. In Aim 3, administrative data were linked to survey data from the VA Survey of Healthcare Experiences of Patients to identify patient-reported experience variables that were predictive of increased VA reliance. Traditional regression models were applied to examine which patient-reported variables were associated with greater VA reliance. In addition, we measured the proportion of variation in reliance explained by patient-reported experiences and to determine the extent to which adding these variables to a prediction model could improve the ability to accurately predict future reliance. Several machine learning algorithms including regularized regression, decision trees and neural networks were applied to understand whether alternative models could help improve predictive accuracy.

FINDINGS/RESULTS:
Not yet available.

IMPACT:
A more comprehensive understanding of the factors predictive of dual use, including patient-reported perceptions of VA care and policy changes such as ACA, is significant for fiscal planning and the efficient allocation of health resources in VA. VA currently incorporates Veterans' expected reliance on VA care in projection models used for developing annual budget requests to Congress. Better predictions of dual use will ensure VA receives funding levels that balance Veterans' health needs and fiscal responsibility. Determining the role of primary care perceptions in patients' healthcare choices will also help VA improve healthcare delivery through the Patient Aligned Care Team initiative.

---

External Links for this Project

NIH Reporter

Grant Number: IK2HX001389-01
Link: https://reporter.nih.gov/project-details/8595842

---

PUBLICATIONS:

None at this time.

DRA: Health Systems Science
DRE: Treatment - Preclinical, Prevention, Technology Development and Assessment
Keywords: none
MeSH Terms: none