June 2022

In This Issue: HSR&D Research on Community Care
» Table of Contents

Opioid Misuse: Coordination of Fragmented Care

Feature Article

Takeaway:This study will leverage several non-VA data sources, along with VA data, to study opioid use and misuse among Veterans who use VA healthcare only – and those who use both VA and non-VA (community) healthcare in order to improve care coordination and reduce opioid misuse.

The US opioid epidemic has put a significant burden on Veterans and the VA healthcare system. Veterans often suffer from chronic pain due to war injuries, toxic exposures, and deployment-related psychiatric comorbidities and, thereby, are vulnerable to opioid use/misuse. Both opioid misuse and unrelieved pain have been linked to a higher risk of suicide among Veterans. While the majority of Veterans enrolled in VA healthcare have their own insurance (~80%), VA researchers and operations depend mostly on VA data to assess opioid use and evaluate policy. Meanwhile, Veterans who are dual users of VA and community care are at higher risks of opioid use and misuse. Also, while opioid overdose rates have been increasing in VA enrollees, preliminary data (from this investigator) show a decline in opioid prescriptions, but less of a decline in the diagnosis of opioid use disorder. The absence of community data is mentioned as a significant deficiency in the study.

This ongoing study (May 2021 – April 2024) is examining factors associated with prescription opioid misuse, specifically the guideline-discordant use of opioids, in three Veteran groups: 1) Veterans using VA healthcare only; 2) Veterans using both VA and non-VA care – paid for by VA (dual users); and 3) non-VA paid dual users. Investigators are conducting an interview/focus group study of VA and non-VA community health providers perspectives on:

• Barriers and facilitators in providing guideline-concordant care to the dual users, and
• Coordination strategies to reduce opioid misuse in the dual user groups.

Investigators will analyze complex data using deep learning and natural language processing methods in addition to state-of-the-art statistical methods. The main goal is to evaluate and improve care coordination – and reduce opioid misuse in Veterans who are dual users of VA and community care.

Methods

Data will include the VA and MedStar Health (largest healthcare system in the Mid-Atlantic region) electronic health record (EHR), the Chesapeake Regional Information System for Patients (CRISP), and Medicare databases. Preliminary studies have been conducted and preliminary data gathered to demonstrate the feasibility of the proposed deep learning and natural language processing methods, as well as access to VA and non-VA EHR data.

Findings

Using data from DC and Baltimore VA with 217,745 unique Veterans, investigators have found:

• The percentage of dual spell out first time and VA) users has increased from almost 0% before 2015 to about 24% in 2019.
• The VCP (Veteran Community Partnerships) users are about three years younger (58 vs 61) – and have more female Veterans (24% vs 13%), more African Americans (52% vs 42%) and Latinos (4% vs 2%). The VCP group also had a much higher prevalence of PTSD (31% vs 19%).
• The VCP group is associated with higher rates of opioid initiation, opioid use, and opioid misuse. After adjusting for age, gender, race, ethnicity, and selected comorbidities, the odds ratios of VCP vs non-VCP users are 23 (1.18-1.28), 1.17 (1.15-1.19), and 1.07 (1.04-1.09) for opioid initiation, opioid use, and opioid misuse, respectively. 

Anticipated Impact

This study will leverage several non-VA data sources along with VA data to study opioid use and misuse among Veterans who use VA healthcare only – and those who use both VA and non-VA (community) healthcare, with the goal of improving care coordination and reducing opioid misuse.

Principal Investigators: Joel Kupersmith, MD, Professor of Medicine and Director, Georgetown University Veterans Initiatives, and former Director of VA’s Office of Research and Development (2005-2013); and Qing Zeng, PhD, part of VA Washington DC health care and is Director of the Biomedical Informatics Center at the GW School of Medicine & Health Sciences in Washington, DC.

Publications

None to report at this time.

View study abstract