Initiating Care for Low Back Pain in the Emergency Department of the Veterans Health Administration, Impact on the Location of Subsequent Visits.

Objectives: Our study objective was to evaluate the location and timing of subsequent Veterans Health Administration care for low back pain (LBP) based on the initial clinic. Veterans can initiate care for LBP through different entry points such as primary care, emergency department (ED), urgent care, or other service line clinics. We examined how different starting points impacted the location and timing of the next LBP encounter. Methods: Veteran Affairs (VA) electronic health records were used to identify veterans with an initial visit for LBP, defined as a visit with an International Classification of Diseases, 10th Revision (ICD-10) LBP-related code after a 365-day period without such a visit, between October 1, 2015, and September 30, 2016. Primary and secondary clinic stop codes were used to identify the initial entry point and stratify the clinics into 3 groups: ED/urgent care, primary care, and all other service lines grouped into "other." The location of the first subsequent visit within 365 days with an ICD-10 code for LBP, if any, was similarly identified. Logistic regression was used to assess the impact of entry point on the location of the first subsequent LBP visit and identify demographic and clinical characteristics of veterans whose first subsequent LBP visit was in the ED/urgent care clinic. Results: We identified 55,612 veterans, 88% of whom were male, 61% of whom identified as White, 54% of whom were married, and 9% of whom had co-occurring neck pain, who presented to VA care settings for an initial visit for LBP. Of those, 44,956 (80.8%) were initially seen in primary care, 4652 (8.4%) in the ED/urgent care, and 6004 (10.8%) in other clinics. First subsequent LBP visits occurred in primary care clinics (39%), other clinics (37%), no subsequent LBP visit within 365 days (21%), or ED/urgent care clinics (3%). The odds of the first subsequent LBP visit occurring in the ED/urgent care clinics were lower for veterans initially seen in the primary care clinic (odds ratio [OR], 0.12; 95% CI, 0.11-0.14) or other clinics (OR, 0.12; 95% CI, 0.09-0.15) compared with veterans initially seen in the ED/urgent care settings. Pain, a history of a serious LBP diagnosis, and prescription of opioids were associated with the first subsequent LBP care in the ED settings. Conclusion: In the VA care setting, 1 in 10 initial LBP visits occur in the ED/urgent care setting. This initial site of care is associated with an 8-fold increase in the odds that the next medical visit for LBP will also occur in the ED/urgent care setting. Future research is needed to explore patient and system factors contributing to the first subsequent LBP care occurring in the ED/urgent care setting after their initial LBP visit.