Leveraging Natural Language Processing to Identify Veterans Who Inject Drugs to Assess Preexposure Prophylaxis and Sexually Transmitted Infection Testing Services at the Veterans Health Administration - Summary - MDSpire
Advertisement
Leveraging Natural Language Processing to Identify Veterans Who Inject Drugs to Assess Preexposure Prophylaxis and Sexually Transmitted Infection Testing Services at the Veterans Health Administration
To evaluate an NLP dashboard designed to identify Veterans with evidence of injection drug use and assess the delivery of preventive care services, including STI testing and PrEP.
Key Findings:
507 potential PWID identified, with 78 (15%) confirmed through chart review, indicating a need for improved identification methods.
49% of confirmed PWID injected opiates, 41% cocaine, and 37% methamphetamines, highlighting the diversity of substance use.
HIV prevalence was 6%, hepatitis C antibody positivity 45%, and hepatitis B exposure 13%, underscoring the health risks faced by this population.
Only 29% of PWID saw infectious disease specialists despite high engagement with mental health services, indicating a gap in care access.
Interpretation:
The NLP dashboard effectively identified PWID within the VA system, revealing significant gaps in preventive care delivery despite high engagement with mental health services, suggesting a need for targeted interventions.
Limitations:
The study was limited to 6 VA facilities and may not represent all veterans, potentially affecting the generalizability of the findings.
Confirmation of PWID relied on chart reviews, which may introduce bias and affect the accuracy of the identified population.
Conclusion:
The findings suggest opportunities to enhance collaboration among mental health, social work, and infectious disease services, including specific strategies such as integrated care pathways and shared training initiatives to improve care for PWID.
