Identifying Transportation Needs in Ophthalmology Clinic Notes Using Natural Language Processing: Retrospective,...

Area of Science:

• Ophthalmology
• Health Informatics
• Natural Language Processing

Background:

• Transportation insecurity is a significant barrier to accessing eye care, negatively impacting patient visual outcomes.
• Electronic health records often lack structured data on transportation challenges, hindering patient identification and support.
• Free-text clinical notes offer a potential avenue for capturing transportation-related barriers more effectively.

Purpose of the Study:

• To develop and validate a natural language processing (NLP) algorithm for identifying transportation insecurity in free-text ophthalmology clinic notes.
• To assess the feasibility of using NLP to detect transportation barriers in electronic health records.
• To improve the identification of patients facing challenges in accessing eye care due to transportation.

Main Methods:

• A retrospective, cross-sectional study analyzing 1,801,572 ophthalmology clinic notes from adult patients (2016-2023).
• Development of a rule-based NLP algorithm to detect transportation insecurity within deidentified free-text clinical documentation.
• Validation of the NLP algorithm against a gold-standard expert review, using precision, recall, and F1-scores to evaluate performance.

Main Results:

• The NLP algorithm successfully identified 726 patients (0.6%) with transportation insecurity, demonstrating high performance (precision 0.860, recall 0.960, F1-score 0.778).
• Older patients (≥80 years) were significantly more likely to experience transportation insecurity compared to younger adults.
• Asian patients were less likely to have transportation insecurity identified compared to White patients; no significant differences were found by sex or race (Black vs. White).

Conclusions:

• Natural language processing (NLP) is a viable tool for accurately identifying transportation insecurity from unstructured ophthalmology clinical notes.
• This NLP approach can aid in proactively identifying patients with transportation barriers, facilitating timely referrals to necessary resources.
• Improving the capture of transportation insecurity can lead to better patient support and potentially enhance visual outcomes in eye care settings.