Studying Veteran food insecurity longitudinally using electronic health record data and natural language processing

Area of Science:

• Public Health
• Health Informatics
• Social Determinants of Health

Background:

• Food insecurity is a significant social risk factor affecting patient health.
• The Department of Veterans Affairs (VA) seeks to address food insecurity through interventions.
• Evaluating intervention effectiveness is hindered by a lack of longitudinal food insecurity data for Veterans.

Purpose of the Study:

• To develop and validate a natural language processing (NLP) system for identifying food insecurity from electronic health records (EHRs).
• To analyze longitudinal food insecurity trajectories in a large cohort of Veterans.
• To provide a scalable method for assessing the impact of food insecurity interventions and VA policies.

Main Methods:

• Development and validation of an NLP system to extract food insecurity status from clinical notes.
• Application of the NLP system to a large cohort of Veterans' EHR data.
• Analysis of longitudinal data to understand the timing and persistence of food insecurity.

Main Results:

• A validated NLP system capable of identifying food insecurity from clinical documentation.
• Insights into the temporal patterns and duration of food insecurity among Veterans.
• Demonstration of a method to leverage EHR data for studying social determinants of health.

Conclusions:

• NLP analysis of EHRs offers a viable solution for tracking Veteran food insecurity.
• The developed system can support the evaluation of VA's food insecurity interventions and policies.
• This approach enables a deeper understanding of the long-term impact of food insecurity on Veteran health.