Extracting Housing and Food Insecurity Information From Clinical Notes Using cTAKES

Area of Science:

• Health Informatics
• Natural Language Processing
• Social Determinants of Health

Background:

• Electronic Health Records (EHRs) are increasingly used to capture patient data.
• Health-related social needs (HRSNs) significantly impact health outcomes, especially in older adults.
• Natural Language Processing (NLP) offers potential for extracting unstructured data from EHRs.

Purpose of the Study:

• To evaluate the effectiveness and challenges of using NLP, specifically the Clinical Text Analysis and Knowledge Extraction System (cTAKES), to identify HRSNs in older adults.
• To assess the performance of cTAKES in recognizing food and housing insecurity within EHRs.
• To identify barriers to NLP-driven HRSN ascertainment in clinical notes.

Main Methods:

• Utilized cTAKES with Concept Unique Identifiers and Systematized Nomenclature for Medicine codes to extract HRSN information from de-identified EHRs.
• Validated cTAKES performance through manual chart review for food insecurity (included in screening) and housing insecurity (not included).
• Analyzed 1,385,259 clinical notes from 119,127 patients aged 55+ in a large California healthcare system (2013-2022).

Main Results:

• cTAKES demonstrated a moderate positive predictive value (77.5%) for housing insecurity, but often failed to recognize relevant concepts.
• Performance for food insecurity was poor (18.5% positive predictive value) due to incorrect flagging of structured screening tool data.
• Significant discrepancies were observed between the algorithm's recognized concepts and actual patient challenges.

Conclusions:

• NLP systems like cTAKES face substantial challenges in accurately identifying food and housing insecurity in older adults.
• Barriers include unstandardized terminology within EHRs and inadequate integration of HRSN screening tools.
• Improved NLP models and EHR integration are crucial for effectively leveraging clinical notes to address social needs.