Supporting the Capture of Social Needs Through Natural Language Processing.

Area of Science:

• Clinical Informatics
• Public Health
• Health Services Research

Background:

• Social factors significantly influence patient morbidity and mortality.
• Family physicians commonly document social needs in clinical notes.
• Unstructured data on social factors in electronic health records (EHRs) impedes clinical action.

Purpose of the Study:

• To propose and evaluate natural language processing (NLP) for identifying social needs within EHRs.
• To enable structured and reproducible documentation of social determinants of health (SDOH).

Main Methods:

• Utilizing NLP techniques to parse unstructured clinical notes.
• Developing algorithms to extract predefined social needs categories.
• Assessing the consistency and reproducibility of extracted information.

Main Results:

• NLP successfully identified social needs from unstructured clinical text.
• The proposed method facilitates structured data capture without increasing physician workload.
• Potential for improved identification and addressing of patient social needs.

Conclusions:

• NLP offers a viable solution to overcome unstructured data challenges in documenting social needs.
• This approach can enhance clinical decision-making by providing structured SDOH data.
• Automated social needs extraction can streamline documentation and improve patient care.