What Can We Learn about Fall Risk Factors from EHR Nursing Notes? A Text Mining Study.

Area of Science:

• Health Informatics
• Clinical Nursing Research
• Natural Language Processing (NLP)

Background:

• Hospital falls are a significant clinical concern, costing billions annually and impacting patient safety.
• Existing fall risk assessments lack empirical validation, and current reduction strategies show limited sustained success.
• Registered nurses' (RNs') narrative notes are an under-examined source of actionable patient fall data.

Purpose of the Study:

• To explore the presence and utility of meaningful fall risk and prevention information within RNs' electronic narrative notes.
• To investigate whether NLP can extract actionable data from clinical notes for fall risk assessment.

Main Methods:

• Utilized a natural language processing (NLP) design on deidentified electronic health record (EHR) data.
• Extracted and analyzed RN narrative notes from the Medical Information Mart for Intensive Care (MIMIC-III) database (2001-2012).
• Applied text mining procedures to identify documented fall risk factors and prevention interventions.

Main Results:

• Over one million RN notes were analyzed, identifying thousands of notes with explicit fall risk/prevention documentation.
• A significant portion of notes mentioned intrinsic (patient-related) and extrinsic (environmental/organizational) factors associated with fall risk.
• Intervention information potentially impacting patient falls was found in a notable percentage of notes.

Conclusions:

• RNs' narrative notes represent a rich, underutilized source of actionable data for patient fall risk and prevention.
• NLP can effectively identify clinical, environmental, and organizational factors influencing fall risk from unstructured text.
• Further research is warranted to validate the predictive value of these identified factors in clinical practice.