Electronic Surveillance For Catheter-Associated Urinary Tract Infection Using Natural Language Processing

Area of Science:

• Healthcare-associated infections
• Clinical informatics
• Natural language processing

Background:

• Catheter-associated urinary tract infection (CAUTI) is a significant healthcare burden.
• Accurate CAUTI measurement is challenging and resource-intensive.
• Existing electronic surveillance methods often use simplified CAUTI definitions.

Purpose of the Study:

• To develop and validate an electronic surveillance tool for CAUTI using a complete definition.
• To incorporate subjective elements from clinical notes via natural language processing (NLP).
• To improve the objectivity and efficiency of CAUTI detection.

Main Methods:

• Applied a comprehensive CAUTI definition, including subjective symptoms identified by NLP.
• Developed a rule-based algorithm analyzing clinical documentation for CAUTI indicators.
• Trained the algorithm on 1421 catheterized patients and validated prospectively on 1567.
• Compared algorithm performance against gold standard chart review.

Main Results:

• The developed tool demonstrated high performance metrics.
• Sensitivity: 97.1%
• Specificity: 94.5%
• Positive Predictive Value (PPV): 66.7%
• Negative Predictive Value (NPV): 99.6%

Conclusions:

• The NLP-based electronic surveillance tool accurately identifies CAUTI using a complete definition.
• This approach offers a more objective and efficient method for CAUTI surveillance.
• Further refinement of computable definitions is suggested for future improvements.