From Cues to Nudge: A Knowledge-Based Framework for Surveillance of Healthcare-Associated Infections

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Ontology Engineering

Background:

• Healthcare-associated infections (HAIs) pose significant risks to patient safety and quality of care.
• Existing methods for classifying and managing HAIs can be inconsistent, leading to challenges in standardization and diagnosis coding.
• The HAIKU (Hospital Acquired Infections - Knowledge in Use) framework aims to standardize HAI classification.

Purpose of the Study:

• To propose an integrated semantic web framework for recommending appropriate patient care levels.
• To enhance the classification and detection of hospital-associated infections (HAIs) using formal ontologies and semantic rules.
• To improve the risk assessment and early detection of surgical site infections (SSIs), particularly in high-risk surgeries like coronary artery bypass graft (CABG).

Main Methods:

• Development of a semantic web framework incorporating formal ontologies (HAI ontology - HAIO), web services, a reasoner, and a rule engine.
• Utilizing the HAIKU framework for classifying HAIs, focusing on key types: SSIs, catheter-associated urinary tract infections (CAUTIs), hospital-acquired pneumonia, and bloodstream infections.
• Employing statistical inferencing and heuristic-based rule axioms for improved SSI case detection and identifying SSIs using semantic e-triggers.

Main Results:

• The proposed framework integrates formal ontologies and semantic rules to recommend patient care.
• The HAI ontology (HAIO) provides a structured hierarchy for thousands of HAI-related codes, enhancing consistency.
• Semantic e-triggers were demonstrated to identify SSI occurrences, aiding in risk assessment for specific surgical procedures.

Conclusions:

• The integrated semantic web framework offers a standardized approach to patient care recommendations and HAI management.
• The HAIKU framework and HAIO ontology facilitate consistent clinical practice and diagnosis coding, improving patient safety.
• The use of semantic e-triggers shows promise for proactive risk assessment and early detection of SSIs, ultimately enhancing surgical outcomes.