Hybrid collaborative filtering methods for recommending search terms to clinicians.

Area of Science:

• Health Informatics
• Artificial Intelligence in Medicine
• Clinical Decision Support

Background:

• Electronic health records (EHR) present challenges for efficient information retrieval.
• Clinicians face difficulties in finding relevant patient data for diagnosis.
• Repetitive searching for similar patient cases is time-consuming.

Purpose of the Study:

• To develop an effective recommender system for suggesting accurate search terms to clinicians.
• To address the challenge of efficient information retrieval in EHR systems.
• To reduce the burden of repetitive searches for clinicians.

Main Methods:

• Developed a hybrid collaborative filtering model using patient clinical encounter data and search history.
• Recommendations are generated based on co-occurring ICD codes and relevance to recent searches.
• Two model variations were explored: HCFMH (most recent ICD codes) and cpHCFMH (all ICD codes).

Main Results:

• The proposed hybrid model demonstrates superior performance in top-N search term recommendation.
• Experimental results show outperformance against state-of-the-art baseline methods.
• The model effectively leverages patient data and search patterns for accurate recommendations.

Conclusions:

• The developed hybrid collaborative filtering model significantly enhances search term recommendation for clinicians.
• This approach improves diagnostic efficiency by providing relevant information retrieval.
• The recommender system offers a valuable tool for navigating complex EHR data.