Improving physician practice efficiency by learning lab test ordering pattern.

Area of Science:

• Medical Informatics
• Health Services Research
• Clinical Decision Support

Background:

• Physician practice faces time pressures, particularly in China, necessitating efficiency improvements.
• Ordering laboratory tests is a frequent action within electronic medical records (EMR) systems.
• Current EMR systems may not fully optimize the lab test ordering process for physician efficiency.

Purpose of the Study:

• To reduce physician time spent on ordering lab tests during patient encounters.
• To develop a system that facilitates efficient lab test ordering within EMRs.
• To investigate the impact of physician-specific patterns on lab test ordering.

Main Methods:

• Developed a weight-based multi-label classification framework.
• Trained models using historical electronic medical record data.
• Incorporated physician-specific lab test ordering patterns into the models.

Main Results:

• Physician-specific models demonstrated superior performance compared to baseline approaches.
• The framework successfully learned to suggest relevant lab tests based on historical EMR data.
• The proposed method showed potential for time savings in clinical practice.

Conclusions:

• A physician-specific, weight-based multi-label classification approach can enhance EMR efficiency.
• Tailoring lab test ordering suggestions to individual physicians improves performance.
• This methodology offers a promising direction for mitigating physician time pressure.