Automated Surgical Term Clustering: A Text Mining Approach for Unstructured Textual Surgery Descriptions.

Area of Science:

• Health care informatics
• Medical text mining
• Predictive analytics in surgery

Background:

• Rising healthcare costs necessitate quality improvement in care delivery.
• Surgical services significantly impact hospital costs, revenue, and patient care quality.
• Effective operational planning and inventory management are crucial for surgical unit efficiency.

Purpose of the Study:

• To develop a novel preprocessing framework for surgical text data.
• To address challenges in utilizing unstructured text from electronic health records for text mining.
• To improve the accuracy of predictive models in surgical settings by cleaning text data.

Main Methods:

• Proposed a novel preprocessing framework for surgical text data.
• Focused on detecting misspellings and ad hoc abbreviations in unstructured text.
• Reduced the dimensionality of the raw feature set from principal procedure and additional notes.

Main Results:

• The framework effectively identifies and corrects misspellings and abbreviations in surgical notes.
• The transformed text feature set enhances subsequent prediction tasks.
• Validated the approach using datasets from multiple hospital surgical departments.

Conclusions:

• The proposed framework significantly improves the quality of surgical text data for analysis.
• Accurate text preprocessing is essential for leveraging unstructured data in health informatics.
• This method supports better operational planning and quality improvement in surgical services.