Scoring Patient Fall Reports Using Quality Rubric and Machine Learning.

Area of Science:

• Healthcare quality improvement
• Patient safety research
• Machine learning in healthcare

Background:

• Patient falls are a significant safety concern in healthcare, leading to adverse outcomes.
• Effective analysis of patient fall reports is essential for identifying root causes and preventing recurrence.
• Current patient fall reports often lack the necessary quality and detail for thorough analysis.

Purpose of the Study:

• To develop and evaluate a method for assessing the quality of patient fall reports.
• To identify the most effective machine-learning model for scoring fall reports using the Agency for Healthcare and Quality (AHRQ) rubric.
• To enhance the learning from patient fall reports to improve overall patient safety and reduce fall incidents.

Main Methods:

• Utilized the Agency for Healthcare and Quality (AHRQ) rubric for evaluating patient fall report quality.
• Compared the performance of three distinct machine-learning models in scoring fall reports based on the AHRQ rubric.
• Assessed the effectiveness of each model in capturing report quality and detail.

Main Results:

• Identified a specific machine-learning model as the most effective for scoring patient fall reports according to the AHRQ rubric.
• Demonstrated the potential of machine learning to objectively assess and improve the quality of patient safety event reporting.
• Quantified the performance differences between the evaluated machine-learning models.

Conclusions:

• The selected machine-learning model offers a reliable method for scoring patient fall reports, thereby improving data quality.
• Implementing this scoring method in healthcare facilities can lead to better identification of fall-related safety issues.
• Enhanced report quality will facilitate more effective root cause analysis, ultimately contributing to the prevention of patient falls.