Assisting nurses in care documentation: from automated sentence classification to coherent document structures with...

Area of Science:

• Health Informatics
• Natural Language Processing

Background:

• Nurses spend up to 35% of their time on patient care documentation.
• Current documentation methods involve manually selecting from over 500 subject headings, which is time-consuming.
• A system to streamline nursing note creation is needed to reduce workload.

Purpose of the Study:

• To evaluate a system designed to assist nurses with patient care documentation.
• To enable nurses to dictate or write notes naturally without manual structuring.
• To reduce the overall documentation time and effort for nurses.

Main Methods:

• A neural network-based text classification model is used for sentence-level classification.
• Sentences are grouped into paragraphs based on topic coherence.
• A paragraph merging step is implemented to consolidate related content.

Main Results:

• The system achieved topic coherence in approximately 69% of generated paragraphs.
• Assigned paragraph headings accurately reflected the content in 69% of cases.
• A paragraph merging step reduced the number of paragraphs by 23% without performance loss.

Conclusions:

• The developed system effectively structures free-text nursing narratives.
• This technology has the potential to significantly decrease the time nurses spend on documentation.
• The system offers a promising solution for improving efficiency in hospital care documentation.