Leveraging user's performance in reporting patient safety events by utilizing text prediction in narrative data entry

Area of Science:

• Health Informatics
• Human-Computer Interaction
• Clinical Data Management

Background:

• Narrative data entry is crucial for electronic health records and patient safety event reporting.
• Efficient and high-quality clinical data entry is essential for accurate diagnosis and treatment.
• Text prediction technology offers potential to improve data entry performance in healthcare.

Purpose of the Study:

• To evaluate the impact of text prediction functions on the efficiency and data quality of patient safety event reporting.
• To assess how text prediction influences the rate of text generation and the completeness of reported information.

Main Methods:

• A two-group randomized controlled trial involving 52 nurses reporting patient fall cases.
• Participants were assigned to either a treatment group (with text prediction) or a control group (without text prediction).
• Statistical analyses included t-tests, chi-square tests, and linear regression to compare free-text data entry outcomes.

Main Results:

• The treatment group demonstrated a 70.5% increase in text generation rate and a 34.1% rise in reporting comprehensiveness.
• A 14.5% reduction in non-adherence within comment fields was observed in the treatment group.
• The treatment group showed a progressive increase in text generation over time, unlike the control group.

Conclusions:

• Text prediction functions effectively assist healthcare professionals in generating comprehensive free-text reports for patient safety events.
• This strategy shows promise for enhancing clinical data entry across various healthcare settings requiring free-text documentation.