Predicting Objective Performance Using Perceived Cognitive Workload Data in Healthcare Professionals: A Machine...

Area of Science:

• Human Factors and Ergonomics
• Health Informatics
• Human-Computer Interaction

Background:

• Cognitive Workload (CWL) is crucial for predicting healthcare professionals' (HCPs) objective performance.
• The NASA Task Load Index (NASA-TLX) is a common tool for measuring CWL.

Purpose of the Study:

• To compare the predictive accuracy of classical (six-dimension) and novel (four or five-dimension) NASA-TLX models for HCP objective performance.
• To develop and evaluate data-driven computational models using supervised machine learning.

Main Methods:

• Utilized a dataset from prior human factors research studies.
• Applied a wide range of supervised machine learning classification techniques.
• Developed computational models to predict objective performance based on CWL measures.

Main Results:

• Classical models, incorporating all six NASA-TLX dimensions, demonstrated superior accuracy in predicting HCP objective performance compared to novel models.
• Supervised machine learning techniques successfully developed predictive models for objective performance.

Conclusions:

• The classical six-dimension NASA-TLX model is a more effective predictor of HCP objective performance than reduced-dimension models.
• Findings have significant implications for health informatics, human factors, and human-computer interaction in healthcare settings.
• Results are preliminary due to a small dataset, limiting generalizability; future research should include additional CWL measures.