Using machine learning to modify and enhance the daily living questionnaire.

Area of Science:

• Cognitive assessment
• Machine learning applications
• Psychometrics

Background:

• The Daily Living Questionnaire (DLQ) is a functional cognitive measure used in medical and rehabilitation settings.
• Its length presents challenges for efficient public screening, leading to potential inaccuracies due to subject fatigue.

Purpose of the Study:

• To utilize Machine Learning (ML) to shorten the DLQ without sacrificing accuracy or fidelity.
• To develop a more efficient and user-friendly version of the DLQ for broader application.

Main Methods:

• An ML-based Computerized Adaptive Testing (ML-CAT) algorithm was applied to DLQ data from studies in the USA and Israel.
• The ML-CAT algorithm created an adaptive testing instrument with a shortened form tailored to individual scores.

Main Results:

• The ML-CAT approach reduced the number of required tests by 25% for individual DLQ scores and over 50% when predicting all seven scores concurrently.
• Accuracy was maintained at 95% (5% error) across all subject scores.
• The study identified specific DLQ items most predictive of overall scores.

Conclusions:

• The ML-CAT model offers a method to modify, refine, and abridge the DLQ.
• This abbreviated DLQ can facilitate wider community screening and enhance clinical and research utility.