A predictive model for assistive technology adoption for people with dementia

Area of Science:

• Gerontology
• Human-Computer Interaction
• Health Informatics

Background:

• Assistive technology can improve independence for people with dementia, supporting home-based care.
• Individuals with dementia often resist change, necessitating careful selection of suitable technologies.
• Predictive models are needed to assess an individual's likelihood of adopting specific technologies.

Purpose of the Study:

• To develop and evaluate predictive models for assistive technology adoption in people with dementia.
• To identify key features influencing technology adoption among this population.
• To provide healthcare professionals with interpretable tools for technology selection.

Main Methods:

• Development of predictive models using data mining classification algorithms.
• Consideration of personal characteristics: ability, living arrangements, and preferences.
• Evaluation of models based on prediction performance, robustness, bias, and usability.

Main Results:

• The k-Nearest-Neighbour algorithm, using seven features, demonstrated optimal classification performance.
• Achieved a prediction accuracy of 0.84 ± 0.0242 for assistive technology adoption.
• Models were assessed for ease-of-use and clear decision-making processes for healthcare professionals.

Conclusions:

• A k-Nearest-Neighbour based predictive model effectively identifies assistive technology adoption potential in people with dementia.
• This approach supports informed technology selection, enhancing independence and home-based care.
• The developed models offer practical, interpretable tools for healthcare providers.