Primary Care Physician Perspectives on Artificial Intelligence for Dementia Screening: A Qualitative Study

Area of Science:

• Medical Informatics
• Clinical Ethics
• Geriatric Medicine

Background:

• Automated cognitive impairment (CI) screening shows promise for clinical use but raises significant ethical concerns.
• Successful implementation hinges on understanding the perspectives of clinicians, the intended users of this technology.

Purpose of the Study:

• To identify the ethical challenges faced by physicians regarding the implementation of automated CI screening.
• To explore physician attitudes towards automated CI screening using machine learning models and electronic medical records or audio data.

Main Methods:

• A qualitative study involving 22 physicians from internal medicine, geriatrics, and family medicine in New York City.
• Four focus groups were conducted to discuss manual and automated CI screening, using hypothetical scenarios.
• Grounded theory was applied to analyze transcribed focus group recordings.

Main Results:

• Physicians currently screen for CI primarily during Medicare annual wellness visits, often avoiding it due to limited treatment options and poor access to specialist services.
• Perceived benefits include improved clinical efficiency, patient engagement in cognitive healthcare, and expanded access to care.
• Identified ethical challenges include concerns about accuracy, bias, patient communication impacting informed consent, and risks such as privacy, stigmatization, and insurability.

Conclusions:

• Physicians highlighted critical ethical challenges for automated CI screening.
• Addressing these challenges is essential for ensuring patient safety, autonomy, and privacy.
• Health systems must develop supportive infrastructure to maximize the benefits of automated CI screening for patients with dementia.