Perceptions Toward Using Artificial Intelligence and Technology for Asthma Attack Risk Prediction: Qualitative...

Area of Science:

• Digital Health
• Indigenous Health
• Artificial Intelligence

Background:

• Asthma affects over 500,000 New Zealanders, with Māori experiencing disproportionately worse outcomes.
• Digital technologies like AI and ML are used for asthma prediction but risk bias against minority groups.
• Existing health inequities may be exacerbated by AI models that underrepresent Māori.

Purpose of the Study:

• To explore Māori perspectives on AI and ML for asthma self-management.
• To identify key considerations for developing equitable asthma attack risk prediction models.
• To ensure Māori representation in ML models without worsening health disparities.

Main Methods:

• Conducted semistructured interviews with 20 Māori participants with asthma.
• Employed thematic and sentiment analysis to identify key themes and participant sentiments.
• Utilized a Bidirectional Encoder Representations from Transformers model for sentiment analysis.

Main Results:

• Identified four themes: concerns about AI, interest in technology, desired system characteristics, and experiences with asthma care.
• Participants expressed unfamiliarity and trust concerns regarding AI, influenced by historical colonization.
• Most participants showed interest in technology for asthma management, providing insights into preferences and healthcare challenges.

Conclusions:

• Māori communities require more information and trust-building regarding AI and technology in healthcare.
• Participant expectations for computer-based health applications were identified.
• Outcomes will inform future AI and technology development for asthma care, particularly for Indigenous populations in New Zealand.