Developing an Artificial Intelligence-Driven Nudge Intervention to Improve Medication Adherence: A Human-Centred...

Area of Science:

• Digital Health
• Artificial Intelligence in Healthcare
• Human-Computer Interaction

Background:

• Medication adherence remains a significant challenge in healthcare, impacting treatment efficacy and patient outcomes.
• Existing adherence tracking methods often have limitations, necessitating innovative solutions.
• Digital interventions, particularly those leveraging artificial intelligence (AI), show promise for improving patient engagement and adherence.

Purpose of the Study:

• To co-develop a human-centered, AI-driven digital nudge intervention for medication adherence with key stakeholders.
• To identify user needs and preferences for designing an effective and acceptable intervention.
• To report findings from the initial phase of a multi-phase project focused on user needs assessment and solution ideation.

Main Methods:

• A human-centered design approach was employed, involving co-development with patients, providers, and technologists.
• Qualitative interviews were conducted with healthcare providers (n=10) and patients (n=10).
• Healthcare providers also participated in a survey to rate example nudge interventions.

Main Results:

• Stakeholders expressed optimism about the potential of the AI-driven intervention to address medication adherence deficits.
• Flexibility in delivery mode, intensity, and user-specific stratification were identified as critical success factors.
• Reminder nudges and direct contact with healthcare workers were highly favored, while incentive-based nudges were poorly received by patients.
• Minimizing user burden through integration with existing software and simplified data entry was recommended.

Conclusions:

• An AI-driven nudge tool is perceived as an acceptable and appropriate solution for improving medication adherence.
• Incorporating user needs through a participatory approach is crucial for enhancing intervention utility.
• The developed intervention must be flexible to accommodate diverse user requirements and preferences for successful implementation.