JMIR Publications

ABSTRACT

Background:

Generative artificial intelligence (AI) has the potential to impact healthcare and medicine by transforming clinician workflows and improving patient outcomes. Patient-centered clinical decision support (PC CDS) are digital tools that use patient-specific information and patient-centered outcomes research. Generative AI-supported PC CDS can be leveraged to improve healthcare decision making. However, a robust understanding of patient and caregiver perspectives on generative AI-supported PC CDS is imperative.

Objective:

We aimed to generate a prioritized list of patient- and caregiver-informed considerations for the design, implementation, and use of generative AI in PC CDS.

Methods:

We conducted a total of six small group discussions across two phases, January/February 2024 and November 2024, with a total of 16 participants comprising patient and caregiver advocates. Using the first phase discussions, we generated an initial list of seven key considerations for the implementation and use of generative AI-supported PC CDS. During the second phase, we generated a refined list of considerations using a prioritization ranking activity and member checking.

Results:

Participants believed generative AI-supported PC CDS tools have the potential to enhance efficiency, support clinicians, and improve healthcare decision making but recognized they could introduce challenges. Trust and willingness to use these tools are shaped by individuals' healthcare experiences and familiarity with technology. Key concerns included transparency, data security and accuracy, and potential to exacerbate or introduce bias and mistrust in healthcare. Participants emphasized the importance of customizability to different preferences and seamless integration into patient-clinician interactions to avoid disruptions. The tools' success depends on clinicians’ skills and use. Our final list of seven patient- and caregiver-informed considerations include the development of regulatory standards and design principles for generative AI-supported PC CDS tools; codesign of generative AI-supported PC CDS tools with end-users, including patients and caregivers; considerations for mistrust in the healthcare system, especially among vulnerable populations; monitoring and evaluation to ensure these tools are accurate; education and training to understand and use AI; use of generative AI-supported tools that complement clinicians’ work and uphold the patient-clinician relationship; and AI that holistically uses patient data and tailors outputs.

Conclusions:

Our study addresses gaps in existing research by offering novel insights into patient and caregiver preferences for generative AI-supported PC CDS tools. It highlights nuanced perspectives on the tools' impact on patient-clinician dynamics and underscores the importance of participatory design approaches. Future efforts on generative AI-supported PC CDS tools should focus on actionable steps to address identified key considerations. This work contributes to the growing evidence base on AI in healthcare by describing patient and caregiver views on the use of generative AI in PC CDS. The list of patient- and caregiver-informed considerations for the implementation and use of generative AI-supported PC CDS tools can support trust and patient-centeredness in emerging technologies.