Desai, Priyanka J.; Dobes, Angela; Shah, Avantika S.; Ancker, Jessica S.; Abdulhay, Lindsay; Das, Sagarika; Peterson, Caroline; Dullabh, Prashila. (2026).听.听Journal of Medical Internet Research, 28, e75851.听
础蝉听generative artificial intelligence (AI)聽becomes increasingly integrated into healthcare, understanding how patients and caregivers view these technologies is essential. This study explored perspectives on聽patient-centered clinical decision support (PC CDS)鈥攄igital tools that use a patient’s health information and research evidence to help guide healthcare decisions. Through a series of small group discussions with 16 patient and caregiver advocates, the researchers identified key priorities for designing and implementing AI-supported PC CDS tools. Participants recognized that generative AI could improve efficiency, support clinicians, and enhance healthcare decision-making, but they also raised concerns about transparency, data privacy, accuracy, bias, and trust. They emphasized that AI tools should be designed in partnership with patients and caregivers, complement rather than replace clinicians, and strengthen the patient-clinician relationship. Participants also highlighted the need for ongoing monitoring to ensure AI systems remain accurate, along with education and training to help both patients and clinicians use these tools effectively. The study resulted in seven patient- and caregiver-informed recommendations that can help guide the development and implementation of AI-supported clinical decision support tools, with a focus on building trust, promoting human oversight, and tailoring care to individual patient needs.
Figure 1.听Importance rating (on a 5-point Likert scale) for the initial list of 7 considerations (N=9). The numbers in the bars represent the number of participants who selected the Likert scale response for consideration. AI: artificial intelligence; PC CDS: patient-centered clinical decision support.