JMIR Publications

ABSTRACT

Background:

Artificial intelligence (AI) tools are increasingly being integrated into healthcare, offering potential to reduce clinicians' workload and enhance patient care. Older adults, as key healthcare consumers, exhibit unique preferences, including a strong inclination toward shared decision-making and trust in physician-derived information. This research explores critical factors influencing older adults' perceptions of AI-assisted medication decisions, addressing limitations in existing studies while proposing methodologies to enhance clinical relevance. Methodology:The study critiques a vignette-based experimental survey method for its oversimplification of real-world complexities and proposes improvements. These include in-person surveys, multi-center investigations, and incorporation of cultural and psychosocial variables. Additionally, it emphasizes the need to address algorithmic bias, transparency, and long-term dynamics through follow-up studies.

Results:

The research identifies several limitations in current methodologies, including the inability of online surveys to account for cognitive diversity and the neglect of cultural factors shaping perceptions. It highlights that the "black-box" nature of AI undermines trust, especially among older adults unfamiliar with AI technologies. Furthermore, comparative evaluations of AI and physician recommendations could reveal critical insights into trust dynamics.

Conclusions:

To enhance the clinical relevance of AI research in healthcare, methodologies must incorporate real-world complexities, cultural and psychosocial factors, and robust mechanisms for ensuring transparency and addressing algorithmic biases. Future research should adopt longitudinal designs to explore evolving perceptions and their implications for doctor-patient relationships. Addressing these aspects will improve the robustness of AI integration into medication decision-making for older adults.