Transforming Alzheimer’s Digital Caregiving through Large Language Models

Introduction/Objective: Alzheimer's Disease and Related Dementias (AD/ADRD) present significant caregiving challenges, with increasing burdens on informal caregivers. This study examines the potential of AI-driven Large Language Models (LLMs) in developing digital caregiving strategies for AD/ADRD. The objectives include analyzing existing caregiving education materials (CEMs) and mobile application descriptions (MADs) and aligning key caregiving tasks with digital functions across different stages of disease progression. Methods: We analyzed 38 CEMs from the National Library of Medicine’s MedlinePlus, along with associated hyperlinked web resources, and 57 MADs focused on AD digital caregiving. Using ChatGPT 3.5, essential caregiving tasks were extracted and matched with digital functionalities suitable for each stage of AD progression, while also highlighting digital literacy requirements for caregivers. Results: The analysis categorizes AD caregiving into 4 stages-Pre-Clinical, Mild, Moderate, and Severe-identifying key tasks, such as behavior monitoring, daily assistance, direct supervision, and ensuring a safe environment. These tasks were supported by digital aids, including memory-enhancing apps, Global Positioning System (GPS) tracking, voice-controlled devices, and advanced GPS tracking for comprehensive care. Additionally, 6 essential digital literacy skills for AD/ADRD caregiving were identified: basic digital skills, communication, information management, safety and privacy, healthcare knowledge, and caregiver coordination, highlighting the need for tailored training. Conclusion: The findings advocate for an LLM-driven strategy in designing digital caregiving interventions, particularly emphasizing a novel paradigm in AD/ADRD support, offering adaptive assistance that evolves with caregivers' needs, thereby enhancing their shared decision-making and patient care capabilities.