Use of ChatGPT by physicians to build rehabilitation plans for the elderly: A mini-review of case studies

Abstract

This mini-review explores the potential of using ChatGPT, an artificial intelligence language model, to build personalized rehabilitation plans for elderly patients. Creating such plans improve the function of morbid and frail elderly, and can be time-consuming, requiring a multidisciplinary team of health-care professionals. ChatGPT can generate human-like responses to text inputs, making it a valuable tool for health-care professionals in creating personalized rehabilitation plans. The review outlines a case study in which trial and error questioning was done to develop a set of optimal parameters that can be input into ChatGPT to develop personalized rehabilitation plans for patients. Six case scenarios involving different organ systems were assessed by expert geriatricians for quality of advice. ChatGPT use offered several benefits for developing personalized plans, such as its easy and free accessibility, personalized chatbot, ability to integrate complex multiple morbidities, and reduced need for extra personnel. However, its limitations include limited accuracy, no reference of information, bioethical considerations, lack of information storing capabilities, and patient mistrust in machine learning software. Overall, our review suggests that ChatGPT has the potential to be an excellent tool for developing personalized rehabilitation plans for elderly patients. However, it is important to consider the limitations and ensure that health-care professionals review and approve any plans generated by ChatGPT before implementing them. It is crucial to note that ChatGPT should be used to support clinical decision-making rather than replace health-care professionals' expertise and knowledge.