In Japan, it is common for individual elderly patients requiring long-term care to use multiple home care services in order to continue living at home. However, the details of the services provided and patients’ conditions at the time of care are not shared among the service providers.
 To address this issue, our research team aims to develop a website that collectively manages care information, predicts declines in activities of daily living (ADL) based on artificial intelligence (AI) analysis, and if necessary, automatically sends reports to service providers and the care manager in charge of the patient. 
 The target population of this study are elderly people with care needs living at home, their families, care managers, and home care service providers. 
 This research is being conducted with the cooperation of elderly individuals requiring care and their families, while always listening to the opinions of the parties involved. In developing the web system, we collaborated with a researcher specializing in care management and another researcher specializing in website development.
 This report describes a questionnaire survey conducted to examine the assessment items to be used in a “care management support system for prevention of severe disease and continuation of living at home for elderly people requiring long-term care”, which our research team is working to develop. Firstly, the assessment items presented in “Appropriate care management techniques: Basic care and disease-specific care” published by the Japanese Ministry of Health, Labour and Welfare were discussed among the author, co-authors, and care managers who were not the subjects of this study, and 73 assessment items were extracted. For each of the 73 items, a questionnaire was developed asking “How important do you think it is for the elderly requiring care to continue living at home?” Each question was scored on a 5-point scale. Of the 30,631 home care support centers in Japan, 1,000 were randomly selected based on a population-proportional distribution by prefecture and government-designated city. Questionnaires were mailed to a total of 3,000 care managers, three to each center, and 642 valid responses were obtained (valid response rate: 21.4%). 
 Assessment items that the care managers selected as highly important (4.7 points or higher) were “presence or absence of falls”, “status of cognitive impairment if the patient has dementia”, “amount of food intake”, “changes in daily condition”, and “mental and physical statuses of a caregiver (family member)”.
 For the international audience, the development of a “website that predicts declines in ADL based on AI analysis, and if necessary, sends reports to service providers and the care manager in charge of the patient” may provide useful information, as it could be modified to suit the situations in each country and be used effectively in many countries. 
 In the next steps, we plan to conduct interviews with elderly people requiring long-term care, their families, and care managers and determine the data entry items before creating a website and implementing it on a trial basis.