Abstract: Based on a questionnaire survey in Beijing, the desirable caring patterns of urban elderly were investigated. With a decision tree analysis approach, the respondents’choices among four different caring patterns (living independently, family care, community care, and institutional care) in two scenarios were revealed, one in the healthy stage and one that a person was in need of long term care. Then the rationality of the preferred caring patterns was examined. First of all, the study manifested the lifestyle change of Chinese elderly, which was characterized by a tremendous number of no-child families. There was a huge gap between the needs of people in different health stages. In particular, about half of the respondents intended to go to nursing homes if they were in need of care, while only 5.7% intended to do so when they were healthy. However, the severe shortage of caring facilities was a critical issue, especially those for the disabled and semi-disabled people, and it would be unrealistic to provide enough nursing beds in the future. Considering the capacity of service supply, it was proposed that the appropriate ratios for the (semi-)disabled elderly to choose institutional care and community care in the year 2020 could be 35% and 30%, respectively. Furthermore, people aged under 70 should be the main target of demand management, most of whom had demonstrated a strong preference for institutional care in the future.

Key words: aging, Beijing, caring pattern, caring service, community care, decision tree, institutional care