Abstract: Objective: To explore the heterogeneous trajectories and the predicting factors of sleep among the elderly in nursing homes. Methods: A total of 353 elders from 27 nursing homes were surveyed using the Athens Insomnia Scale, Generalized Anxiety Disorder Scale, 9-item Patient Health Questionnaire, Verbal Rating Scale for pain, Lubben Social Network Scale and Activity Participation Questionnaire at baseline and after 6 months and 12 months during the follow-up period. The sleep trajectories of the elderly were explored by Latent Class Growth Analysis. Results: A total of 277 participants completed at least two surveys and the data was included in the analysis. Four sleep trajectories were identified: good sleep (78.0%), insomnia and deteriorated (9.0%), insomnia but improved (7.9%) and maintaining severe insomnia (5.1%); the results of Logistic regression analysis showed that age, depression, pain, activity participation and social network could predict sleep trajectories. Conclusion: The heterogeneous trajectories existed in the sleep of elders in nursing homes. Those who were older, more depressed, with lower activity participation and with more serious pain, were more likely to worsening their insomnia. Participants with less social network tended to suffer maintaining more severe insomnia. It is necessary to identify different sleep trajectories among the elders in nursing homes, and effective targeted interventions should be taken to improve their sleep status and the quality of life.

Key words: nursing home; the elderly; sleep; trajectory