Background: This study identifies latent profiles of sleep in the first trimester of pregnancy using a person-centered method, and investigate the predictive role of demographics, perinatal features, physical activity, depression, and social capital across profiles.
Methods: A total number of 1,066 pregnant women in Shenzhen were invited to participate in this study. Latent profile analysis (LPA) was used to identify sleep profiles. Regression Mixture Modeling (RMM) was used to explore the predictive role of demographic variables, clinical features, physical activity, depression, and social capital among sleep profiles.
Results: Three profiles were identified:(1) good sleep quality (n = 732, 68.7%), (2) poor sleep efficiency (n = 87, 8.2%), (3) daily disturbances (n = 247, 23.2%). Age, education, occupation, gravidity, childbirth, pregnancy BMI, depression, and social capital were the predictive factors among sleep profiles. Compared with good sleep quality group, pregnant woman in poor sleep efficiency group were more likely to be younger, have education of high school or technical secondary school and undergraduate or above, and higher level of depression, but less likely to have twice pregnancy and one childbirth. Those in daily disturbances group were more likely to be older, obesity and have lower lever of social capital, but less likely to be worker and public servant.
Conclusion: This study revealed three sleep profiles using a person-centered method and underlined the predictive role of depression and social capital across profiles. Our results may provide information for tailored interventions that can promote sleep quality of pregnant women and prevent a worsened sleep quality unprecedented situation.