Statistical analysis
Continuous variables were expressed as mean ± standard deviation (SD) or median (P25, P75), and categorical variables were expressed as percentages. We compared the distribution of participants with different demographic characteristics by chronotypes (morning chronotype, neutral chronotype, and evening chronotype) using t-tests and analysis of variance (ANOVA) for continuous variables or Chi-square tests and Fisher’s exact test for categorical variables. Binary logistic regression models were used to examine the relationship between chronotype and depressive symptoms during infertility treatment. Hypothesis testing was performed to ensure that the assumptions of the regression model were met. Since the missing values are few, we handle it by deleting the rows.
We used the evening chronotype as the reference group in the binary logistic regression models to assess the odds ratios (ORs) and 95% confidence intervals (CIs) for chronotype on the depressive status and adjusted for possible confounding variables in the final model, including age, sex, annual income, education, passive smoking, physical activity, living children, infertility treatment time, cause of infertility, frequency of insomnia, nocturnal wake frequency, daytime napping, social jetlag, and nighttime sleep duration. Trend tests were performed using the median of the different categories of chronotypes scores as a continuous variable. Restrictive cubic splines with five knots were conducted to characterize the dose-response curve between sleep and depressive symptoms. We also tested the interaction effect and estimated the relationship between chronotype and odds of depressive symptoms by stratification of the covariates above. A sensitivity analysis was applied to test the robustness of the results. As well, we utilized the Actor- Partner Interdependence Mode (APIM) approach to explore the partner effect of own chronotype on the spouse’s depressive symptoms, as recommended by Kenny et al 36. In our study, the actor effect refers to the influence of a person’s chronotype on his/her own depressive symptoms, and the partner effect refers to the influence of a person’s chronotype on his/her spouse’s depressive symptoms. The APIM framework for a husband-wife dyad is also described in Figure S3 .
All the above statistical analyses were carried out using R software version 4.1.0 (University of Auckland, Auckland, New Zealand), SPSS version 23.0 (SPSS, Chicago, IL, USA), and Mplus software (version 8.3). In current study, all statistical tests were double-tailed and the level of significance was p -value <0.05 which was considered statistically significant.