Table 5.

Association between time-of-day energy intake patterns and insulin resistance (OR, 95% CI) ¹.

Models	“Noon Dominant Pattern” (n = 2479)	“Evening Dominant Pattern” (n = 4887)	“Morning Dominant Pattern” (n = 1360)
Model 1	1	1.16(1.02–1.32) *	1.01(0.85–1.20)
Model 2	1	1.15(1.01–1.31) *	1.05(0.88–1.25)
Model 3	1	1.14(1.00–1.30) *	1.05(0.88–1.26)
Model 4	1	1.21(1.05–1.40) *	1.08(0.89–1.31)

¹ A Multilevel Mixed-effects Logistic Regression Model was constructed to estimate the association between latent time-of-day energy intake patterns and insulin resistance (binary, HOMA-IR ≥ 2.0 as the cutoff), taking household as the second level and individual as the first level. Model 1 adjusted for no covariates. Model 2 adjusted for age (continuous), gender (categorical), education level (categorical), geographic region (categorical), per capita household income (categorical), urbanicity index (continuous), physical activity (categorical), smoking (categorical), alcohol drinking (categorical), and sleep duration (categorical). Model 3 additionally adjusted for total energy intake, and CDGI (2019)-A score. Model 4 additionally adjusted for BMI, SBP, DBP, TC, TG, HDL_C. * p < 0.05.