Descriptive data were compared using Pearson’s chi-square test and one-way analysis of variance (ANOVA). Categorical data such as gender, age range, ethnicity, low SES, smoking status, hypertension occurrence, BMI, diabetes, and CVD were compared with shift-type using Pearson’s chi-square test for independence. A one-way ANOVA was used to compare shift schedule types (categorical) and shift duration (continuous) with the outcome of interest.
Binary and multinomial logistic regression models were used to examine associations between shift type and shift duration (in years) with CMD risk factors (smoking status, hypertension, and BMI) and CMD (diabetes and CVD) with the relevant odds ratio (OR) and 95%CI. All analyses were adjusted for gender, age range, ethnicity, and low SES as potential confounders [19 (link),20 (link)]. We created the models in the following fashion: in the crude model, no covariates were added in the model; Model 1 was adjusted for sociodemographic factors of age, gender, ethnicity, and SES; Model 2 was adjusted for model 1 and work arrangement.