In a study to determine the predictors of Gestational diabetes mellitus (GDM) in pregnancy, the diagnosis of GDM was used as the outcome variable ( Dependent variable). This was a dichotomous (dependent) variable with positive cases, i.e diagnosed cases of GDM ( targeted group) coded as 1, while women who were negatives were coded as 0 (reference group). There were 21 cases of GDM and 249 women were screened negative.
Is it statistically feasible to use this dichotomous variable (dependent variable) in binary logistic regression despite the low frequency of the targeted group (21) to the reference group (249)?