If this is set up as a standard epidemiological case referent (case control) study) with cases and controls (non cases) appropriately sampled, then standard binomial regression would be suitable-- include as independent variables all your continuous variables plus the two categorically defined confounder variables .

then the exp (beta coefficient) gives the multiplicative increase in odds ratio (relative risk) for unit change in each continuous independent variable (risk factor) ,adjusted for confounding.

if you are just comparing some findings in a group of patients with "normal " controls. Then I would recommend separate t tests for each variable between cases and controls in each confounder category (eg city male, city female,

rural male, rural female) if you have enough data.

@james Leigh. Thank you sir. I shall use the binomial logistics regression then.

Hello Sahithya,

Assuming the outcome of interest is a categorical variable, then logistic regression, as suggested by James Leigh would work (whether binomial or multinomial depends on the number of categories into which participants may be classified).

However, if the outcome is a measured variable (e.g., white blood cell count), then multiple linear regression (or, if you prefer, hierarchical MLR) would work.

Good luck with your study.