Best practice(in Rapidminer) on avoid confounding variable?

After search our community, I find few post on confounding variable in regression analysis. Confounding variable is often met on observational data. In this attachment, Eggs Meat Fish are Confounding variables, while Age_at_Collection is independent variable(x) and IAA_Level(IAA_Level_Bin) are dependent variable(y).Could you kindly suggest best practice on avoid those confounding variables?