Text this: Confounding and effect modification: possible explanation for variation in the results on the association between oral and systemic diseases