"Proof" of cause and effect in epidemiologic studies: criteria for judgment.

A review of epidemiologic criteria for judging cause and effect emphasizes their empirical basis and points to their limitations for establishing "scientific proof." The criteria are applied to a controversy concerning a putative detrimental effect of spermicides on spontaneous abortion and congenital malformations. The criteria indicate a lack of firm evidence for the contention that spermicides have an adverse effect on pregnancy outcome.