The rationale of calculating predictive values to interpret the amniotic fluid alpha-fetoprotein (AFP) test has been examined and applied to amniotic fluid AFP testing from one Canadian center. Such predictive value ccalculations may be misleading if they fail to make use of the actual magnitude of the test result or the results of other investigations. The alculation of predictive values has thus been extended to take into account magnitude of the test results, the clinical history, and the results of other investigations. The interpretation of an abnormal amniotic fluid AFP test that is followed by a normal result of a careful ultrasound scan of the fetal back is that there is a 54.5% chance that the fetus has spina bifida if there is a previous history of spina bifida. There is a 12.5% chance if there is a negative family history. These calculations lead to informed genetic counseling and rational decision making with regard to the continuation of the pregnancy.