Cost-benefit analysis of targeted ultrasonography for prenatal detection of spina bifida in patients with an elevated concentration of second-trimester maternal serum alpha-fetoprotein

OBJECTIVE: The objectives of this study were to examine (1) the diagnostic accuracy requirements (from the cost-benefit point of view) of targeted ult rasonography versus genetic amniocentesis for prenatal detection of spina b ifida in women with an elevated level of maternal serum alpha-fetoprotein, (2) the ultrasonographic accuracy of previously published studies from the cost-benefit point of view, and (3) the possible economic impact for the Un ited States of offering targeted ultrasonography instead of routine amnioce ntesis to this group of patients. STUDY DESIGN: Our cost-benefit formula was based on the hypothesis that the cost of universal genetic amniocentesis in patients with an elevated conce ntration of maternal serum alpha-fetoprotein in the second trimester should be at least equal to the cost of universal targeted ultrasonography, with amniocentesis used only for those with abnormalities on a sonogram. The mai n components of the formula included the diagnostic accuracy of targeted ul trasonography (sensitivity and specificity for detecting spina bifida), the cost of the amniocentesis package, the cost of targeted ultrasonography, a nd the lifetime cost of spina bifida not detected by targeted ultrasonograp hy. After appropriate manipulation of the formula, a graph was constructed to represent the balance between the sensitivity and false-positive rate of targeted ultrasonography and was used to examine the accuracy of previousl y published ultrasonographic studies from the cost-benefit point of view. S ensitivity analyses included a range of prevalences of spina bifida in wome n with elevated maternal serum alpha-fetoprotein from 1:50 to 1:200 and fal se-positive rates of targeted ultrasonography from 1% to 10%. RESULTS: Assuming overall prevalences of spina bifida of 1:50, 1:100, or 1: 200 among women with elevated maternal serum alpha-fetoprotein, we found ta rgeted ultrasonography to be beneficial only if the overall sensitivities f or detecting fetal spina bifida were >88%, >76%, and >51%, respectively. Al l 17 studies published after the mid-1980s, which used the "cranial signs" for detecting spina bifida, had accuracies compatible with economic benefit s (sensitivities, 92% to 100%; false-positive rates, 0% to 3%). CONCLUSION: The benefit of second-trimester targeted ultrasonography for fe tal spina bifida depends on diagnostic accuracy tie, sensitivity and false- positive rate). Currently achieved ultrasonographic accuracies are compatib le with net benefits. Targeted ultrasonography in patients with an elevated level of second-trimester maternal serum alpha-fetoprotein in the United S tates has the potential for annual savings of approximately $36 million to $49 million and for avoiding 268 fetal deaths. (Am J Obstet Gynecol 1999;18 0:1227-33.).