OBJECTIVE: My purpose was to evaluate the cost-effectiveness of screening for human immunodeficiency virus during pregnancy as part of a protocol in which zidovudine was used to reduce the risk of vertical transmission. STUDY DESIGN: This mathematic model used decision analysis to calculate the marginal cost-effectiveness of screening for human immunodeficiency virus in pregnancy and treating human immunodeficiency virus - positive women with zidovudine. Cost and probability assumptions were drawn from a literature review. Sensitivity analyses were performed for important costs and probabilities. RESULTS: When baseline cost and probability assumptions were used, the marginal cost-effectiveness of human immunodeficiency virus screening was $436,927 when the prevalence of human immunodeficiency virus in the population was low (0.00075) and $198,510 when the prevalence was average (0.0015). Above a prevalence of human immunodeficiency virus of 0.009, testing is both cheaper and more effective than not testing. Of the cost variables examined, the charge for a negative testing sequence had the greatest impact on cost-effectiveness. CONCLUSION: Human immunodeficiency virus testing in pregnancy is cost-effective in populations in which the prevalence of human immunodeficiency virus exceeds 9 per 1000 population. Depending on how individual lives saved are valued, screening may also be warranted in populations with lower prevalences of infection. (AM J OBSTET GYNECOL 1996;174:716-21.)