OBJECTIVES Nationwide, human immunodeficiency virus type 1 (HIV-1) seroprevalence surveys using dried neonatal blood specimens are critical to estimating HIV-1 seroprevalence among childbearing women. However, the noninclusion of blood specimens deemed "quantity not sufficient" (QNS) for HIV-1 antibody testing potentially introduces bias. In Wisconsin beginning in 1990, we modified the survey protocol to reduce QNS rates and assess bias introduced by QNS specimens. METHODS The HIV-1 antibody assay was modified to use four 1/8-in blood spots when a single 1/4-in blood spot could not be obtained. Both methods obtain identical blood volumes for testing. RESULTS During a 27-month period, 7396 (4.8%) of 154,683 specimens were deemed QNS using 1/4-in blood spots. Of these, 6590 (89%) were of sufficient quantity to be tested using four 1/8-in blood spots; 6 (0.09%) specimens tested with 1/8-in blood spots were HIV-1 Western blot assay positive compared with 44 (0.03%) of 147,287 1/4-in specimens (odds ratio = 3.0; 95% confidence interval = 1.2, 7.4). CONCLUSIONS Because noninclusion of QNS specimens potentially introduces bias, incorporating the results of HIV-1 antibody testing of QNS specimens using four 1/8-in blood spots can improve the accuracy of HIV-1 seroprevalence estimates in these serologic surveys.