Sera were collected from 426 volunteers in Uganda at high and low risk for acquisition of hepatitis C virus (HCV). All samples were tested by the Ortho HCV second generation ELISA (S1) and by the INNOTEST HCV Ab second generation enzyme immunoassay, (S2), (Innogenetics, Antwerpen, Belgium). Sera that were repeatedly reactive by either screening assay were further tested by each of two different HCV supplemental/confirmatory assays: a second generation recombinant immunoblot assay (RIBA, Ortho Diagnostics), (C1), and a line immunoassay (INNO-LIA HCV-Ab, Innogenetics), (C2). In these populations there were 16 true positives, 351 true negatives, and 59 indeterminate results. Fifty-nine point four percent (253/426) of the samples were repeatedly reactive by the S1 test, while only 2.6% (11/426) were repeatedly reactive by S2. Test S1 produced a high false positive rate, a low positive predictive value, a specificity of 49.3%, and had a sensitivity of 100%. In contrast, the S2 screening assay had much higher specificity (98.8%), but lacked in sensitivity (31.3%). This poor sensitivity of S2 was based almost exclusively on the fact that the C2 supplemental test classified 9 samples as confirmed positive when the homologous screening assay classified these samples as negative; several of these were not confirmed when using a new generation INNO-LIA. Both the screening tests S1 and S2, and the supplemental assays C1 and C2 exhibited a significant degree of discordance, and neither of the screening tests alone would be adequate for use in these populations.