BACKGROUND: Cryptococcal antigen (CrAg) screening at the point of care could improve cryptococcal meningitis prevention where laboratory resources are limited. We evaluated the accuracy of Immunomycologics (IMMY, Norman, OK) CrAg lateral flow assay (LFA) using different techniques at point of care. SETTING: Two tertiary-level hospitals in Johannesburg and a community health clinic in Soweto, South Africa. METHODS: A case-control diagnostic validation study and a prospective clinic-based implementation study using the IMMY CrAg LFA on finger-prick blood. Accuracy, using direct application of LFA to sample, or pipette to transfer sample to diluent, and reading after 10 and 20 minutes, was compared with laboratory-based plasma testing. RESULTS: The validation study tested 64 CrAg-positive and 152 CrAg-negative patients with no symptoms or signs of meningitis, identified by routine laboratory screening, recruited by convenience sampling. Consecutively diagnosed HIV-infected adults (n = 654) were included in the implementation study. Sensitivity was 82% and 20% when the LFA was read 10 minutes after direct application to finger-prick blood in the validation and implementation studies, respectively. Using a pipette to transfer blood and reading after 20 minutes improved sensitivity to 100%, while retaining 100% specificity, in both studies. CONCLUSIONS: Although the IMMY CrAg LFA performs well when applied directly to finger-prick blood for diagnosing cryptococcal meningitis, this technique may not provide adequate volume to detect low concentrations of CrAg when screening asymptomatic patients. Using a pipette to transfer larger volumes of blood to diluent before CrAg LFA testing and reading results after 20 minutes is a more reliable point-of-care method.