Diagnostic tests that require human interpretation will always be susceptible to error. This may lead to delays in initiating effective therapy for potentially fatal conditions. The Malaria Control Programme in South Africa has historically relied on the microscopic examination of Giemsa-stained thick blood smears to confirm the clinical diagnosis of malaria. The level of agreement of malaria laboratory diagnosis performed by the four laboratories serving the Mpumalanga Province Malaria Control Programme was investigated. The disagreement between the laboratories argues in favour of a uniform training and quality assurance programme or the introduction of alternative diagnostic modalities. It should also serve as a caution against diagnostic complacency in other public health programmes.