Validity of Lot Quality Assurance Sampling to optimize falciparum malaria surveys in low-transmission areas.

To control the reappearance of malaria in the Madagascan highlands, indoor house-spraying of DDT was conducted from 1993 until 1998. Before the end of the insecticide-spraying programme, a surveillance system was set up to allow rapid identification of new malaria epidemics. When the number of suspected clinical malaria cases notified to the surveillance system exceeds a predetermined threshold, a parasitological survey is carried out in the community to confirm whether or not transmission of falciparum malaria is increasing. Owing to the low specificity of the surveillance system, this confirmation stage is essential to guide the activities of the control programme. For this purpose, Lot Quality Assurance Sampling (LQAS), which usually requires smaller sample sizes, seemed to be a valuable alternative to conventional survey methods. In parallel to a conventional study of Plasmodium falciparum prevalence carried out in 1998, we investigated the ability of LQAS to rapidly classify zones according to a predetermined prevalence level. Two prevalence thresholds (5% and 15%) were tested using various sampling plans. A plan (36, 2), meaning that at least 2 individuals found to be positive among a random sample of 36, enabled us to classify a community correctly with a sensitivity of 100% and a specificity of 94%. LQAS is an effective tool for rapid assessment of falciparum malaria prevalence when monitoring malaria transmission.