Optimizing malarial epidemiological studies in areas of low transmission.

Malaria risk factor studies have traditionally used microscopy readings of blood slides as the measure of malaria infection in humans, although alternatives are available. There is the need for an assessment of how the use of these alternative diagnostic approaches will influence the efficiency and significance of epidemiological studies. In an area of Sri Lanka with known risk factors for malaria, two cross-sectional surveys were done at the start and at the peak of transmission season. Microscopy was compared with enzyme-linked immunosorbent assays (ELISA) and polymerase chain reaction (PCR). The major risk factor in this area was the location of houses relative to confirmed vector breeding sites. At the peak of the transmission season, the results pointed in the same direction, irrespective of the diagnostic method used. However, the importance of distance from the breeding site was not statistically significant when microscopy was used, which can be explained by the lower prevalence of microscopy positivity in comparison to the prevalence of ELISA- and PCR-positivity. This study suggests that in low-transmission areas, such as Sri Lanka, smaller sample sizes can be used for epidemiological research studies using PCR instead of microscopy to estimate parasite prevalence. This efficiency gain has to be weighed against the higher cost and complexity of the PCR. PCR cannot replace microscopy as the standard diagnostic procedure at the field level. ELISA is not directly comparable with microscopy and PCR but it can also be a useful tool in malaria epidemiological studies. This study indicates that cross-sectional surveys are only efficient if they take place during peak transmission season. Cross sectional surveys currently implemented by the Sri Lankan government in response to local malaria outbreaks can form the basis for valid epidemiological studies and be used for the generation of malaria risk maps if samples were also analyzed using PCR.