Reactive case-detection of malaria in Pailin Province, Western Cambodia: lessons from a year-long evaluation in a pre-elimination setting.

BACKGROUND: As momentum towards malaria elimination grows, strategies are being developed for scale-up in elimination settings. One prominent strategy, reactive case detection (RACD), involves screening and treating individuals living in close proximity to passively detected, or "index" cases. This study aims to use RACD to quantify Plasmodium parasitaemia in households of index cases, and identify risk factors for infection; these data could inform reactive screening approaches and identify target risk groups. METHODS: This study was conducted in the Western Cambodian province of Pailin between May 2013 and March 2014 among 440 households. Index participants/index cases (n = 270) and surrounding households (n = 110) were screened for Plasmodium infection with rapid diagnostic tests (RDT), microscopy and real-time polymerase chain reaction (PCR). Participants were interviewed to identify risk factors. A comparison group of 60 randomly-selected households was also screened, to compare infection levels of RACD and non-RACD households. In order to identify potential risk factors that would inform screening approaches and identify risk groups, multivariate logistic regression models were applied. RESULTS: Nine infections were identified in households of index cases (RACD approach) through RDT screening of 1898 individuals (seven Plasmodium vivax, two Plasmodium falciparum); seven were afebrile. Seventeen infections were identified through PCR screening of 1596 individuals (15 P. vivax, and 22 % P. falciparum/P. vivax mixed infections). In the control group, 25 P. falciparum infections were identified through PCR screening of 237 individuals, and no P. vivax was found. Plasmodium falciparum infection was associated with fever (p = 0.013), being a member of a control household (p ≤ 0.001), having a history of malaria infection (p = 0.041), and sleeping without a mosquito net (p = 0.011). Significant predictors of P. vivax infection, as diagnosed by PCR, were fever (p = 0.058, borderline significant) and history of malaria infection (p ≤ 0.001). CONCLUSION: This study found that RACD identified very few secondary infections when targeting index and neighbouring households for screening. The results suggest RACD is not appropriate, where exposure to malaria occurs away from the community, and there is a high level of treatment-seeking from the private sector. Piloting RACD in a range of transmission settings would help to identify the ideal environment for feasible and effective reactive screening methods.