Fig. 7

Population-level proportion of P. vivax relapses in a the PNG placebo arm; b the PNG primaquine arm; and c Thailand. Solid lines show the proportion of new P. vivax infections due to relapses, and dashed lines show the proportion of total P. vivax infections due to relapses. The coloured lines depict the result for each genotype and the black lines show the average across all genotypes. The difference between the solid and dashed lines is due to relapses that are undetectable because they occur when blood-stage parasites of the same genotype are already circulating. If an individual has not been recently exposed, then their next P. vivax infection is more likely to be from a mosquito bite. For example, in low transmission Thailand, if an individual has had no detectable blood-stage infection for > 9 months, the probability that a new infection is due to a relapse is low, as liver-stage infection of this duration without relapsing is unlikely. However, once a new infection does occur, it is likely to be followed by multiple relapses hence the high proportion of total infections due to relapses