Does a significant reduction in malaria risk make lopinavir/ritonavir-based ART cost-effective for children with HIV in co-endemic, low-resource settings?