Rapid diagnostic tests (RDTs) provide point-of-care medical screening without the need for expensive laboratory equipment. RDTs are theoretically straightforward to use, yet their analog colorimetric output leaves room for diagnostic uncertainty and error. Furthermore, RDT results within a community are kept isolated unless they are aggregated by healthcare workers, limiting the potential that RDTs can have in supporting public health efforts. In light of these issues, we present a system called RDTScan for detecting and interpreting lateral flow RDTs with a smartphone. RDTScan provides real-time guidance for clear RDT image capture and automatic interpretation for accurate diagnostic decisions. RDTScan is structured to be quickly configurable to new RDT designs by requiring only a template image and some metadata about how the RDT is supposed to be read, making it easier to extend than a data-driven approach. Through a controlled lab study, we demonstrate that RDTScan's limit-of-detection can match, and even exceed, the performance of expert readers who are interpreting the physical RDTs themselves. We then present two field evaluations of smartphone apps built on the RDTScan system: (1) at-home influenza testing in Australia and (2) malaria testing by community healthcare workers in Kenya. RDTScan achieved 97.5% and 96.3% accuracy compared to RDT interpretation by experts in the Australia Flu Study and the Kenya Malaria Study, respectively.
We describe a specific implementation of the Alternating Direction Method of Multipliers (ADMM) algorithm for distributed optimization. This implementation runs logistic regression with L2 regularization over large datasets and does not require a user-tuned learning rate metaparameter or any tools beyond MapReduce. Throughout we emphasize the practical lessons learned while implementing an iterative MapReduce algorithm and the advantages of remaining within the Hadoop ecosystem.
There is a growing concern for malaria control in the Horn of Africa region due to the spread and rise in the frequency of Plasmodium falciparum Histidine-rich Protein (hrp) 2 and 3 deletions. Parasites containing these gene deletions escape detection by the major PfHRP2-based rapid diagnostic test. In this study, the presence of Pfhrp2/3 deletions was examined in uncomplicated malaria patients in Kilifi County, from a region of moderate-high malaria transmission. 345 samples were collected from the Pingilikani dispensary in 2019/2020 during routine malaria care for patients attending this primary health care facility. The Carestart™ RDT and microscopy were used to test for malaria. In addition, qPCR was used to confirm the presence of parasites. In total, 249 individuals tested positive for malaria by RDT, 242 by qPCR, and 170 by microscopy. 11 samples that were RDT-negative and microscopy positive and 25 samples that were qPCR-positive and RDT-negative were considered false negative tests and were examined further for Pfhrp2/3 deletions. Pfhrp2/3-negative PCR samples were further genotyped at the dihydrofolate reductase (Pfdhfr) gene which served to further confirm that parasite DNA was present in the samples. The 242 qPCR-positive samples (confirmed the presence of DNA) were also selected for Pfhrp2/3 genotyping. To determine the frequency of false negative results in low parasitemia samples, the RDT- and qPCR-negative samples were genotyped for Pfdhfr before testing for Pfhrp2/3. There were no Pfhrp2 and Pfhrp3 negative but positive for dhfr parasites in the 11 (RDT negative and microscopy positive) and 25 samples (qPCR-positive and RDT-negative). In the larger qPCR-positive sample set, only 5 samples (2.1%) were negative for both hrp2 and hrp3, but positive for dhfr. Of the 5 samples, there were 4 with more than 100 parasites/µl, suggesting true hrp2/3 deletions. These findings revealed that there is currently a low prevalence of Pfhrp2 and Pfhrp3 deletions in the health facility in Kilifi. However, routine monitoring in other primary health care facilities across the different malaria endemicities in Kenya is urgently required to ensure appropriate use of malaria RDTs.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.