The performance evaluation was conducted during a case–control study organized by FIND (global alliance for diagnostics) [16 ] with the help of the Institute of Endemic Diseases, University of Khartoum, Sudan (IEND), to evaluate the Malaria Screener software developed by the National Library of Medicine (NLM) at the National Institutes of Health (NIH). Patients were recruited at two primary hospitals in Sudan, one in the Alsororab (SOR) area and another one in the Gezira Slanj (GS) area, 40 and 50 km north of Khartoum where P. falciparum and Plasmodium vivax are endemic. The patients were recruited during the second malaria season between October 2020 and March 2021.
Sample size calculation was performed according to [17 ]. It was estimated that 100 patients positive for malaria (cases) by on-site microscopy (approx. 1.1xN) would need to be recruited for the evaluation to obtain a reliable estimate of the expected sensitivity, with 95% power of getting a 95% confidence interval of ± 10% or less, while allowing for procedural errors in 10% of all cases. Furthermore, it was estimated that 90 patients negative for malaria (controls) by on-site microscopy (approx. 1.4xN) would need to be recruited for the evaluation to obtain a reliable estimate of the expected specificity with 95% power of getting a 95% confidence interval of ± 10% or less while allowing for procedural errors in 10% of all controls.
Patients were enrolled consecutively until reaching the calculated numbers (190 patients in total, 95 from each site). Patients were five years of age and older. Patients with symptoms and signs of severe disease or comorbidities such as central nervous system or cardiovascular disease, as defined by World Health Organization (WHO) guidelines, were excluded, as were those who had received anti-malarial treatment in the four weeks before enrollment. Patients were enrolled after signing informed consent documents. Finger-prick blood samples were collected by a capillary tube to prepare blood smears, and dried blood spots (DBS) were prepared for PCR analysis. Figure1 describes the procedures that were performed during this study.![]()
Sample size calculation was performed according to [17 ]. It was estimated that 100 patients positive for malaria (cases) by on-site microscopy (approx. 1.1xN) would need to be recruited for the evaluation to obtain a reliable estimate of the expected sensitivity, with 95% power of getting a 95% confidence interval of ± 10% or less, while allowing for procedural errors in 10% of all cases. Furthermore, it was estimated that 90 patients negative for malaria (controls) by on-site microscopy (approx. 1.4xN) would need to be recruited for the evaluation to obtain a reliable estimate of the expected specificity with 95% power of getting a 95% confidence interval of ± 10% or less while allowing for procedural errors in 10% of all controls.
Patients were enrolled consecutively until reaching the calculated numbers (190 patients in total, 95 from each site). Patients were five years of age and older. Patients with symptoms and signs of severe disease or comorbidities such as central nervous system or cardiovascular disease, as defined by World Health Organization (WHO) guidelines, were excluded, as were those who had received anti-malarial treatment in the four weeks before enrollment. Patients were enrolled after signing informed consent documents. Finger-prick blood samples were collected by a capillary tube to prepare blood smears, and dried blood spots (DBS) were prepared for PCR analysis. Figure
Flow chart of the study procedures. For Malaria Screener, P. vivax samples were excluded since it can only process P. falciparum malaria. For PVF-Net, a newly developed deep learning-based algorithm, one mixed infection sample was excluded since it cannot process mixed infection
Free full text: Click here
