Hemoglobin point-of-care testing in rural Gambia: Comparing accuracy of HemoCue and Aptus with an automated hematology analyzer
Pereira, Dora I. A.
Public Library of Science
MetadataShow full item record
Nass, S. A., Hossain, I., Sanyang, C., Baldeh, B., & Pereira, D. I. A. (2020). Hemoglobin point-of-care testing in rural Gambia: Comparing accuracy of HemoCue and Aptus with an automated hematology analyzer. PLOS ONE, 15 (10)https://doi.org/10.1371/journal.pone.0239931
Background: Anemia is one of the most impactful nutrient deficiencies in the world and disproportionately affects children in low-resource settings. Point-of-care devices (PoCDs) measuring blood hemoglobin (Hb) are widely used in such settings to screen for anemia due to their low cost, speed, and convenience. Here we present the first iteration of Aptus, a new PoCD which measures Hb and hematocrit (HCT). Aim: To evaluate the accuracy of Aptus and HemoCue® Hb 301 against an automated hematology analyzer (Medonic®) in Gambian children aged 6–35 months and the Aptus’ usage in the field. Methods: Aptus, HemoCue® and Medonic® were compared using venous blood (n = 180), and Aptus and HemoCue® additionally using capillary blood (n = 506). Agreement was estimated using Bland-Altman analysis and Lin’s concordance. Usage was assessed by error occurrence and user experience. Results: Mean Hb values in venous blood did not significantly differ between Aptus and HemoCue® (10.44±1.05 vs 10.56±0.93g/dl, p>0.05), but both measured higher Hb concentrations than Medonic® (9.75±0.99g/dl, p<0.0001). Lin’s coefficient between Aptus and Medonic® was rc = 0.548, between HemoCue® and Medonic® rc = 0.636. Mean bias between the PoCDs venous measurements was -0.11g/dl with limits of agreement (LoA) -1.63 and 1.40g/dl. The bias was larger for the comparisons between the Medonic® and both Aptus (0.69g/dl, LoA 0.92 and 2.31g/dl) and HemoCue® (0.81g/dl, LoA 0.17 and 1.78g/dl). ROC curves showed an AUC of 0.933 in HemoCue® and 0.799 in Aptus. Capillary Hb was higher with Aptus than HemoCue® (10.33±1.11g/dl vs 10.01±1.07g/dl, p<0.0001). Mean bias was 0.32g/dl with LoA of -1.91 and 2.54g/dl. Aptus‘ usage proved intuitive, yet time-to-results and cuvettes could be improved. Conclusion: Both PoCDs showed a relatively limited bias but large LoA. Aptus and HemoCue® showed similar accuracy, while both overestimated Hb levels. Aptus showed promise, with its operation unimpaired by field conditions as well as being able to show HCT values.
Research Article, Biology and life sciences, Medicine and health sciences, Research and analysis methods
External DOI: https://doi.org/10.1371/journal.pone.0239931
This record's URL: https://www.repository.cam.ac.uk/handle/1810/310988
Attribution 4.0 International (CC BY 4.0)
Licence URL: https://creativecommons.org/licenses/by/4.0/