Non-invasive hemoglobin measurement devices require refinement to match diagnostic performance with their high level of usability and acceptability

• Authors: Melissa Young, Emory University; Kelley Raines, Emory University; Farhad Jameel, DeKalb City Board of Health; Manal Sidi, Emory University; Shaiana Oliveira-Streiff, Emory University; Paula Nwajei, DeKalb City Board of Health; Katherine McGlamry, Emory University; Jiangda Ou, Emory University; Alawode Oladele, DeKalb City Board of Health; Parminder Suchdev, Emory University
• Journal Title: PLOS ONE
• Volume: Volume 16, Number 7
• Publisher: PUBLIC LIBRARY SCIENCE | 2021-07-16, Pages e0254629-e0254629
• Type of Work: Article | Final Publisher PDF
• Publisher DOI: http://doi.org/10.1371/journal.pone.0254629
• Subjects: Health Sciences, Nutrition; Sociology, Public and Social Welfare; Health Sciences, Public Health
• Keywords: Science & Technology; Multidisciplinary Sciences; Science & Technology - Other Topics; NUTRITIONAL-STATUS; SEVERE ANEMIA; CHILDREN; WOMEN; HEMOCUE; PALLOR; INCOME; BLOOD

Abstract

Anemia remains an important global health problem. Inexpensive, accurate, and noninvasive solutions are needed to monitor and evaluate anemia in resource-limited settings. We evaluated the performance of multiple point-of-care hemoglobin devices, including a novel noninvasive smartphone application tested on Apple® and Android® cell phones, Masimo Pronto®, and HemoCue® Hb-301 and Hb-801, against a gold-standard hematology analyzer (reference hemoglobin) using venous blood. We examined correlations between hemoglobin devices and reference hemoglobin, device accuracy (average bias, Bland-Altman plots, clinical performance) and classification bias (sensitivity, specificity) among 299 refugees (10mo-65y) in Atlanta, GA. Semi-structured interviews (n = 19) with participants and staff assessed usability and acceptability. Mean reference hemoglobin was 13.7 g/dL (SD:1.8) with 12.5% anemia. Noninvasive hemoglobin devices were not well correlated with reference hemoglobin (Apple® R2 = 0.08, Android® R2 = 0.11, Masimo Pronto® R2 = 0.29), but stronger correlations were reported with HemoCue® Hb-301 (R2 = 0.87) and Hb-801 (R2 = 0.88). Bias (SD) varied across each device: Apple®: -1.6 g/dL (2.0), Android®: -0.7 g/dL (2.0), Masimo Pronto®: -0.4 g/dL (1.6), HemoCue® Hb-301: +0.4 g/dL (0.7) and HemoCue® Hb-801: +0.2 g/dL (0.6). Clinically acceptable performance (within ± 1 g/dL of reference hemoglobin) was higher for the invasive devices (HemoCue® Hb-301: 90.3%; HemoCue® Hb-801: 93.4%) compared to noninvasive devices (Apple®: 31.5%; Android®: 34.6%; Masimo Pronto®: 49.5%). Sensitivity and specificity were 63.9% and 48.2% for Apple®, 36.1% and 67.6% for Android®, 45.7% and 85.3% for Masimo Pronto®, 54.3% and 97.6% for HemoCue® Hb-301, and 66.7% and 97.6% for HemoCue® Hb-801. Noninvasive devices were considered easy to use and were the preferred method by participants. Among the only studies to compare multiple point-of-care approaches to hemoglobin testing, the diagnostic ability of HemoCue® was comparable to reference hemoglobin, while noninvasive devices had high user acceptability but considerable biases. Improvements in noninvasive device performance and further testing in anemic populations are recommended before broader use.