Evaluation of point‐of‐care testing device for anemia detection: A cross‐sectional method comparison study from Thailand

Abstract Background A comparison study is crucial before launching a new medical device; therefore, we compared the Mission Ultra Hb Testing System with the Sysmex XN‐3000 automated hematology analyzer in Thai adult males and non‐pregnant adult females. Methods Parallel studies were conducted using discarded venous K2‐ethylenediaminetetraacetic acid samples from participants requiring hematological investigations. According to the World Health Organization criteria, the participants were categorized as overall, anemia, and non‐anemia for analysis. Results Three hundred participants were included in this study. In all participants, near‐perfect correlation and agreement were observed between the two methods for Hb measurement (r = 0.963, p < 0.001) with an interclass correlation coefficient (ICC) of 0.981 (95% confidence interval [CI]: 0.976–0.985) and Hct measurement (r = 0.941, p < 0.001) with an ICC of 0.965 (95% CI: 0.956–0.972). The sensitivity and specificity of the device in detecting anemia were 86.2% (95% CI: 79.7–91.2) and 98.6% (95% CI: 95.2–99.8), respectively. The area under the curve was 0.976 (95% CI: 0.963–0.989). The device showed average biases of 0.76 g/dl (95% limits of agreement [LOA]: −1.03 to 2.54) for Hb measurement and −2.73% (95% LOA: −9.28 to 3.82) for Hct measurement in all participants. Conclusion Agreement between the Mission Ultra Hb Testing System and Sysmex XN‐3000 was observed. The device was excellent for detecting anemia. However, the essential evidence showing biases of the Hb and Hct measurements obtained from the device was revealed. Laboratory interpretation should be carefully performed, particularly at the near cut‐off values.


| INTRODUC TI ON
Anemia is a serious public health problem worldwide. Based on data from the Global Burden of Diseases, Injuries, and Risk Factors Study, it affected approximately 32.9% of the world's population in 2010. 1,2 Interestingly, it has been reported that low-and middle-income countries have the highest prevalence of anemia. 2 In Thailand, the prevalence of anemia varies from 11.9% to 53.3% according to population and geographic distribution. 3 Nevertheless, the reported prevalence may be underestimated due to poverty and the lack of opportunities for people residing in remote areas to access the public health system.
Several methods are commonly used to assess an individuals' hemoglobin (Hb) and hematocrit (Hct) levels, including automated hematology analyzers, microhematocrit centrifuge, gravimetric copper sulfate method, cyanmethemoglobin method, and color code Hb estimation. [4][5][6][7] Although automated hematology analyzers are the gold standard and produce significantly accurate results for diagnosing anemia, their availability is limited to the regional, provincial, and district hospital levels in Thailand. The microhematocrit centrifuge must be of good quality and should be well maintained to provide accurate results. Moreover, it is not practical to transport the centrifuges to the remote areas of Thailand where an electrical power network supply may be inaccessible. The cyanmethemoglobin method is time-consuming and labor-intensive, 6,7 whereas both the gravimetric copper sulfate method and color code Hb estimation are subjective as estimator bias may be introduced. 4,5,7 Therefore, in resource-poor settings, a handheld or portable point-of-care testing (POCT) device for Hb and Hct level assessment potentially plays a key role as a tool for anemia detection, improving the problems of delayed diagnosis and interventions that lead to increased morbidity and mortality. Additionally, it could be a powerful tool used in demographic and health surveys for a nationwide epidemiological study.

| Study participants and design
The study subjects were adult males and non-pregnant adult females aged 18 years and older who required hematological investigations during their routine visits to the Internal Medicine Outpatient Clinic.
The sample size was calculated using the prevalence of anemia in the Thai working-age population of 17.3% 8 ; estimated sensitivity and specificity of 98% and 95%, respectively 7 ; the level of significance of 5%; and the level of estimation error of 5%. We estimated that at least 175 participants were needed, and at least 89 anemic patients were included. Venous K2-ethylenediaminetetraacetic acid (EDTA) samples (Greiner Bio-One, Chon Buri, Thailand) were sent to the Department of Clinical Pathology to be tested for clinical care using the Sysmex XN-3000, then the discarded samples were immediately given to medical technologists for further testing, so that no extra blood was collected. Before participating in the study, the medical technologists were trained on the use of a Mission Ultra Hb meter based on the manufacturer's specifications and were blinded to the results from the Sysmex XN-3000. Using 4 µl of discarded blood, parallel tests were performed using the Mission Ultra Hb meter within a maximum of 4 h of sample collection each day. For data analysis, the participants were categorized as overall, anemia, and non-anemia. Anemia was defined according to the World Health Organization (WHO) criteria, that is, Hb level below 12.0 g/dl in nonpregnant adult females and below 13.0 g/dl in adult males 9 based on the results from the Sysmex XN-3000, which was the reference method in this study.

| Point-of-care testing method
The Mission Ultra Hb Testing System (ACON Laboratories, Inc., San Diego, CA, USA) consists of a handheld meter and disposable test strips. One lot of Mission Ultra Hb test strips was used (REF. C131-6011, LOT. BHB0050001) throughout the study. The device uses the principle of electrochemistry for Hb detection, 10 while Hct is measured by electrical impedance. 11,12 There are three modes available for blood sample testing, including whole blood anticoagulated with K2-or K3-EDTA, or capillary blood.

| Automated reference method
The Sysmex XN-3000 hematology analyzer (Sysmex Corp., Kobe, Japan) is an automated blood cell counter for diagnostic use in clinical laboratories. Sysmex XN-3000 uses cyanide-free sodium lauryl sulfate (SLS) to determine Hb levels, whereas Hct level is measured based on the principle of hydrodynamic focusing. 13

| Quality control
The function of the Mission Ultra Hb meter was checked daily by measuring the three-set standards of known concentration (low, normal, and high) provided by the manufacturer. Three-set controls (low, normal, and high) were run daily to ensure adequate functionality of

| Statistical analyses
The data were analyzed using Stata Statistical Software Release 17 Hb + Sysmex)/2]) on the horizontal axis, and the mean ± standard deviation (SD) of the differences and 95% confidence interval (95% CI) of the limits of agreement (LOA) (95% LOA) were calculated.
The diagnostic accuracy of the device for detecting anemia was determined by sensitivity, specificity, predictive values, and the area under the receiver operating characteristic (ROC) curve (AUC). The level of statistical significance was set at p < 0.05.

| Summary statistics of the participants and the hemoglobin and hematocrit levels
The summary statistics of the participants and the Hb and Hct levels are presented in Table 1. Anemia was defined according to the WHO criteria, 9 based on the results from the Sysmex XN-3000. A total of 300 participants were involved in the study and were aged between 20 and 94 years, with 50.7% of them being anemic. Males accounted for 60.7% of the participants. Unsurprisingly, anemia was more prevalent among females (63.6%) than among males and it was more commonly found in older participants. Considering each subpopulation, the results of the Hb level (g/dL) obtained from the Mission Ultra Hb meter were higher than those obtained from the Sysmex XN-3000 in all subpopulation categories (p < 0.001). In contrast, compared with the Sysmex XN-3000, the percentages of Hct obtained from the Mission Ultra Hb were lower in all subpopulation categories (p < 0.001).

| Correlation and agreement between the mission ultra Hb testing system and the sysmex XN-3000
As measured by the Mission Ultra Hb meter and the Sysmex XN-3000, in all participants, near-perfect correlation and agreement were observed for Hb measurement ( Figure 1A) and Hct measurement ( Figure 1D). In anemic participants, excellent correlation and agreement were also observed for Hb measurement ( Figure 1B) and Hct measurement ( Figure 1E). Nevertheless, the correlation and agreement slightly decreased in non-anemic participants for Hb measurement ( Figure 1C) and for Hct measurement ( Figure 1F). Compared with the Sysmex XN-3000, the Bland-Altman plots proficiency testing for Hb and Hct measurements was defined as ±4% of the target. 15 Therefore, the Mission Ultra Hb meter failed to meet the indicated cut-off value for both Hb and Hct measurements.

| Accuracy of the mission ultra Hb testing system as a tool for detection of anemia
Using the Sysmex XN-3000 as the reference method and the WHO The majority of the Thai population is working age 18 with a continuously growing elderly population. Therefore, we focused on young, middle-aged, and older adults in the present study.
A variety of statistical tests were used to determine the clin- Consistently, a high AUC value indicated that the accuracy of the device was outstanding in its overall performance for detecting

anemia. The expense is significantly lower for the Mission Ultra
Hb test compared with a complete blood count from an automated analyzer. Therefore, in terms of cost-effectiveness, it is worthwhile to use a POCT device as an initial screening tool for anemia in resource-poor settings.  Figure 2B) above the Hb cut-off values for anemia according to the WHO criteria. 9 In contrast, in the case of serial Hct monitoring in critical care settings, a blood transfusion may be prescribed unnecessarily using a device with negative bias ( Figure 2F). Another point of concern is pre-analytical errors.

ACK N OWLED G EM ENTS
The authors would like to thank Supak Ukritchon for statistical analysis and data visualization, Dollapak Apipongrat for technical support, staff members of the Department of Pathology, Phramongkutklao

College of Medicine and the Department of Clinical Pathology, and
Phramongkutklao Hospital for their support in completing this study.

CO N FLI C T O F I NTE R E S T
The authors declare that no competing financial interests exist. The database on Anaemia includes data by country on prevalence of anaemia and mean haemoglobin concentration. Thailand 2006. https://www.who.int/vmnis/ anaem ia/data/datab ase/count ries/ tha_ida.pdf?ua=1. Accessed June 1, 2021.

F I G U R E 3
The receiver operating characteristic curve of the Mission Ultra Hb meter as a detective tool for anemia based on the hemoglobin level in all participants