Evaluation of Different Methods of Haemoglobin Estimation Among Malaria‐Infected Patients Authors: Mina Ofosu, Jessica Ntiamoah, Patricia Sam Barko, Prince Elikem Torli Published: 2026 ABSTRACT ABSTRACTBackgroundHaemoglobin (Hb) estimation is essential for assessing anaemia, particularly in malaria‐endemic settings. Automated Full Blood Count (FBC) analysers, cyanmethemoglobin spectrophotometry, and portable Hb meters are commonly used, but differences in analytical performance may affect measurements.MethodsThis cross‐sectional study compared Hb measurements obtained using an automated FBC analyser, the cyanmethemoglobin method, and a portable Hb meter among 100 malaria‐infected patients aged 16–65 years at Kumasi South Hospital, Ghana. Measurements were compared using Pearson's correlation coefficient, intraclass correlation coefficient (ICC), Lin's concordance correlation coefficient (CCC), Bland–Altman analysis, percentage error, and Passing–Bablok regression.ResultsThe three methods showed very strong correlations (r = 0.955–0.992, p < 0.001), excellent ICC (single measures = 0.969; average measures = 0.990), and high concordance (CCC = 0.95–0.99). Mean biases were −0.06 g/dL for FBC analyser versus portable Hb meter, −0.25 g/dL for FBC analyser versus cyanmethemoglobin, and −0.20 g/dL for portable Hb meter versus cyanmethemoglobin. The corresponding 95% limits of agreement were −0.59 to 0.47, −1.48 to 0.98, and −1.35 to 0.95 g/dL, respectively. Passing–Bablok slopes were 1.00 for all comparisons. The FBC analyser and portable Hb meter showed the closest agreement.ConclusionThe three methods demonstrated good agreement and broadly comparable Hb measurements. The findings support the practical use of portable Hb meters where point‐of‐care testing is required, while recognising that agreement does not establish diagnostic accuracy or analytical validation.Keywords: agreement, cyanmethemoglobin, full blood count analyser, haemoglobin estimation, haemoglobin meter, malaria Introduction Haemoglobin (Hb) is an iron‐containing porphyrin protein responsible for transporting oxygen from the lungs to body tissues, thereby supporting cellular energy metabolism [1]. In addition to oxygen transport, haemoglobin contributes to carbon dioxide transport, hydrogen‐ion buffering, and nitric oxide metabolism, highlighting its physiological importance and the central role of haemoglobin estimation in the diagnosis and management of anaemia and other haematological disorders [2]. Anaemia, defined as a haemoglobin concentration below the level required to meet an individual's physiological needs, remains the most common haematological disorder worldwide, affecting approximately one‐third of the global population [3]. According to the World Health Organization (WHO), haemoglobin concentration remains the primary laboratory indicator for diagnosing and classifying anaemia and for monitoring treatment response and disease severity [4, 5]. Although clinical features such as fatigue, pallor, and tachycardia may suggest anaemia, laboratory confirmation is essential for accurate diagnosis and appropriate clinical management [6]. Anaemia remains an important public health concern in Ghana, particularly in malaria‐endemic settings where malaria‐associated reductions in haemoglobin make reliable haemoglobin estimation important for patient assessment and management [7]. Following transmission by infected Anopheles mosquitoes, Plasmodium parasites undergo hepatic and erythrocytic stages of development, ultimately invading and destroying red blood cells while impairing erythropoiesis [8, 9]. These pathophysiological processes result in haemolysis, splenic sequestration of infected and uninfected erythrocytes, and reduced haemoglobin concentrations, making accurate haemoglobin estimation an integral component of malaria diagnosis, clinical assessment, and therapeutic monitoring [10, 11]. Furthermore, malaria‐associated alterations in blood composition may influence haemoglobin measurement and contribute to variation between analytical methods [12]. Several methods are currently available for haemoglobin estimation, including automated Full Blood Count (FBC) analysers, the cyanmethemoglobin method, and portable haemoglobin (Hb) meters. The cyanmethemoglobin (Drabkin's) method remains the internationally recognised established reference method for haemoglobin determination and is recommended by both the World Health Organization and the International Council for Standardization in Haematology because of its standardized spectrophotometric principle and established analytical performance [13, 14]. Automated FBC analysers provide rapid, precise, and reproducible measurement of haemoglobin together with comprehensive haematological parameters using technologies such as electrical impedance, flow cytometry, and light scattering [15, 16]. However, their widespread use in resource‐limited settings may be constrained by equipment costs, infrastructure requirements, maintenance demands, and the need for trained personnel [17, 18]. Conversely, portable Hb meters have become increasingly popular because they provide rapid, convenient, and point‐of‐care haemoglobin estimation with minimal infrastructure requirements, although concerns remain regarding their comparability with laboratory‐based methods [19, 20]. Previous studies have compared different haemoglobin estimation methods in clinical and field settings, demonstrating varying levels of correlation and agreement depending on the analytical method, sample type, and study population [21, 22]. However, evidence comparing these methods specifically among malaria‐infected patients remains limited, particularly in malaria‐endemic and resource‐limited settings where accurate, affordable, and practical haemoglobin estimation is essential for patient care. Given that malaria‐associated physiological changes may influence haemoglobin measurement, further evaluation of method agreement in this population is warranted. Therefore, this study compared haemoglobin concentrations obtained using an automated FBC analyser, the cyanmethemoglobin method, and a portable Hb meter among malaria‐infected patients to evaluate their agreement, comparability, and practical applicability. Materials and Methods Study Design and Setting A hospital‐based cross‐sectional study was conducted among malaria‐infected patients attending Kumasi South Hospital, Kumasi, Ghana, between April 2025 and June 2025. Kumasi South Hospital is a secondary‐level public healthcare facility that serves as a referral centre for surrounding communities and routinely provides malaria diagnosis, treatment, and comprehensive laboratory services. The study was designed to compare haemoglobin (Hb) concentrations obtained using an automated Full Blood Count (FBC) analyser, the cyanmethemoglobin‐based established reference method, and a portable Hb meter. Study Population The study included laboratory‐confirmed malaria‐infected patients aged 16–65 years who attended Kumasi South Hospital during the study period and provided written informed consent. Patients who had received blood transfusions within the preceding 3 months, those with known haematological disorders unrelated to malaria, or individuals with incomplete clinical or laboratory data were excluded from the study. Sample Size and Sampling The minimum sample size was determined using Cochran's formula: where n represents the sample size, Z is the standard normal deviate at a 95% confidence level (1.96), p is the estimated prevalence (0.5), and e is the margin of error (0.095) n = 106.4. The calculated minimum sample size was 106 participants. A total of 100 eligible participants were ultimately recruited, representing approximately 94% of the calculated sample size. This slight shortfall was due to participant availability during the predefined study period. Participants were recruited using a consecutive convenience sampling approach, whereby all eligible malaria‐infected patients who presented to the hospital during the study period and met the inclusion criteria were invited to participate until the target sample size was achieved. Although probability sampling may reduce selection bias, this approach was considered appropriate because recruitment was restricted to patients with confirmed malaria presenting during the study period. Blood Sample Collection Approximately 5 mL of venous blood was collected aseptically from each participant by trained laboratory personnel using standard phlebotomy procedures. Blood samples were dispensed into ethylenediaminetetraacetic acid (EDTA) anticoagulant tubes, labelled with unique identification numbers, and transported promptly to the laboratory for analysis. All samples were analysed as soon as possible after collection to minimise pre‐analytical variation. Laboratory Analysis Haemoglobin Estimation Using the Automated Full Blood Count (FBC) Analyser Approximately 2 mL of EDTA‐anticoagulated whole blood was gently mixed by repeated inversion to ensure sample homogeneity before analysis. Haemoglobin concentration was measured using the Sysmex XN‐550 automated Full Blood Count analyser (Sysmex Corporation, Kobe, Japan) according to the manufacturer's operating instructions. The analyser determines haemoglobin concentration using the sodium lauryl sulfate (SLS)‐haemoglobin method while simultaneously providing complete blood count parameters through impedance and flow cytometric technologies. Following sample aspiration through the analyser probe, haemoglobin values were automatically generated and recorded in g/dL for statistical analysis. Haemoglobin Estimation Using the Cyanmethemoglobin Method Haemoglobin concentration was determined using the cyanmethemoglobin (Drabkin's) reference method in accordance with the recommendations of the International Council for Standardization in Haematology. Three test tubes labelled Blank (B), Standard (S), and Test (T) were prepared, and 5 mL of Drabkin's reagent was dispensed into each tube. Twenty microlitres (20 μL) of whole blood was added to the tube labelled Test and mixed thoroughly. The reaction mixture was allowed to stand at room temperature for 15 min to ensure complete conversion of haemoglobin derivatives (except sulfhaemoglobin) to cyanmethemoglobin. The spectrophotometer was calibrated at 540 nm using the reagent blank before measuring the absorbance of the standard and test samples. Haemoglobin concentration was calculated using the equation: Hb (g/dL) = Absorbance of testAbsorbance of standard× Concentration of standard. The procedure was repeated for all study samples. Haemoglobin Estimation Using the Portable Hb Meter Haemoglobin concentration was measured using a portable Hb meter (HemoCue Hb 301, HemoCue AB, Ängelholm, Sweden) according to the manufacturer's instructions. Fresh EDTA‐anticoagulated whole blood was applied to a compatible disposable test strip inserted into the analyser, and the haemoglobin concentration was displayed electronically within seconds and recorded in g/dL. The device was checked using the manufacturer's quality‐control procedures before routine sample analysis and maintained according to the manufacturer's recommendations throughout the study. Anaemia Classification Anaemia severity was classified according to the World Health Organization (WHO) haemoglobin cut‐offs for adults aged 15–65 years. The WHO thresholds were applied directly to haemoglobin values obtained from all three methods (Sysmex‐FBC analyser, portable Hb meter, and cyanmethemoglobin method) to grade anaemia severity [4]. Quality Assurance All laboratory analyses were performed by qualified laboratory personnel following standard operating procedures. The automated FBC analyser underwent routine maintenance, calibration, and internal quality‐control verification according to the manufacturer's recommendations before participant sample analysis. For the cyanmethemoglobin method, freshly prepared Drabkin's reagent, reagent blanks, and haemoglobin standards were used to ensure assay accuracy, and the spectrophotometer was calibrated and blanked before sample analysis. The portable Hb meter was operated according to the manufacturer's instructions, and quality‐control procedures were conducted before each testing session to ensure consistent instrument performance. Duplicate measurements were performed as part of routine quality assurance, and their mean values were used for analysis; formal repeatability and coefficient‐of‐variation analyses were not undertaken. Data Entry and Statistical Analysis Data were entered into a computerized database, cleaned, and validated before analysis. Statistical analyses were performed using IBM SPSS Statistics version 26.0 and R software version 4.2. Continuous variables were assessed for normality and summarized as mean ± standard deviation, while categorical variables were presented as frequencies and percentages. Pearson's correlation coefficient (r) was used to assess the strength of the linear relationship between haemoglobin measurements obtained by the different methods. Agreement among the methods was assessed using the intraclass correlation coefficient (ICC) based on a two‐way mixed‐effects model with absolute agreement and Lin's concordance correlation coefficient (CCC). Pairwise differences between methods were evaluated using one‐way analysis of variance (ANOVA), followed by appropriate post hoc pairwise comparisons. Bland–Altman analysis was performed for each pair of haemoglobin estimation methods to assess systematic bias and agreement at the individual measurement level. Mean bias and the corresponding 95% limits of agreement (LoA) were calculated and presented graphically using Bland–Altman plots. Percentage error was also calculated, and Passing–Bablok regression analysis was performed for each pairwise comparison. The regression slope and intercept were reported for each comparison. Statistical significance was set at p < 0.05. Results were presented in tables and figures, as appropriate. Results Participants' Characteristics A total of 100 malaria‐infected participants were enrolled in the study (Table 1). Females constituted the majority of the study population (61.0%, n = 61), whereas males accounted for 39.0% (n = 39). The age distribution showed that participants aged 20–29 years formed the largest proportion of the cohort (32.0%, n = 32). Participants aged < 20 years and 30–39 years each represented 19.0% of the study population. Smaller proportions were observed among participants aged 40–49 years (11.0%, n = 11), ≥ 60 years (11.0%, n = 11), and 50–59 years (8.0%, n = 8). TABLE 1Participants' characteristics.Variables n = 100Percent (%)GenderMale3939.0Female6161.0Age groups (years)< 201919.020–293232.030–391919.040–491111.050–5988.0≥ 601111.0 Participants' characteristics. WHO Classification of Anaemia Severity Participants were classified according to the World Health Organization (WHO) haemoglobin cut‐offs for adults aged 15–65 years, using haemoglobin concentrations obtained by the three methods independently. The distribution of anaemia severity across the methods is graphically represented in Figure 1. The majority of participants had no anaemia, with 61 participants (61.0%) classified as having no anaemia by all three methods. Mild anaemia was observed in 17 participants (17.0%) by the FBC analyser and cyanmethemoglobin method, and 16 participants (16.0%) by the portable Hb meter. Moderate anaemia was observed in 20 participants (20.0%) by the FBC analyser and 22 participants (22.0%) by both the cyanmethemoglobin method and the portable Hb meter. Severe anaemia was uncommon, with 2 participants (2.0%) classified as severely anaemic by the FBC analyser, 1 participant (1.0%) by the portable Hb meter, and none by the cyanmethemoglobin method. Distribution of anaemia severity by haemoglobin estimation method. Age‐Stratified Comparison of Haemoglobin Estimation Methods Table 2 presents the mean haemoglobin concentrations obtained using the automated FBC analyser, portable Hb meter, and cyanmethemoglobin method across different age groups. Mean haemoglobin concentrations obtained by the three methods were generally comparable across all age categories. No statistically significant differences were observed among participants aged < 20 years, 20–29 years, 40–49 years, 50–59 years, and ≥ 60 years (p > 0.05). However, a statistically significant difference was observed among participants aged 30–39 years (p = 0.003), with the cyanmethemoglobin method yielding the highest mean haemoglobin concentration (13.20 ± 2.22 g/dL), followed by the portable Hb meter (12.85 ± 2.21 g/dL) and the automated FBC analyser (12.71 ± 2.19 g/dL). TABLE 2Age‐stratified comparison of haemoglobin estimation methods.VariableFBC analyserPortable Hb meterCyanmethemoglobin method p‐valueMean ± SD (g/dL)Age groups (years)< 2013.08 ± 2.0113.06 ± 2.0713.27 ± 2.010.17420–2912.18 ± 2.4112.24 ± 2.4612.42 ± 2.410.09930–3912.71 ± 2.1912.85 ± 2.2113.20 ± 2.22 0.003 40–4913.65 ± 1.3113.72 ± 1.4213.87 ± 1.380.35350–5912.58 ± 1.5112.60 ± 1.5312.85 ± 1.390.088≥ 6012.25 ± 1.8912.26 ± 1.8612.26 ± 1.750.941 Note: Values presented in bold indicate statistically significant results, with statistical significance defined at the predetermined threshold of p < 0.05. Age‐stratified comparison of haemoglobin estimation methods. Note: Values presented in bold indicate statistically significant results, with statistical significance defined at the predetermined threshold of p < 0.05. Comparison of Haemoglobin Estimation Methods According to Gender and Age Group Table 3 presents haemoglobin concentrations stratified by gender and age group. Across most gender–age subgroups, haemoglobin concentrations obtained by the three methods were comparable, with no statistically significant differences observed (p > 0.05). Statistically significant differences were identified among males aged ≥ 60 years (p = 0.040) and females aged 30–39 years (p = 0.022). Among males aged ≥ 60 years, mean haemoglobin concentrations were 14.33 ± 1.01 g/dL, 14.20 ± 1.04 g/dL, and 14.07 ± 1.07 g/dL for the automated FBC analyser, portable Hb meter, and cyanmethemoglobin method, respectively. Similarly, among females aged 30–39 years, the corresponding haemoglobin concentrations were 11.90 ± 1.67 g/dL, 12.06 ± 1.68 g/dL, and 12.31 ± 1.60 g/dL, respectively. TABLE 3Comparison of haemoglobin estimation by gender and age groups.GenderAge ranges (years)FBC analyserPortable Hb meterCyanmethemoglobin method p‐valueMean ± SDMale< 2013.60 ± 1.5813.57 ± 1.6113.88 ± 1.620.13220–2915.14 ± 1.2215.27 ± 1.2715.33 ± 0.970.63630–3914.96 ± 2.0015.06 ± 2.1215.66 ± 1.870.10740–4914.13 ± 1.3014.23 ± 1.4914.29 ± 1.480.77250–5912.50 ± 1.9312.50 ± 1.9312.80 ± 1.760.300≥ 6014.33 ± 1.0114.20 ± 1.0414.07 ± 1.07 0.040 Female< 2012.20 ± 2.4612.20 ± 2.5912.24 ± 2.320.89020–2911.19 ± 1.8211.22 ± 1.8411.45 ± 1.900.09430–3911.90 ± 1.6712.06 ± 1.6812.31 ± 1.60 0.022 40–4912.80 ± 0.9012.83 ± 0.7813.15 ± 0.930.28950–5912.65 ± 1.2512.70 ± 1.3112.90 ± 1.190.401≥ 6011.46 ± 1.4911.54 ± 1.5611.58 ± 1.440.647 Note: Values presented in bold indicate statistically significant results, with statistical significance defined at the predetermined threshold of p < 0.05. Comparison of haemoglobin estimation by gender and age groups. Note: Values presented in bold indicate statistically significant results, with statistical significance defined at the predetermined threshold of p < 0.05. Pairwise Comparison of Haemoglobin Estimation Methods Pairwise comparisons of haemoglobin concentrations obtained using the three methods are presented in Table 4. No statistically significant difference was observed between the automated FBC analyser and the portable Hb meter (mean difference = −0.056 g/dL, p = 0.123). However, statistically significant differences were observed between the automated FBC analyser and the cyanmethemoglobin method (mean difference = −0.254 g/dL, p < 0.001), as well as between the portable Hb meter and the cyanmethemoglobin method (mean difference = −0.198 g/dL, p = 0.003). These findings indicate that the cyanmethemoglobin method consistently produced slightly higher haemoglobin concentrations than the other two methods. TABLE 4Pairwise comparison of hemoglobin estimation by each method.MethodMean difference (g/dL) p‐valueFBC analyserPortable Hb meter−0.0560.123Cyanmethemoglobin method−0.254 < 0.001 Portable Hb meterFBC analyser0.0560.123Cyanmethemoglobin method−0.198 0.003 Cyanmethemoglobin methodFBC analyser0.254 < 0.001 Portable Hb meter0.198 0.003 Note: Values presented in bold indicate statistically significant results, with statistical significance defined at the predetermined threshold of p < 0.05. Pairwise comparison of hemoglobin estimation by each method. Note: Values presented in bold indicate statistically significant results, with statistical significance defined at the predetermined threshold of p < 0.05. Correlation Analysis of Haemoglobin Estimation Methods Pearson correlation analysis demonstrated very strong positive linear relationships between haemoglobin concentrations obtained using the three estimation methods (Table 5). The strongest correlation was observed between the automated FBC analyser and the portable Hb meter (r = 0.992, p < 0.001), followed by the portable Hb meter and the cyanmethemoglobin method (r = 0.961, p < 0.001), and the automated FBC analyser and the cyanmethemoglobin method (r = 0.955, p < 0.001). TABLE 5Correlation analysis of haemoglobin estimation.ComparisonsCorrelation coefficient (r) p‐valueFBC analyser &portable Hb meter0.992 < 0.001 FBC analyser & cyanmethemoglobin method0.955 < 0.001 Portable Hb meter & cyanmethemoglobin method0.961 < 0.001 Note: Values presented in bold indicate statistically significant results, with statistical significance defined at the predetermined threshold of p < 0.05. Correlation analysis of haemoglobin estimation. Note: Values presented in bold indicate statistically significant results, with statistical significance defined at the predetermined threshold of p < 0.05. Agreement Analysis Using the Intraclass Correlation Coefficient Agreement among the three haemoglobin estimation methods was further evaluated using the intraclass correlation coefficient (ICC) based on a two‐way mixed‐effects model with absolute agreement (Table 6). The single‐measures ICC was 0.969 (95% CI: 0.958–0.978; F = 95.846; p < 0.001), indicating excellent agreement among individual measurements obtained using the three methods. Similarly, the average‐measures ICC was 0.990 (95% CI: 0.985–0.993; F = 95.846; p < 0.001), demonstrating excellent overall agreement across the haemoglobin estimation methods. TABLE 6Agreement analysis using the intraclass correlation coefficient.Intraclass correlation coefficient95% CI F‐test value p‐valueSingle measures0.9690.958–0.97895.846 < 0.001 Average measures0.9900.985–0.99395.846 < 0.001 Note: Values presented in bold indicate statistically significant results, with statistical significance defined at the predetermined threshold of p < 0.05. Agreement analysis using the intraclass correlation coefficient. Note: Values presented in bold indicate statistically significant results, with statistical significance defined at the predetermined threshold of p < 0.05. Agreement and Method‐Comparison Analysis of Haemoglobin Estimation Methods Agreement between the three haemoglobin estimation methods was assessed using Lin's concordance correlation coefficient (CCC), Bland–Altman analysis, percentage error, and Passing–Bablok regression as shown in Table 7. TABLE 7Agreement analysis and method‐comparison statistics for haemoglobin estimation methods.ComparisonLin's CCCMean bias (g/dL)95% LoA (g/dL)Percentage error (%)PB slopePB interceptFBC analyser vs Portable Hb meter0.99−0.06−0.59 to 0.474.21.000.10FBC analyser vs cyanmethemoglobin0.95−0.25−1.48 to 0.989.61.000.30Hb meter vs cyanmethemoglobin0.96−0.20−1.35 to 0.959.01.000.25Abbreviations: CCC, Concordance Correlation Coefficient; LoA‐Limits of Agreement; PB, Passing–Bablok. Agreement analysis and method‐comparison statistics for haemoglobin estimation methods. Abbreviations: CCC, Concordance Correlation Coefficient; LoA‐Limits of Agreement; PB, Passing–Bablok. The automated FBC analyser and portable Hb meter demonstrated the highest concordance, with a Lin's CCC of 0.99. The mean bias between the two methods was −0.06 g/dL, with 95% limits of agreement ranging from −0.59 to 0.47 g/dL. The corresponding percentage error was 4.2%. The comparison between the automated FBC analyser and the cyanmethemoglobin method yielded a Lin's CCC of 0.95, with a mean bias of −0.25 g/dL and 95% limits of agreement ranging from −1.48 to 0.98 g/dL. The percentage error was 9.6%. Similarly, the comparison between the Hb meter and cyanmethemoglobin method demonstrated a Lin's CCC of 0.96, a mean bias of −0.20 g/dL, and 95% limits of agreement ranging from −1.35 to 0.95 g/dL, with a percentage error of 9.0%. Passing–Bablok regression produced slopes of 1.00 for all three pairwise comparisons, with intercepts of 0.10 g/dL for FBC analyser versus portable Hb meter, 0.30 g/dL for FBC analyser versus cyanmethemoglobin, and 0.25 g/dL for portable Hb meter versus cyanmethemoglobin. These findings indicate close linear concordance between the measurement methods. Bland–Altman plots illustrating the pairwise differences between the methods are presented in Figures 2, 3, 4. Bland–Altman plot of haemoglobin measurements obtained using the automated FBC analyser and Portable Hb meter. Bland–Altman plot of haemoglobin measurements obtained using the automated FBC analyser and cyanmethemoglobin method. Bland–Altman plot of haemoglobin measurements obtained using the Hb meter and cyanmethemoglobin method. Discussion Principal Findings This study compared haemoglobin concentrations obtained using an automated FBC analyser, a portable Hb meter, and the cyanmethemoglobin method among malaria‐infected patients. The three methods demonstrated strong linear relationships and high overall agreement. The closest agreement was observed between the automated FBC analyser and the portable Hb meter, while the cyanmethemoglobin method produced slightly higher haemoglobin concentrations on average than the other two methods. Bland–Altman analysis demonstrated small mean biases across the three pairwise comparisons, although the width of the limits of agreement differed between method pairs. These findings indicate that the three methods produced broadly comparable haemoglobin measurements in the study population, while also demonstrating that correlation alone is insufficient to establish agreement between analytical methods or to infer interchangeability of measurements. Age, Gender, and Anaemia Severity Specific Performance Haemoglobin measurements obtained using the three methods were generally comparable across the age groups examined. The statistically significant differences observed in the 30–39‐year group and among males aged ≥ 60 years should, however, be interpreted cautiously because some age‐ and sex‐specific subgroups contained relatively few participants, reducing the precision of subgroup estimates and increasing the possibility of chance findings. Among females aged 30–39 years, the cyanmethemoglobin method produced slightly higher haemoglobin values than the automated FBC analyser and portable Hb meter. This age group includes many women of reproductive age, among whom anaemia and iron deficiency remain important public health concerns, partly because of menstrual iron loss and increased iron requirements [2]. The observed differences in this subgroup should not, however, be interpreted as evidence of a systematic age‐ or sex‐specific measurement effect because the study was not primarily powered for stratified comparisons. Moreover, differences between haemoglobin measurement methods have been reported in previous method‐comparison studies, supporting the need to interpret subgroup variations within the broader analytical agreement between methods [23]. Anaemia severity classification showed that the distribution of anaemia categories was consistent across the three methods, with the majority of participants having no anaemia (61.0% across all methods) and severe anaemia being uncommon (0–2 participants per method). The consistency of classification across methods supports the overall agreement observed in the study. The small number of participants with severe anaemia prevents a formal subgroup‐specific agreement analysis for this category, as such analyses would be underpowered and at increased risk of Type I error. Among males aged ≥ 60 years, the automated FBC analyser produced slightly higher haemoglobin values than the cyanmethemoglobin method. Age‐related changes in haematopoiesis and haemoglobin concentrations have been documented in older adults and may contribute to biological variation in haemoglobin measurements [24, 25]. However, given the relatively small number of participants in this subgroup, this finding should be regarded as exploratory rather than evidence of a consistent age‐related analytical effect. Overall, the broadly comparable measurements across most age and sex groups and the consistent distribution of anaemia severity categories support the comparability of the three methods across the demographic and clinical spectrum of the study population, while the subgroup differences highlight the importance of cautious interpretation when sample sizes are limited. Agreement, Correlation, and Concordance Among Methods The overall pattern of agreement indicates that the three haemoglobin estimation methods produced broadly comparable measurements, with the closest agreement observed between the automated FBC analyser and portable Hb meter. The greater variability observed in comparisons involving the cyanmethemoglobin method may reflect differences in analytical principles, calibration, and measurement platforms, as reported in previous haemoglobin method‐comparison studies [12]. The small mean biases and percentage errors observed across the comparisons further indicate limited systematic differences between the methods, although the wider limits of agreement involving cyanmethemoglobin suggest greater variability at the individual measurement level. The clinical relevance of these differences should therefore be considered in relation to the intended application and relevant allowable‐error criteria [26, 27]. The Passing–Bablok regression results provide additional support for the absence of proportional differences between the methods, as the slope was 1.00 for all three pairwise comparisons. The observed intercepts, particularly in comparisons involving cyanmethemoglobin, suggest the presence of some fixed differences between methods. Together with the Bland–Altman findings, these results indicate that the differences between methods were largely systematic rather than progressively increasing across the haemoglobin measurement range. The strong Pearson correlations, ICC, and Lin's concordance correlation coefficients are consistent with previous evaluations of laboratory‐based and point‐of‐care haemoglobin systems [2, 13]. However, correlation measures the strength of association rather than the closeness of individual measurements and therefore does not establish agreement or interchangeability [26]. Likewise, concordance and reliability do not establish diagnostic accuracy against an independent reference standard. This distinction is important because analytical agreement and clinical accuracy represent different concepts. Although the three methods demonstrated good agreement and strong correlations, the small systematic differences observed between methods—particularly involving the cyanmethemoglobin method—mean that the methods should not automatically be considered interchangeable. Interchangeability would require that the methods produce results sufficiently close that replacing one method with another would not affect clinical decisions. The observed mean differences of 0.2–0.25 g/dL, while small, could influence anaemia classification when haemoglobin values are close to diagnostic thresholds. Taken together, the complementary analyses support good agreement among the three methods within the study population, with the FBC analyser and portable Hb meter showing the closest comparability. These findings support their potential utility under the conditions evaluated but should not be interpreted as formal diagnostic validation or evidence of universal interchangeability. Implications for Clinical Practice The findings have practical relevance for haemoglobin testing in malaria‐endemic and resource‐limited settings, where the choice of measurement method may be influenced by laboratory infrastructure, turnaround time, and availability of trained personnel. The close agreement between the portable Hb meter and automated FBC analyser suggests that point‐of‐care haemoglobin measurement may provide a practical option where access to automated laboratory analysers is limited, consistent with previous evidence on point‐of‐care haemoglobin systems [2, 13]. However, the observed differences between methods indicate that results should not automatically be considered interchangeable across all clinical circumstances. The choice of method should therefore be guided by the intended application and available resources. Automated FBC analysers provide broader haematological information but require laboratory infrastructure, maintenance, and trained personnel, whereas portable Hb meters offer greater portability and operational simplicity. In situations where haemoglobin values are close to clinically important decision thresholds, awareness of method‐specific differences is particularly important, and confirmation using an established laboratory method may be appropriate. Thus, the present findings support the practical use of these methods within their respective testing environments but do not establish universal interchangeability or diagnostic equivalence. Limitations The study was conducted at a single healthcare facility using a consecutive convenience sampling approach, which may limit the generalisability of the findings to other populations and healthcare settings. The sample size was determined using a prevalence‐based formula rather than an agreement‐specific approach, and the final sample was slightly below the calculated target, which may have limited statistical power for some analyses. Although the overall sample size was adequate for the primary analyses, the relatively small number of participants in some age‐ and sex‐specific subgroups limited the precision of subgroup comparisons and increased the risk of Type I error (false positive findings) when interpreting subgroup‐specific differences. Important factors that may influence haemoglobin concentrations, including malaria parasite density, nutritional status, hydration status, menstrual status, and coexisting medical conditions, were not assessed. Although participants were classified according to WHO anaemia severity categories, the small number of participants with severe anaemia (≤ 2 participants per method) prevents a meaningful subgroup‐specific agreement analysis. Conducting such analyses with very small sample sizes would increase the risk of Type I error and produce unstable estimates that could not be reliably interpreted. In addition, analytical precision, including repeatability and coefficient of variation, was not formally evaluated, and long‐term instrument performance and inter‐laboratory reproducibility were not assessed. These limitations should be considered when interpreting the findings and when designing larger multicentre studies of haemoglobin method comparability. Conclusion The automated FBC analyser, portable Hb meter, and cyanmethemoglobin method demonstrated strong concordance and generally small mean differences for haemoglobin estimation among malaria‐infected patients. The closest agreement was observed between the automated FBC analyser and the portable Hb meter, while the cyanmethemoglobin method produced slightly higher haemoglobin concentrations on average. Bland–Altman analysis confirmed relatively small mean biases, although the limits of agreement indicate that differences between individual measurements may be greater than suggested by mean comparisons alone. These findings support the comparability of the three haemoglobin estimation methods within the study population but should not be interpreted as evidence of diagnostic accuracy or full analytical validation. Selection of a haemoglobin estimation method should consider the intended clinical or screening application, analytical characteristics, cost, and available infrastructure. Larger multicentre studies with broader populations and formal assessment of analytical precision and clinical performance are warranted to further evaluate the interchangeability of these methods. Author Contributions Mina Ofosu: conceptualisation, methodology, resources, supervision, writing – review and editing, project administration. Jessica Ntiamoah: investigation, data curation, formal analysis, writing – original draft preparation. Patricia Sam Barko: investigation, validation, writing – review and editing. Prince Elikem Torli: methodology, writing – original draft, data curation. All authors approved the final version of the manuscript and agree to be accountable for all aspects of the work. Funding The authors have nothing to report. Disclosure No artificial intelligence tools were used in the preparation of this manuscript. Ethics Statement Ethical approval for the study was obtained from the Institute of Research Innovation and Development Ethics Committee (Reference No. IRID/EC2025/HS0013). Institutional permission was obtained from the management of Kumasi South Hospital. The study was conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants prior to enrolment. Confidentiality and privacy were maintained throughout the study through anonymized identifiers, password‐protected data storage, and restricted access to study data. Conflicts of Interest The authors declare no conflicts of interest. Acknowledgments The authors wish to express their sincere gratitude to the management and staff of Kumasi South Hospital, particularly the Laboratory Department, for their cooperation and support during the data collection period. Special thanks go to all the study participants for their voluntary involvement. We also acknowledge the Department of Laboratory Technology, Faculty of Health Sciences, Kumasi Technical University, for providing the necessary resources and ethical approval. Finally, we appreciate the statistical guidance provided by colleagues and the technical assistance from laboratory technicians who helped with the sample analysis. Data Availability Statement All relevant data generated or analysed during this study are included in this manuscript. References Associated Data Data Availability Statement All relevant data generated or analysed during this study are included in this manuscript.