Accuracy in measuring hemoglobin concentration using portable hemoglobinometer method

Introduction: Anemia is defined as a condition that arises due to as low concentration of hemoglobin in blood. Anemia affects a variety of population groups, cutting across different geographical boundaries and income groups, affecting all, including low-, middle-, and high-income countries. Anemia can cause cognitive disorder, affecting performance in both work and school and reduce the earning capacity of the affected individuals. In Indonesia, based on RISKESDAS data in 2007, it was found that 11.3% women, 12.2% men, and 12.8% children suffered anemia. And the WHO data collected between 1993 and 2005 globally revealed that 47.4% preschool-age children, 41.8% pregnant women, 30.2% non-pregnant women, 25.4% school-age children, 23.9% older adults, and 12.7% men suffer anemia. Anemia is a public health challenge in Indonesia and mostly affects preschool children and women of reproductive age. Numerous methods have been used and developed to measure hemoglobin concentration in human blood, but not all of the methods are easy and sensitive and effective in diagnosing anemia. The objective of this study is to know the extent of accuracy in measuring hemoglobin concentration using the well-known portable hemoglobinometer method. Research Design and Methods: This study was a diagnostic test and cross-sectional methods were used. The sampling technique used was total sampling. The total study sample comprised 48 participants. The study was conducted in the Public Elementary School 101747, Klumpang Kebun, Hamparan Perak District, between March and December 2016. Analysis was performed using a 2×2 table, in order to calculate a variety of item measures, including the following: sensitivity, specificity, positive predictive value, negative predictive value, positive likelihood ratio, negative likelihood ratio, and an ROC graph was made to determine the cut-off point of Hemoglobinometer. Results: The study results were as follows: sensitivity 0%, specificity 95.8%, positive predictive value 0%, negative predictive value 95.8%, positive likelihood ratio 0, negative likelihood ratio 1, and area under the curve of ROC graph 50%. Conclusion: The study findings encourage us to conclude that the results arrived at were specific, given the high specificity (95.8%) and high negative predictive value (95.8%). However, the sensitivity of this method as well as its positive predictive value were zero.