(P29-048-23) Concordance Between Hemoglobin and Hematocrit Among 1 - 5-Year-olds in NHANES Data From 1999 – 2020

Objectives: The aim of this study is to assess the concordance between hemoglobin and hematocrit to identify anemia among children.

Methods: In this study, we included children aged 1-5 years with both hemoglobin and hematocrit lab results. We examined the concordance of anemia identification between hemoglobin ( < 11 g/dl), and hematocrit ( < 33%) across nine cycles of the National Health and Nutrition Examination Survey (NHANES) (1999-2020) overall and by age (under and over 2 years), sex, race (non-Hispanic white, non-Hispanic black, Mexican American, other), and income level. Low income was considered 130% of the poverty threshold as defined using the poverty income ratio according to Department of Health and Human Services guidelines. The analysis accounted for the continuous, complex survey design. We applied inverse probability reweighting with each survey cycle’s sampling weight for all pooled analyses. Cohen’s weighted kappa statistic (ranging from -1.0 to +1.0) was used as a measure of concordance between hemoglobin and hematocrit to identify anemia.

Results: There were 9,420 children aged 1-5 in the NHANES survey from 1999 – 2020. Of these, 6,240 met the inclusion criteria for age and lab results. The mean age was 2.4 years, mean hemoglobin was 12.4 g/dl and mean hematocrit was 36.4%. Male and female children were evenly split (52% and 48%, respectively). Across age and race groups, sex, and income levels, hematocrit measurement consistently showed higher levels of anemia relative to hemoglobin. The kappa coefficient in the overall analytic sample was 0.67. Concordance was highest among children aged under two years (0.76) and non-Hispanic black (0.70). Concordance was lowest among Mexican Americans (0.57).

Conclusions: An overall kappa coefficient of 0.67 indicates substantial concordance but is below the preferable concordance threshold of 0.81 indicating interchangeability. The 0.67 concordance is problematic because hemoglobin and hematocrit are often used interchangeably in the clinical setting for the identification of anemia, as current guidance does not specify the use of one measure over the other. This may lead to patients with anemia being missed and patients without anemia receiving unnecessary testing and treatment.

Funding Sources: No funding was used for this analysis.