The prevalence of anaemia depends on the definition: an example from the Aboriginal Birth Cohort Study

Objective: The cutoffs defining anaemia based on haemoglobin and haematocrit recommended by the World Health Organization (WHO) and the United States (US) are different. We compared the prevalences resulting from these definitions and explored the reasons for the variation observed. Design: A cross-sectional survey within a cohort study of children recruited at birth at Royal Darwin Hospital. Setting: Subjects were interviewed in their homes or other convenient location and had a blood sample drawn. Subjects: Five hundred and seventeen Australian Aboriginal children aged 9-13 years inclusive. Results: Depending on the criterion used, the prevalence of anaemia in the total group ranged between 6% (95% confidence interval: 4.1-8.4%) and 24.4% (20.7-28.3). Using the WHO criteria, girls aged 12-13 years were identified as a target group, having a two- to-six fold higher prevalence than the other groups. When compared to the US criteria, boys aged 12-13 years had the highest prevalence of anaemia, although this was not significant. Simulations show that the WHO cutoffs are based on inconsistent centiles of the age-sex haemoglobin and haematocrit distributions, and that this largely explains the discrepant results. Conclusions: This variability in definition could lead to inappropriate identification of target groups for intervention programmes.