Iron deficiency is a pressing global health problem and the subject of a 2023 World Health Organisation call to action. It is the leading cause of anaemia worldwide, which in turn is the third most common cause of years lived with disability globally. The World Bank estimates that every dollar invested in anaemia returns $12 by improving school completion, raising adult wages, increasing escape from poverty and increasing gross domestic product. Despite this, progress towards tackling anaemia is slow. One of the barriers to tackling iron deficiency is the poor sensitivity of current screening tests and the complexity of the diagnostic pathway, increasing the risk of missed diagnosis.
Iron deficiency is associated with a characteristic pattern on the full blood count, making it amenable to a machine learning solution. The Full Blood Count is also widely available, reducing the barriers to implementing a test based on it. We are developing machine learning models to detect iron deficiency signatures in the full blood count with the aim of simplifying and improving the diagnosis of iron deficiency.