19 February 2026

From Routine to Revolutionary: How BloodCounts! Is Set to Transform Global Health

By Indrani Kashyap – Associate Director, Regional Communications (Asia, Africa, Lat Am), Global Engagement and Communications Office, Jhpiego

Imagine a world where the most common and routine blood test, the Complete Blood Count (CBC), could do far more than flag an infection. What if this test, performed about 3.6 billion times every year, could predict disease outbreaks, diagnose health conditions early, and save millions of lives at minimal cost? That is the vision driving BloodCounts!, an international consortium that began during the COVID-19 pandemic to help detect outbreaks early and is now set to enable early diagnosis of non-infectious conditions across three continents.

Demystifying the potential of the CBC

The CBC is the most commonly performed medical test for decision-making in clinical and community settings worldwide. During the test, a blood sample is collected and placed in an automated machine called a hematology analyzer (e.g., Sysmex XN-2000), which uses a laser to measure hundreds of cell characteristics. In current medical practice, only a limited number of these measurements are considered, while the raw laser measurements are discarded. That is where BloodCounts! saw an opportunity to mine rich, unused data and use generative artificial intelligence models, advancing biomedical research to predict and diagnose pervasive health conditions.

“The machine actually measures thousands of characteristics for every cell, but we’ve only been distilling it down to 18 numbers,” explains Dr. Joseph Taylor, a hematology consultant at Barts Health NHS Trust, a consortium partner. “BloodCounts! goes back to the raw measurements, revealing signatures that have always been there.”

The Trinity Challenge: Catalyzing individual efforts into a global consortium

In 2021, BloodCounts! won the prestigious Trinity Challenge Prize for its AI-driven approach to outbreak
detection. The award galvanized individual efforts by Cambridge mathematicians, such as Michael Roberts, Professor of Applied AI in the Department of Applied Mathematics and Theoretical Physics and the Department of Medicine at the University of Cambridge. “I’d say winning the Trinity Challenge put a rocket trip under the idea that we can develop data-driven approaches for the CBC. Before that, we’d done a lot of stuff on top of our regular day jobs. But this prize allowed us to recruit an entire team and gain the credibility to attract many people from around the world to join this consortium. It allowed us to really invest in the governance as well” says Michael, a founding member of BloodCounts! and a Chair of the consortium’s steering committee.

Members of the consortium include Amsterdam University Medical Centers (Netherlands), Apollo Hospitals (India, Barts Health NHS Trust (UK), Cambridge University Hospitals NHS Foundation Trust (UK), Digital Environment Research Institute (UK), Health Services Authority (Singapore), KU Leuven (Belgium), MRC Unit The Gambia at the London School of Hygiene and Tropical Medicine (The Gambia), Maastricht University Medical Centre (The Netherlands), NHS Blood and Transplant (UK), St George’s University Hospitals NHS Foundation Trust (UK), UZ Leuven (Belgium), University College London (UK), University College London Hospitals NHS Foundation Trust (UK), University of Cambridge (UK), VU Amsterdam (The Netherlands), West African Centre for Cell Biology of Infectious Pathogens (Ghana), and Zuyderland Medical Centre (The Netherlands).

Since its initial work in detecting public health emergencies like COVID-19 without any prior knowledge of the underlying virus, BloodCounts! has today vastly expanded its scope to examine how CBC tests can be used to identify patients with heart and brain diseases earlier, how the rich CBC data can be used to improve the diagnosis of blood disorders and early detection and response to sepsis; whether the algorithms can be applied to other infectious diseases, such as malaria, or detect high risk pregnancies early, or identify individuals at high risk of renal cell carcinoma. The consortium is also implementing a test for iron deficiency using the rich CBC data.

Anemia: A global priority

Anemia is a major public health concern that has a disproportionate impact on young girls and childbearing women in low and middle-income countries. Anemia caused 50 million years of healthy life lost due to disability in 2019.

Iron deficiency is the leading cause of anemia worldwide, but not the only cause. Dr. Joseph Taylor explains, “Many people make a false equivalence—they think anemia equals iron and iron equals anemia. But that is not true. That’s a misunderstanding. Anemia is not a diagnosis. Anemia is more like a symptom. It just tells you that you don’t have enough hemoglobin or red blood cells to carry your oxygen, but it doesn’t tell you why.” One of the barriers to tackling iron deficiency is the limited sensitivity of current screening tests and the complexity of the diagnostic pathway, which increases the risk of missed diagnosis. BloodCounts! aims to change that. By analyzing the full spectrum of CBC data, its models can distinguish between iron deficiency, thalassemia, sickle cell disease, and other causes—at the first point of care. “If we can tell a clinician, ‘This looks like iron deficiency’ or ‘This does not,’ that’s transformative,” says Taylor. “It means the first treatment is more likely to be right, which matters enormously when healthcare access is limited.”

The hematology analyzer and the need for robust data

Central to BloodCounts! is the hematology analyzer, in this case, Sysmex. Norbert de Wit, a Laboratory
Specialist in Clinical Chemistry, works at the Maastricht University Medical Centre, a consortium partner of BloodCounts! He specializes in hemato-oncology and hemocytometry and explains the importance of this machine. “It’s the machine that measures the complete blood count that we use to train the foundation model. So, for every patient that we measure, we get a file. It’s what we call raw CBC data. And that file and data can be used to train the foundational models being developed,” he explains and adds, “You need a lot of data. And you need data from all over the world to make this model applicable to everybody—data from people of different ancestries. And that’s what the consortium contributes.”

The consortium partnered with Umberto D’Alessandro, Professor of Epidemiology at the MRC Unit in Banjul, The Gambia. Affiliated with the London School of Hygiene and Tropical Medicine, the Unit provides rich CBC data to BloodCounts! as part of the consortium. “The African population is the most diverse in the world in terms of genetic variation. We are excited to include more African countries, more centers in this initiative,” says Umberto.

Protecting Client Privacy While Saving Lives

With data coming in from three continents, BloodCounts! took special care to ensure patient privacy. Instead of sending sensitive data to one central location, BloodCounts! uses federated learning – a breakthrough approach that lets hospitals collaborate without sharing patient data. Each hospital trains the AI model on its own rich CBC data and securely communicates only the encrypted model back. These updates are combined to make the global model smarter and then shared back with all participants. This means the system learns from millions of blood tests globally while keeping personal data safe. “No one’s done it before, using a federated learning infrastructure in real-world hospitals at scale across three continents. We train our models in an ethically compliant manner using real patient data, whilst hospitals always retain control of their sensitive data”, states Professor Roberts.

A hopeful future

From its roots during a global crisis, BloodCounts! is now shaping a future in which a routine blood test could transform healthcare worldwide. By unlocking previously unused signatures in the CBC and combining them with cutting-edge AI, the consortium is building the ability to diagnose health conditions earlier, faster, and more equitably. Excited about what the future holds, Professor Michael Roberts concludes, “What excites me most is the translational side—that we’re going to get real digital diagnostics into these CBC machines, changing practice globally. All of these tools we’re developing have a route to market, a route to getting into clinics and making a difference to people around the world.”