The American Red Cross recently declared a blood crisis for only the second time in its history.
This news immediately caught my attention. It’s hard to imagine hospitals running out of an adequate supply of blood for the purposes of transfusions. Blood is one of healthcare’s most valuable resources, but its supply is for the most part unpredictable.
I read that the American Red Cross announced its first national blood crisis around the time of the Covid-19 pandemic. This occurred in January 2022 and was driven by pandemic-related blood drive cancellations, a drop in the number of donors (as high as ten percent), and a sharp decline in school/college drives.
The most recent ‘emergency blood shortage’ is the result of national blood supplies falling by about 25 percent, in June. This is a result of distributions running around 3 500 units higher per week than expected amid summer trauma demand, and lower donations (from school breaks, busy schedules, illness, and so forth). Blood O Positive supplies have been the worst affected and are at critically low levels.
However, sometimes it takes a crisis to bring about strategic change.
The situation reminds me of what I have heard about the Chinese word for ‘crisis.’ It is written with two characters: 危 (wēi), means danger, peril, or risk; and 机 (jī), means critical point, pivotal moment, or opportunity.
‘Pivotal moment’ is exactly what this recent blood crisis has become, and it seems that healthcare professionals are seeing this as a turning point, in terms of altering how supplies are managed in the future.
Traditionally, blood donation campaigns have rallied communities to donate blood. However, experts are now looking beyond recruitment efforts to harvest this valuable resource. They believe that data could be the ‘powerful tool’ needed to transform hospital readiness.
Advances in artificial intelligence (AI), predictive analytics, and diagnostic technologies are creating new opportunities for hospitals to anticipate blood demand before shortages occur. This will help healthcare systems become more resilient during emergencies.Rather than simply reacting to a crisis, hospitals may soon be able to forecast one.
Hospitals have in traditionally managed blood inventories using historical trends, seasonal patterns, and current stock levels. Although these methods remain effective under normal circumstances, they can struggle during unexpected events such as pandemics, natural disasters, mass casualty incidents, or periods of declining donor participation.
Predictive analytics offers a different approach. By analysing thousands of data points simultaneously, machine learning models can estimate future blood requirements days or even weeks in advance.
These systems consider variables that humans cannot easily process together, including emergency department admissions, scheduled surgeries, trauma trends, infectious disease outbreaks, weather events, holiday periods, and regional population changes.
Instead of asking, “How much blood do we have today?” healthcare leaders can begin asking, “How much blood will we likely need next week?”
AI is also playing its role. AI excels at recognising complex patterns hidden within enormous datasets. In healthcare, AI is already being used to predict patient deterioration, identify sepsis earlier, optimise emergency department workflows, and improve radiology reporting. Applying similar technology to blood management is the next logical step.
Machine learning systems can continuously analyse hospital activity to estimate demand for different blood products, including red blood cells, plasma, platelets, and specialised blood types.
Unlike traditional forecasting methods, these algorithms improve over time. As more patient data become available, predictions become increasingly accurate, allowing hospitals to adapt purchasing, storage, and distribution strategies more effectively.
For regional blood services, predictive models could also identify where shortages are most likely to develop, allowing supplies to be redistributed before critical shortages occur.
Identifying Patients Who Are At Greatest Risk
One of the most promising applications of predictive analytics involves identifying patients who are likely to require transfusions before they deteriorate. Electronic health records already contain vast amounts of clinically useful information, including laboratory values, medical histories, medications, imaging findings, and physiological observations.
AI systems can analyse these variables to identify patients with an elevated likelihood of significant bleeding or transfusion needs.
For example, predictive models may recognise patterns associated with:major trauma, gastrointestinal bleeding, high-risk cardiac surgery, liver disease, cancer treatment, obstetric haemorrhage, and complex orthopaedic procedures
Earlier identification allows clinical teams to prepare blood products in advance, assemble appropriate specialists, and implement blood conservation strategies sooner.
This proactive approach may improve both efficiency and patient safety.
Similarly, using the findings of diagnostic imaging can help estimate the severity of trauma, identify patients at high risk of ongoing haemorrhage, and guide decisions regarding surgery versus image-guided intervention.
When imaging data are combined with laboratory values, vital signs, and clinical history, predictive models become substantially more powerful.
Supporting Patient Blood Management
Many healthcare organisations have adopted Patient Blood Management (PBM) programmes aimed at reducing unnecessary transfusions while maintaining patient safety. Predictive analytics could significantly strengthen these programmes.
Algorithms may identify patients with untreated anaemia weeks before elective surgery, allowing clinicians to optimise haemoglobin levels through iron therapy or other interventions before the operation takes place.
Similarly, predictive tools could recommend evidence-based transfusion thresholds based on individual patient characteristics rather than relying solely on broad clinical guidelines.
The goal is not to reduce transfusions indiscriminately, but to ensure that every blood product is used appropriately.
Then, blood shortages rarely affect a single hospital. Regional healthcare systems must coordinate blood distribution across multiple facilities, balancing local demand against available supplies.
Predictive analytics can assist at this broader level by combining data from hospitals, ambulance services, blood collection agencies, weather forecasts, public events, seasonal illness trends, and even transportation networks.
If a major sporting event, severe weather warning, or influenza outbreak is expected to increase emergency admissions in one region, blood inventories could be repositioned proactively.
The future of hospital readiness
The irony is that as advanced as we might become in terms of technology and information mapping and sharing; blood, a key component of life, remains the same as it has been since time immemorial. No algorithm can replace the body’s need for blood and the necessity of a willing donor.
However, better forecasting may ensure that donated blood reaches the right patients, at the right place, and at the right time.
Here are some other crises faced by healthcare systems in recent years, and examples of how these crises became ‘pivotal moments’ or turning points towards better healthcare:
Global Blood Deserts
In many low- and middle-income countries, ‘blood deserts’ exist where clinical need for blood cannot be met in at least 75 percent of cases. These chronic shortages continue to place trauma patients, women experiencing obstetric haemorrhage and surgical patients at significant risk.
Turning point: The crisis also highlighted the importance of building self-sufficient national blood systems rather than relying on replacement donors or emergency imports, leading to greater investment in sustainable blood collection and transfusion services.
2011 East Africa Drought and Famine
Severe drought led to a devastating food and water crisis across Somalia, Ethiopia and Kenya. The United Nations declared famine in parts of Somalia, while hundreds of thousands of people faced acute malnutrition, disease outbreaks and displacement following widespread crop failures and water shortages.
Turning point: The humanitarian response expanded beyond emergency food aid to include early warning systems for drought, climate monitoring, improved nutritional surveillance, resilient agricultural practices and stronger international coordination. Governments and aid organisations increasingly recognised the importance of anticipating food crises before famine conditions developed, leading to greater investment in disaster preparedness and climate resilience.
California/Western U.S. Water Crises (2010s to 2020s)
Prolonged drought, groundwater depletion and climate change resulted in severe water shortages across the western United States. Reservoirs reached historic lows, mandatory water restrictions were introduced, farmland was left uncultivated and emergency declarations became increasingly common.
Turning point: The crisis accelerated investment in water recycling, desalination, groundwater recharge projects, smart irrigation technologies and more sustainable water management policies. Governments also introduced stricter conservation measures and expanded climate adaptation planning, fundamentally changing how water resources are managed in drought-prone regions.
2021 to 2023 Global Gas Crisis
Triggered by the Russia-Ukraine war and major supply disruptions, natural gas and electricity prices rose sharply across Europe and other parts of the world. Governments faced the prospect of energy rationing, industrial shutdowns and challenges in maintaining reliable heating and electricity supplies during winter.
Turning point: Countries rapidly diversified their energy supplies by increasing imports of liquefied natural gas (LNG), expanding renewable energy projects, improving energy efficiency and strengthening strategic energy reserves.