Every other essential medicine can be stockpiled. Blood cannot. The most-transfused product on Earth
begins to expire the moment it leaves a donor's arm — which means the global supply has to be rebuilt,
person by person, roughly six times a year, forever.
42days
The maximum a unit of red cells may be stored, refrigerated at 1–6 °C in an additive
solution, before it must be discarded. Platelets last five. There is no strategic reserve,
no synthetic substitute in routine use, and no way to buy time.
120.4M
Donations collected worldwide, latest year
36%
Of them from high-income countries — 15% of people
119
Of 195 countries short of modelled need
~3%
Of eligible Americans who donate in a year
01 · The clock
Everything in the bag is already dying
A donation is not one product but several, separated by centrifuge — and each fraction runs on its
own timer. Plot them on a logarithmic scale and the supply chain's whole problem appears at once: the most
urgently needed component has the shortest life.
Log-scaled time. Storage limits per the AABB Circular of Information for the Use of Human Blood and
Blood Components and FDA storage rules. Platelets alone are held at room temperature, under constant agitation,
which is exactly why they spoil fastest.
The consequence
5 days
A platelet unit collected on Monday is unusable by the weekend. Hospitals cannot buffer against a bad week —
they can only ask for more donors.
The consequence
6× / year
At 42 days maximum, a nation's entire red-cell inventory turns over roughly six times a year. The supply is
less a stockpile than a river.
The consequence
~6%
Share of blood units issued to Rwandan hospitals that expired before use, per an internal transfusion-service
report cited in The Lancet Global Health. Programme estimate, not a measured national rate.
02 · Who gives
The donation rate follows the money
The WHO collects donation data from 168 countries covering 97% of humanity. Sorted by national
income, the giving rate falls off a cliff — a sixfold gap between the richest and poorest groups, in a
resource that cannot be traded across the gap because it spoils.
WHO, Blood safety and availability fact sheet, updated 12 June 2026; data reported by 168 countries,
most for the year 2023. One hue, light to dark — no rainbow, no second axis.
Concentration
36 / 15
36% of the world's 120.4 million donations are collected in high-income countries, home to 15% of the world's
population. Because blood expires, that surplus cannot simply be shipped where it is short.
Processing
98% vs 52%
Share of donations separated into components — red cells, plasma, platelets — in high-income versus
low-income countries. Separation is what lets one donation treat three patients. Where it is missing,
each bag helps one.
Also reported: 10 countries are still unable to screen every donation for HIV, hepatitis B,
hepatitis C and syphilis; 119 countries import all of their plasma-derived medicines.
03 · The gap
A hundred million units that were never there
In 2019 a modelling study in The Lancet Haematology put global need against global supply for
the first time, country by country. The world collected enough in aggregate. It did not collect it in the
right places.
Schematic map — hand-simplified outlines, not survey-accurate, and not a country-level choropleth.
Shading marks the GBD regions in which every country's supply fell below modelled need.
Modelled figures, not measured shortfalls: Roberts et al., The Lancet
Haematology, 2019, for the year 2017.Aggregate supply nearly matches aggregate need, which is the trap: blood does not travel and does not
keep. Roberts et al., The Lancet Haematology, 2019 — modelling study.
The term of art
75%
A “blood desert” was formally defined in 2023 as a region where blood cannot be delivered in a timely,
affordable way in at least three-quarters of the cases that need it. Blood DESERT Coalition,
The Lancet Global Health.
Who it reaches
1,849
Units of unmet need per 100,000 people, worldwide, in the modelled estimate. In high-income systems blood
mostly supports surgery and cancer care; in low- and middle-income ones it is spent on childbirth
haemorrhage and childhood anaemia.
04 · The thinning queue
Rich countries have the blood. They are running out of donors.
American collections fell for a decade and then flattened — partly good news, because restrictive
transfusion guidelines genuinely cut demand. The worrying half of the story is who is still showing up.
National Blood Collection and Utilization Survey (NBCUS), biennial, 2013–2023; 2023 results published
in Transfusion, March 2025. Endpoints labelled directly. Collections fell 11.6% between 2013 and 2015
alone, then stabilised; 10.33 million units were transfused in 2023.
Twenty-year drift
−40%
Decline in the number of people donating through the American Red Cross over two decades — the lowest donor
count in 20 years when it declared an emergency shortage in January 2024.
The youngest cohort
−61%
Fall in donations from 16-to-18-year-olds, alongside a 32% fall among adults under 25, as school and campus
blood drives thinned out. Donations from those 65 and over rose 41%.
Participation
3%
About 62% of Americans are eligible to give blood. Roughly 3% actually do in a given year — and someone in
the U.S. needs blood about every two seconds.
05 · The price of a bag
Free to give, expensive to keep alive
The donation is voluntary and unpaid almost everywhere it is safest. Everything after it — testing,
separation, refrigeration, delivery, the cold chain that fails quietly — is not.
Supplemental findings of the 2021 NBCUS, Transfusion, 2023. Prices for all three components
rose again into 2023, with apheresis platelets up a further $65 per unit.
Rwanda, 2017–2019
12,733
Blood-product orders delivered by autonomous drone to 20 health facilities — 43% of them emergencies —
replacing road runs over difficult terrain.
Delivery time
−79 min
Mean drone delivery took 49.6 minutes, about 79 minutes faster than estimated road delivery. On the worst
routes the saving reached 211 minutes.
Waste
−67%
Reduction in blood-unit expirations at 12 months after drone delivery began — 7.1 fewer units binned per
month. Speed is, functionally, extra shelf life.
Nisingizwe et al., The Lancet Global Health, 2022. Interrupted time-series analysis of
administrative data from 20 Rwandan facilities.
Why it matters
The only manufacturing plant is people
Blood is the one medicine that no factory makes, no reserve holds, and no forecast can smooth. A
transfusion system is really a promise that enough strangers will show up this week — and the queue is getting
older and shorter in precisely the countries that can afford everything else.
01 GIVE ON A CLOCK
Whole blood can be given every 56 days, up to six times a year; platelets as often as every seven days.
A standing appointment is worth far more to a blood centre than a crisis-week surge.
02 GIVE WHAT IS SCARCE
Platelets and type O blood run out first. So does blood from under-represented donor groups, which matters
for patients with sickle cell disease who need closely matched units for life.
03 BRING SOMEONE UNDER 25
The steepest decline is in the youngest cohort — the donors who would otherwise give for the next fifty
years. Replacing them is the whole game.
Sources & notes
World Health Organization, “Blood safety and availability” fact sheet, updated 12 June 2026 — 120.4M
donations; 36% from high-income countries with 15% of population; donation rates of 28.9 / 18.2 / 8.5 / 4.5 per
1,000; component separation 98% / 94% / 81% / 52%; 168 reporting countries covering 97% of world population;
10 countries unable to screen all donations; 119 countries importing all plasma-derived medicines.
AABB, Circular of Information for the Use of Human Blood and Blood Components, and FDA blood
component storage rules — red cells 42 days at 1–6 °C in additive solution; whole blood 21–35 days;
apheresis platelets 5 days at 20–24 °C (7 with approved bacterial testing); plasma 12 months frozen at
≤ −18 °C; pooled cryoprecipitate 4 hours after thawing.
Roberts N. et al., “The global need and availability of blood products: a modelling study,”
The Lancet Haematology, 2019 — need 304,711,244 units vs supply 272,270,243 units for 2017; 119 of 195
countries short; 102,359,632 units of unmet need; 1,849 units per 100,000 population.
Modelled, not measured.
Blood DESERT Coalition, consensus statement, The Lancet Global Health, 2023 — definition of a blood
desert as a region where demand cannot be met in a timely, affordable way in at least 75% of cases.
McDavid A. et al., “Have we reached a new baseline for blood collection and transfusion in the United
States? National Blood Collection and Utilization Survey, 2023,” Transfusion, March 2025 — 11,586,000
red-cell units collected and 10,328,000 transfused in 2023; whole blood collections 9.57M; 2,618,000 platelet
units distributed; outdated red-cell units up 13.8% from 2021.
NBCUS earlier survey years (AABB / CDC summaries) — 14.24M units collected in 2013, 12.59M in 2015,
12,211,000 in 2017, 11,590,000 in 2019, 11,784,000 in 2021. 2013 and 2015 values are stated to three
significant figures. Trend-line values rounded.
Kracalik I. et al., “Supplemental findings of the 2021 National Blood Collection and Utilization
Survey,” Transfusion, 2023 — median hospital price per unit: leukoreduced red cells $214,
fresh frozen plasma $52, leukoreduced apheresis platelets $567. Griffin (2023 supplement) reports a further
$65 rise for apheresis platelets by 2023.
American Red Cross — emergency blood shortage declaration, January 2024: donors down about 40% over 20 years,
lowest in two decades; donations from 16–18-year-olds down 61% and under-25s down 32%; roughly 62% of the U.S.
population eligible and about 3% donating; someone needs blood every two seconds; whole-blood donation interval
of 56 days.
Nisingizwe M.P. et al., “Effect of unmanned aerial vehicle (drone) delivery on blood product delivery
time and wastage in Rwanda,” The Lancet Global Health, 2022 — 12,733 drone orders (43% emergency),
mean delivery 49.6 min, 79 min faster than road estimates (up to 211 min), 7.1 fewer monthly expirations
(a 67% reduction at 12 months). The ~6% expiry figure is from an unpublished Rwanda Biomedical Centre report
cited in that paper. Programme estimate.
Note on the map: outlines are hand-simplified schematic polygons, not a survey-accurate projection or a
country-level choropleth. Shading marks Global Burden of Disease regions in which every country fell below
modelled need.