The World Health Organization has prioritized the use of blood transfusions for treating Ebola, and three patients have received blood donated by an American doctor who survived the disease. But a lack of scientific study means that a lot is still unclear about the method and whether it could feasibly scale up to treat the thousands of Ebola patients in West Africa.

Using blood transfusion to treat a viral disease is “a simple concept which has been around for at least 100 years,” says Margaret Harris, a doctor and spokeswoman for the World Health Organization currently based in Sierra Leone. The idea is this: The immune system of a patient fighting a viral disease will produce antibodies that attack the virus in an attempt to clear it from the patient’s system. If successful, about two weeks later, a survivor’s bloodstream will contain these antibodies but no trace of the virus itself.

In transplanting a survivor’s blood to another patient battling the same disease, doctors hope that the antibodies in the donated blood — which have already successfully fought off the disease — will do the same again.

While the theory is straightforward and transfusions for other diseases have succeeded in the past, a whole lot remains unknown about the treatment.

Since transfusion is generally used as an emergency measure, there’s almost no scientific research about its use, especially when it comes to Ebola. This means that even the WHO still has a lot of questions about it: “Who does it work on, how does it work exactly, when doesn’t it work, what’s the right amount to give, all those things—that information really isn’t clear,” says Harris.

The first and last time transfusions were used to treat Ebola was in 1995, during the most recent outbreak of the disease, in what is now the Democratic Republic of the Congo. Bob Colebunders was part of a team of international doctors who were treating Ebola patients in Kikwit, the epicenter of the 1995 outbreak. When a nurse who was tending to Ebola patients began to exhibit the symptoms of the disease, she was given a transfusion of blood from a survivor. She got better, so the team tried the same treatment on seven others. Of the eight who received transfusions, only one died — that’s a much better survival rate than the 20% overall rate during the Kikwit outbreak.

Colebunders, who is now a professor of tropical diseases at the University of Antwerp in Belgium, says the results he saw in 1995 were hopeful but not entirely conclusive. “We don’t know exactly why it had an effect,” Colebunders says now of the transfusions. “It could be that it’s the antibodies, but … certainly it’s clear that other factors were playing a role.”

He says he is wary of using blood transfusion as a treatment for Ebola on a large scale, especially before further tests are conducted. He favors a gradual approach: The treatment could be rolled out “first on a relatively limited scale — well-documented — and then, depending on the results, we could consider scaling it up.”

Transfusion, Colebunders says, is a stopgap measure until a more robust solution is available. “It’s only in the meantime, because there is nothing, and this is readily available, that it should be done. Once there is better treatment, we should go to something else.”

And even if it were proven that blood transfusion is an effective way of treating Ebola, it brings a number of obstacles that could be particularly hard to navigate because of the poor quality of some West African nations’ health care systems.

There’s the issue of voluntary consent: “You have to find people who have survived and who understand what they’re doing, and willingly give that blood,” says Harris. And if transfusions take off as a treatment, she says, there could be a risk to survivors, whose blood would suddenly become a valuable resource. “They could be actually at threat of being seen as a commodity.”

Once willing volunteers are lined up, there are potential pitfalls associated with transferring the blood. It must be collected safely, using a sterile needle in a clean environment, and then screened for blood type (which must match that of the recipient) and the presence of other blood-borne viruses such as hepatitis and HIV. These are all basic requirements that almost any laboratory in the U.S. would meet, but such labs are not widely available in West Africa. “We’re talking about some countries that have very, very limited facilities,” says Harris.

Receiving blood also comes with risks. Reactions can include rapid changes in temperature, fever, or a drop in blood pressure, which is why most hospitals require a doctor to be present at the beginning of a transfusion, she says.

Despite these unknowns, WHO experts agreed in early September that blood transfusion “needs to be considered as a matter of priority” for treating Ebola. The first batches of survivors’ plasma to be used in transfusion treatments, if they’re successful, could be ready by the end of this year. Alternative treatments, such as experimental drugs such as Zmapp, have also not yet been rigorously tested on humans.

The most recent transfusion took place over the weekend. A nurse who was tending to Thomas Duncan, an Ebola patient at Texas Health Presbyterian Hospital, tested positive for Ebola. She received a transfusion of blood from Kevin Branly, an American missionary and doctor who overcame the disease in August. The nurse, Nina Pham, is in stable condition.

The method holds a lot of promise, and, if effective, could become a relatively low-cost treatment. At this point, there are enough Ebola survivors in West Africa that donations would not be difficult to come by. For now, until we know more about the treatment, it’s the least-bad option available. Faced with an outbreak that carries a 70% mortality rate, any treatment that’s been successful before is worth exploring.

• Google Has Received Nearly 145,000 Requests to Be ‘Forgotten’
• Obama Demands Tough Net-Neutrality Rules That Ban ‘Fast Lanes’
• Al-Qaida and ISIS Use Twitter Differently. Here’s How and Why.
• Google Is Gone, but ALEC Is Still Winning

This article originally published at National Journal here