LAKELAND, Fla. -- The Food and Drug Administration's announcement on Aug. 23 announcement that it was issuing an emergency use authorization of convalescent plasma to treat COVID-19 was met with words of caution from some in the medical community. The World Health Organization warned that it's still an "experimental therapy, and the University of Oxford reminded the public that there's no hard proof it works.

"Controversy abounds, but what we have seen is that we have seen improvement in our patients at AdventHealth Tampa by using convalescent plasma," Dr. Douglas Ross, chief medical officer of AdventHealth Tampa, told members of Hillsborough County's emergency policy group at its Aug. 27 meeting.

Ross reported that some COVID-19 patients who received the treatment saw improved blood oxygenation, and three out of five recovered enough to be taken off ventilators. He said how early patients receive the plasma and how many antibodies it contains seem to determine how well it works.

"It is an exciting treatment option for patients that's showing a lot of promise," said Lakeland Regional Health Chief Medical Officer Dr. Tim Regan.

Regan is also president of Lakeland Regional Health Medical Center, which he said has treated nearly 250 patients using convalescent plasma.

"What we've seen in general is that there's an improvement in the amount of time people have to stay in the hospital, there's a reduction in mortality," Regan said.

Like AdventHealth and Lakeland Regional, BayCare is involved with the Mayo Clinic's ongoing study into convalescent plasma. About 700 patients there have received the therapy.

"Convalescent plasma is quite complicated, which is why you're seeing so many controversies in the media," said Dr. Paul Lewis, evidence based medicine director for BayCare.

The WHO's chief scientist called current data on convalescent plasma's effectiveness "very low quality". Doctors said one issue is the study doesn't involve a control group that's not receiving treatment.

"The challenge has been for them, as well as, frankly, for us, is to say, 'Okay, since there really was no non-intervention group, how can you distinguish how beneficial the plasma has been?" said Lewis. "Looking at our data, it suggests there's a slight benefit in the groups that have gotten plasma. Again, it's quite difficult to differentiate or distinguish, quite frankly."

Lewis said more randomized, controlled trials on the treatment are needed.