This paper has a good overview, they say the rate of adverse events is around 1/1700. So if it reduces your risk of death from respiratory failure in the next week by almost any positive amount, it's worth doing.

This is only going to be used in hospitalized patients who are already developing respiratory failure (low blood oxygen levels), not in random people off the street with a cough.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3356109/#S1titl...