Access to oxygen - and the ability of existing oxygen systems to meet demand - has been one of the major global challenges of responding to COVID-19. As a respiratory therapist, this is at the core of my clinical practice, and has worried me for years. Some thoughts:

Oxygen supplies in many health systems are inadequate for meeting daily hospital needs, never mind those created by COVID-19. There are many reasons for this, but probably this is largely because it's a big infrastructure investment and can require significant retrofitting.

There are essentially four different ways in which a patient can receive oxygen, each with their own benefits and limitations:
- Portable oxygen concentrators
- Oxygen cylinders (portable tanks, or cylinder banks)
- Liquid oxygen storage
- On-site oxygen generating facilities

All of these serve different functions. Some are only appropriate for individual patients requiring lower-flow oxygen (e.g. most oxygen concentrators can reliably deliver 3-5L/min for one patient - about enough for nasal prongs) whereas others are more large-scale solutions.

I've been in many hospitals where the hospital's oxygen system is a series of bedside oxygen concentrators - this is fine for routine, low-flow oxygen therapy. But COVID and critical care, generally, requires something different and these just don't cut it. Here's why:

Portable oxygen concentrators generally can't deliver high-flow oxygen beyond 10L/min. Ventilators require a high driving pressure to work that can't be supplied by most (perhaps all) portable concentrators; high-flow oxygen therapy requires exactly that: high flows.

Providing high-flow, pressurized oxygen supply throughout a hospital is an infrastructure issue - sure, you can bring in cylinders and can create cylinder banks, but those need to be replenished frequently. Depending on flow rate, a tank could last a few hours or a few minutes.

What hospitals really need, particularly for critical care, is a piped-in oxygen supply, delivered throughout the hospital at a high pressure, which essentially needs to come through either on-site production facilities or liquid/cryogenic storage.

The problem with piped-in oxygen systems is that they are a fairly significant infrastructure investment. I understand there are some 'turn-key' options available, but you still need to get oxygen throughout the hospital and to patients from the storage or generating facility.

With high oxygen demands from large numbers of patients in a hospital, the problem becomes more complex.
It's not just an oxygen supply problem, it's an oxygen delivery problem.
Whereas low-flow oxygen might run at 3-5L/min, high flow oxygen therapy can require 40-60L/min of gas.

Basically no hospital's medical gas infrastructure is designed to support the volume of gas delivery required to provide hundreds of patients with 40-60L/min of compressed gas. It's a fluidics problem: you can increase a hospital's gas supply, but not pipe sizes.

This is not a problem that's specific to any one country, and oxygen supply problems have occurred in many hospitals, including in Toronto yesterday: https://www.thestar.com/opinion/star-columnists/2021/04/29/michael-garron-hospital-scrambles-to-transfer-covid-19-patients-amid-worries-over-low-oxygen-supply.html

But the availability of oxygen in low- and middle-income countries where investments in health infrastructure are less robust has been a persistent problem before COVID-19, but acutely during the pandemic. Unfortunately, it's not an easily-solved problem.

Cylinders are being used as a way of providing oxygen in places where piped-in oxygen is unavailable or insufficient. This is exactly what cylinders are for (they're portable and versatile) but the logistics of this are huge. Cylinders run out and need to be refilled.

I would also add that cylinders are invaluable in places where electricity supply is variable or questionable - oxygen concentrators need some form of electricity to run, yet another infrastructure problem.

Trying to run a hospital's oxygen system off of a cylinder bank can be a full-time job. At high-flows or with high-demand - both are hallmarks of COVID care - cylinders can run out quickly and need to be replaced. That means bringing in new cylinders and refilling empty ones.

The point here is that providing oxygen is a logistically complex undertaking. Liquid oxygen systems and oxygen cylinders need refilling, and therefore access to a production and distribution system. Oxygen concentrators need stable electricity and have functional limitations.

Oxygen was recently added to the WHO Essential Medicines list, and it's essential for respiratory care, anesthesia, emergency medicine, etc. My hope is that the world is realizing that we need to make major investments in oxygen infrastructure across health systems and will act.