This was the main reason why temporary Nightingale hospitals, built in the first Covid wave at a cost of more than £500m, only ever treated a handful of patients. It was possible to build the critical care infrastructure almost overnight, but quite another thing to find trained medics to work in them.

To help plug these staff shortages in ICU, volunteers were frequently brought in from other parts of the hospital, often with no experience of intensive care medicine or of dealing with that level of trauma and death.

“They were being exposed to things which they wouldn’t necessarily be [exposed to] in their normal jobs, people deteriorating and dying in front of them, the emotional distress of that,” said Dr Ganesh Suntharalingam, an ICU doctor and former president of the Intensive Care Society.

Another hospital doctor said he felt some junior members of staff were “thrown in at the deep end” with little training and no choice about where they were sent.

The inquiry heard that all this “inevitably” had an impact on some of the sickest patients.

At no point did the NHS have to impose a formal ‘national triage’, where someone was refused treatment because they could not get a hospital bed.

But using that as measure of health system collapse may be too simplistic anyway.

Prof Summers said it would be mistake to think of “catastrophic failure” as a switch that goes “from everything being okay to everything not being okay the next second.”

“It is in the dilution of a million and one tiny little things, particularly in intensive care.”

She said when the system becomes so overstretched it feels like “we are failing our patients” and not providing the care “that we would want for our own families”.

New research suggests those hospital units under the greatest pressure also saw the highest mortality rates for both Covid and non-Covid cases., external

Difficult decisions were having to be made about which of the sickest patients to move up to intensive care.

Those Covid patients who needed CPAP, a form of pressurised oxygen support, rather than a ventilator, often had to be cared for in general wards instead, where staff may have been less used to the technology.

One anonymous ICU doctor in Wales, external said: “We didn’t have enough space to ‘give people a go’ who had a very remote chance of getting better. If we had had more capacity, we might have been in a position to try.”

The inquiry was also told that at least one NHS trust was under so much pressure it implemented a blanket “do-not-resuscitate order” at the height of the pandemic. If a patient went into cardiac arrest or stopped breathing, it would mean they should not be given chest compressions or defibrillation to try to save their life.

In normal times, that difficult decision should only be made after an individual clinical assessment, external, and a discussion with the patient or their family.

But Prof Jonathan Wyllie, ex-president of the Resuscitation Council, said he knew of one unnamed trust that put in place a blanket order based instead on age, condition and disability.

Groups representing bereaved families said they were horrified, adding it was “irrefutable evidence the NHS was overwhelmed”.