Jeremy Groves considers the facts behind the recent Independent article on critical care winter activity and NHS England’s response.

You may have seen the article about critical care in the  Independent a couple of weeks ago.  In it our president, Gary Masterson, outlined the pressure units were under this winter.  NHS England’s press office went into overdrive.  Tweeting via NHS Media they said “It’s simply not true that intensive care beds are full”.  So where do they get their information from?

I’ve always been intrigued by the complexity of defining a critical care bed.  A unit may have 20 beds, but it may have only 14 ‘open’.  So does that make it a 14 or 20 bed unit? What do we mean by ‘critical care bed’.  It can encompass both L2 (HDU) and L3 (ITU) patients and makes it difficult for a ‘mixed’ critical care unit to define its precise number of beds.  Our hypothetical 20 bed unit could, with 10, ideally 11, staff take twenty L2, but only ten L3 patients.  So what size is it?


I’ve always been intrigued by the complexity of defining a  critical care bed.

What do we mean by ‘open’? Well, this is often dependent on whether we can staff a bed.  Stress and sickness play a big role in that and staff tend to be ill when the rest of the population are.  After all, those working in critical care probably have greater exposure to seasonal pathogens than the rest of the population by virtue of them being cohorted in our work-space.  This is compounded by the stress of working in a high pressure, emotionally charged, and under resourced environment.

Throughout the year NHS Trusts are required to return both the total number, and the number of occupied, general critical care ‘beds’ to NHS England.  The ‘sitrep’. I assume these were the figures that NHS England referred to in their Twitter statement.  In the summer the data is recorded on the last Thursday of each month; during the winter it is done daily at midnight. The figures are a blunt tool: they don’t differentiate between L2 and L3, we don’t know what the Trusts report as ‘Critical Care’ beds, and we don’t know how they deal with open and closed beds.


we don’t know what the Trusts report as ‘Critical Care’ beds, and we don’t know how they deal with open and closed beds

The national bed state, NHS Pathways Directory of Services, doesn’t give a much better picture.  It could, but many units don’t keep it up to date, particularly if the units are busy and staffing is under pressure.  I haven’t been able to find accessible historical data either.

So we and our masters are left with the NHS England’s data to give an ‘objective’, and those are heavy inverted commas, insight into how busy we are.  So when NHS England tweeted that not all the beds were full, were they right?

The data is freely available from the NHS England website.  I’ve reproduced it below in the form of a chart showing the headline occupancy for England.


This is derived from the total number of critical care beds Trusts were reporting as open vs the number of occupied critical care beds.  As you can see we ran at a higher occupancy in January than December or November, but there were spare beds in the system.  NHS England was factually correct in their tweet.

It is interesting to see what happens over Christmas, presumably elective activity shutting down, and I’m not sure about that outlier in mid December.

Now these numbers are for England as a whole and don’t give a flavour of individual Trust’s pressures.  Lets drill down.

Drilling down begs the question as to an acceptable occupancy level for a critical care unit.  The answer to that question depends on your perspective but I think we would all agree that it should be full enough to ensure appropriate resource utilisation but not so full that patients are not properly cared for; an issue that has perhaps been brought into focus by the Bawa-Garba case.

The conclusion of Gutiez and Ramaiah, published in Critical Care (2), was that it should be at about 70%.  In discussing the implications of a unit that had a high occupancy they highlight the adverse consequences: refused admissions, early discharges both in and out of hours, increased levels of non-clinical transfers and surgical cancellations.

Now 70% is perhaps a luxury but I’ve included the percentage of Trusts reporting 70% occupancy and above in the next chart.  I’ve also put in the percentage reporting 85% and above and those reporting 100% occupancy.

JEremy Blog Graph

You can see the peak in early January when we really felt the pressure.  It is particularly evident from the red line, Trusts with 85% occupancy or above.  About 40% were at this level in mid November and it rose to between 60 and 70% of Trusts in early January.  That leaves capacity very tight.  Remember, 85% occupancy is 15% above the level at which quality really starts to suffer.  If we take 70% occupancy or below as being reasonable, only 20-35 percent of units were hitting this in Nouvember and 10% for the first half of January.

More worrying is those units really feeling the pinch.  Even in November 10% to 15% of Trusts were reporting no spare capacity.  In early January this was between 20% and 25%.  For those units it must have been really tough.

We can all argue about the semantics, but when a unit full, and its neighbours are full, that poses the system with a problem.  It’s no consolation to those units if they are in the north, and there are spare beds in the south west.

I don’t know what the answer is to winter pressures, but I do know that at a unit level, high occupancy, even if not every bed in the country is occupied, is bad for patients, bad for relatives, bad for staff and bad for the system.

breaker only

1) https://www.england.nhs.uk/statistics/statistical-work-areas/winter-daily-sitreps/winter-daily-sitrep-2017-18-data/

2) Tanaka Gutiez, R Ramaiah. Demand versus supply in intensive care: an ever-growing problem  Crit Care. 2014; 18(Suppl 1): P9. doi:  10.1186/cc13199