What’s really happening with hospital bed numbers?

The King's Fund report is a useful addition to the current debate on beds in the NHS. But the problem with a focus on the number of available beds is that it distracts from thinking about how to use those beds more effectively. Or whether adding more physical beds would actually lead to significant improvement.

If we took the headlines and slogans in this debate seriously we would conclude that the NHS has been cutting beds for years (true) because of government cuts (well, the budget hasn't been cut, it just recently stopped growing and bed cuts were also happening during the Blair years when the NHS budget doubled) and this has led to catastrophic decline in emergency performance because the older population needs more beds (dubious as the population is also healthier). The headlines and many campaigners want us to conclude that we just need to open more beds and things like emergency performance would get better. Simples, as the Meerkats say.

But it isn't so simple and, even if the NHS opened more beds, things might not improve as much as expected. There are much better ways to improve the way we use beds that are too often neglected in the debate.

Our ability to think about the problems, like A&E crowding, caused by lack of beds (and, therefore, our ability to think of practical actions that might generate improvements, is greatly inhibited by the way we think about bed utilisation. It is currently at very high levels across NHS hospitals (often greater than 90% which makes finding free beds for new admissions pretty hard). But the metric we use is distractingly irrelevant and gets in the way of better thinking about how to use beds more effectively. The actual public utilisation metric is the number of occupied beds at midnight (often only measured on a single day of the week). This has, at best, a vague correlation with what hospitals actually need to know to get the best use from their beds.

To get an idea of why a focus on bed numbers isn't the most useful thing in finding ways to improve consider two hospitals. Let's keep things simple and imagine both deal only with emergencies. The first hospital has an average length of stay (LOS) of 5 days. but it has a very disciplined discharge process and all discharges happen before 12 noon. The second hospital has the same LOS but the typical discharge happens late in the afternoon. Both hospitals have the same average demand for emergency admissions. Both have the same bed occupancy at midnight of 90%.

Now consider the situation at noon in both hospitals. The need for emergency admissions peaks in the morning (often at 9 or 10am) and those patients will either wait in A&E or get admitted to a free bed. In both hospitals, about 20% of all the beds occupied at midnight will be freed up at some point in the day (a consequence of the LOS of 5 days). In the first hospital occupancy at midday will be 70% assuming no new patients have been admitted but in the second the occupancy will still be 90%. In the first hospital essentially every patient in the A&E needing an emergency bed will get one as soon as the A&E department need it. In the second some might get a bed that quickly, but many will not as there are only half as many free beds at midday as the expected number of admissions during the day (which is, on average, 20% of the total number of beds). Many patients will be stuck in the A&E for far more than 4hrs as the hospital struggles to find the free beds they need.

Both hospitals have the same LOS and the same occupancy at midnight. But one has enough free beds at midday to admit every emergency admission that day; the other spends most of the afternoon struggling to find free beds for half the patients waiting in A&E. The metrics the NHS reports are doing nothing to identify the core problem with bed availability.

On the metrics the NHS uses about beds the hospitals are indistinguishable. But one will report excellent A&E performance and the other will have many patients waiting more than 4hrs and the A&E will spend most of the day in a crowded and possibly unsafe state.

Guess which hospital is typical in the NHS? The second, not the first. Should the first thing we do be to somehow create more beds?

Both hospitals might welcome extra beds as this would help them on the days when things are busier than average. But the second hospital would need 20% more beds to achieve the same performance as the first assuming nothing else changes. But past experience would suggest that even those 20% more beds would not help as much as expected because something else would change. LOS would go up as the second hospital would feel less pressure to discharge its patients. Too often in reality discharges are driven not by the state of the patient but by the pressure to find new free beds. When there are more beds this pressure reduces and occupancy and LOS goes up.

The way hospitals use beds has an enormous effect on the number that are available at the point of the day when patients need them. If every hospital had a rigorously disciplined discharge process, the NHS would have as many free beds at the point of need as it would get if it restored every closed bed for the last 20 years (assuming nothing else changed in LOS or discharge discipline).

We are having a debate that focusses on how many beds there are as though that was the entire problem. But we are ignoring the more important issue of how beds are used. By focussing on the wrong things policy makers, the central NHS bodies that collect the statistics and campaigners are all having about as much effect on practice as a homeopathic poisoner. We need to do better if we really want to improve the performance of the system.