Five issues every bed plan should consider, and how to make sure plans get delivered

There is nothing quite as unifying for a local community than the loss of beds at the local hospital. So it is not surprising that research published by the King’s Fund, which showed that the number of hospital beds in England have more than halved in the past 30 years, caught the headlines last week.

But focusing on bed numbers alone does not tell the whole story. There are many reasons why fewer hospital beds may not be a bad thing – especially if it means people are (safely) back home quicker. But in the context of a financially squeezed NHS, with people waiting more than a year for routine operations and others sitting on trolley beds in hospital corridors, that is a hard message to get across.

Given these difficult constraints, it is critical to build up an evidence base for enabling bed capacity to be as best aligned as possible to demand. Fortunately the data exists to help answer this question.

In particular, there are five issues which need to be considered:

1. Beds mean cost…

To properly nurse for patients, a ward should have a good staffing ratio to each bed – this might be 8 patients to 1 nurse (depending on the sickness of the patients). So more beds means more staff and this means more cost – an average ward costs £2 million a year to run.

2. Are the beds in the right place?

One of the challenges many hospitals face is that they have “enough beds”, but they are in “the wrong place”. So there may be more than enough beds for patients with respiratory conditions, but not enough for post-operative patients. So you may be able to get into a hospital bed if you are sick from breathing difficulties, but at the same time have your operation cancelled due to a lack of beds…

Or patients end up on “the wrong ward”. This is not a great outcome as the type of care needed is very different and specialist doctors need to run from one side of the hospital to the other to see their patient (not a great use of time).

It also means that different departments rightly or wrongly see patients in “their beds” that should be elsewhere in the hospital. This leads to the inevitable suggestion of more ring-fencing beds for specific types of patients.

3. Fewer beds aren’t efficient

The challenge with a limited number of beds is varying occupancy rates – patients tend to arrive at unexpected times and stay for unpredictable lengths of stay (1). Interestingly, this means that a ward of 48 beds will usually hold more patients than two separate / ring-fenced wards of 24 beds.

From that perspective, there are economies of scale to hospital bed numbers. More beds can accommodate more patients than the same number of beds across different sites. In reality the type of nursing care varies across different types of patients (a diseconomy of scope). Balancing these two considerations is difficult and really depends on which patients are being nursed together, or how different nursing teams work together.

4. A bed at home is better than one in hospital

There are people in hospital beds that are not acutely unwell and could go home to their own bed. Often this results in better outcomes for the patient – being in a hospital bed too long is not good for recovering.

But there are still patients in hospital beds that do not need to be there. I’ve heard stories of elderly patients arriving with large suitcases and families refusing to have them back until their kitchen has been refurbished. I’ve also seen the consequence of pivotal members of staff taking holiday. Most notably, the discharge coordinator for stroke patients at one hospital who, as an avid skier, took leave every February – guess what happened to the length of stay for the patients on the stroke ward!

There are also other patients that could go home quicker if the way hospitals operate was more streamlined – e.g. with quicker scans.

5. Bed plans

The anecdotal evidence that hospital beds are not used effectively supports the view that hospital beds could be re-planned “around the patient” to align capacity to demand, and reduce costs. The problem is that this planning is often done at a very high level with aggregate data and a poor understanding of what is happening at the hospital level, let alone the ward or team level. This has resulted in a number of regions in England facing a lot of confusion when trying to turn top-down bed plans into a workable reality.

There is a big difference between high level bed plans and making these work at the ward level. This is due to the aggregation bias (mentioned above) – fewer beds are needed for the same number of patients when they are considered as one single group. Planning with averages simple does not work and it is why we tend to model bed occupancy at the hourly level (sometimes minute-by-minute).

Handling fluctuations: an example output from the Edge ward simulator shows how even a ward that runs well below capacity on average can see capacity constraints when faced with fluctuations in demand.

How best approach the problem

The reality is that the NHS does not have enough money to not challenge the use of hospital beds when they are not effectively used. Even if it did, the lack of nursing staff means staffing additional beds would not be possible. There are of course some areas where more beds may be needed, but there are also others where there are probably too many. It isn’t a good outcome to ignore these discrepancies as it leads to inequalities between different groups of people, and worse outcomes for patients.

But nor is there an optimal solution for bed numbers one that can be achieved in a single motion. Changing how beds are used in a hospital is as much about changing cultures in different teams as it is about planning where beds should be located – at the end of the day, most patients will only go home when the nursing team is ready for them to go home.

So changes to where beds are located has to be driven from the bottom up, albeit with significant challenge. Anecdote needs to be replaced with evidence, and data should be used as an enabler for change that results in a better alignment of demand and capacity.

This is something we have enjoyed helping a number of hospitals with over the past few months. One recently achieved 100% on its 4 hour A&E target off the back of some of this work.

Over the next four weeks, we plan to publish blogs on the topics covered in this blog in more detail. Let us know if there are specific aspects of interest and we will try to cover.

=========================================

(1) In reality, both of these factors can be predicted with varying levels of success depending on what data have been collected historically

#NHS #datascience #machinelearning #artificialintelligence #ml