Understanding Capacity and Demand: A resource pack for healthcare professionals

Term

Meaning

80% rule

By monitoring the total demand on a service over a period of time it is possible to calculate the theoretical capacity that would give the process the flexibility to cope with the variation in demand and maintain a manageable queue.

theoretical capacity = minimum demand + [(maximum demand - minimum demand) * 80/100]

This figure is a guide; if you need no wait then you should be prepared to deal with 100% of your demand; if you are prepared to have a slightly longer wait for the service then this figure could be reduced.

Activity

The work done; the patients seen through the process; the actual capacity that is used to see patients. Aim to reduce lost activity due to patient DNAs and cancellations.

To measure activity multiply the number of patients seen by the time it took to process them.

Bottleneck

A bottleneck is where queues form; they slow down the whole process. There are two types of bottleneck: the process bottleneck which limits the rate of the process, i.e. the step that takes a significant amount of time; and the functional bottleneck which occurs when a resource is used by more than one process, e.g. porters. It is often relatively cheap or free to remove or reduce the effect of a bottleneck.

Capacity

All the resources required to do the work; including rooms, staff and equipment; theoretical capacity takes into account the availability of rooms, downtime for maintenance of equipment, staff holidays and study leave etc. Theoretical capacity is the maximum capacity that could be utilised if lists were run to time, fully booked, equipment was available etc. Actual capacity may be lower due to staff unavailability, equipment breakdown, adverse incidents etc.

To calculate theoretical capacity multiply the number of rooms and pieces of equipment by the time available to the people with the necessary skills to use them, e.g. a theatre may be physically available 24 hours a day, but if it is only staffed for 7 hours a day, 5 days a week for 42 weeks then this would allow a theoretical capacity of 1470 hours per year. Theoretical capacity should ideally be calculated using your demand (see the 80% rule above).

Carve out

Capacity that is retained for a particular type of patient or test ( e.g. urgent slots on a clinic template, slots set aside for knee operations). Carve out causes queues to lengthen as slots go unused or are misused (patients jumping the queue) which leads to increased variation. The flow of one group of patients is improved at one bottleneck to the detriment of others.

Churn

Churn starts when patients are waiting a long time and are not called in date order. More and more referrals may be flagged as urgent, and even those who are flagged as routine may jump to the top of the queue by phoning the secretary for an earlier appointment. Patients who don't chase the system fall further and further down the list, to the point that their health may deteriorate, resulting in an emergency admission, or they may get treated privately.

Constraint

Ultimately restricts the capacity of the system, e.g. the number of theatres, the number of surgeons. These are not easily increased, so you should aim to utilise them effectively.

DCAQ chart

A graph plotting the demand, capacity (theoretical and actual), activity and queue on a single graph in the same unit of measurement - they are all inter-related. If activity is greater than demand for a sustained period, you should see the queue fall. If capacity falls (for example over holiday periods) you can compare it to the demand which may not have such a significant reduction. This will lead to an increased queue.

It is also worth measuring the number of DNAs and cancellations as these will have an impact if not managed.

Demand

All the requests into the service from all sources. Remember the repeat patients who need to be seen albeit at a specified time. Use demand to set the theoretical capacity (see the 80% rule above).

To measure demand multiply the number of patients waiting by the time it will take to process them.

Flow

Improve the flow of patients along a pathway by reducing or removing bottlenecks, reducing hand-offs, removing non-value-added steps, reducing variation, managing capacity and demand, and increasing the probability of getting it right first time.

Mismatch

The mismatch between the variation in demand and the variation in capacity will usually result in a queue.

Non-value-added steps

Steps within a patient pathway or other direct process that could be deemed as being time-wasting and of no benefit to the patient or process, e.g. 6 steps to appoint a patient, 5 steps to vet a request. Aim to eliminate the non value-added steps to streamline the process and shorten the time taken.

Pareto analysis (similar to the Glenday Sieve analysis)

By looking at your service you will find that only 20% of procedures contribute to 80% of the workload; 7% contribute to 50%. Consider whether your service should be performing the very rare procedures. Concentrate on improving the flow for the high-volume procedures - this is where you will have the biggest impact on waiting times.

Pooling

Rather than each consultant maintaining a separate queue, consider pooling all straightforward referrals into one queue based on urgency and date of receipt. By pooling referrals the queue will stabilise and patients will wait equitably. This in turn will reduce churn and may additionally remove the need for an urgent category.

Process-mapping

By looking at the process from the patient perspective as a series of steps it is possible to identify what steps add value to the patient pathway and what simply add time and bureaucracy into the process. Remember to consult all services, staff and other stakeholders who may be affected by changes you make.

Queue

The demand which has not yet been dealt with (also known as backlog), i.e. the patients still waiting to be seen. Queues form whenever demand exceeds capacity, or, more typically, when demand and capacity are mismatched.

To measure the queue multiply the number of patients still waiting by the time it will take to process them.

Run chart

A line graph plotting the item you are interested in measuring as consecutive points ( e.g. time from referral to treatment; number of patients waiting). Helps to identify common and special cause variation.

Segmentation

Segmentation separates the process of care along the whole pathway for one group of patients in order to improve the overall flow of patients but not at the expense of other patients.

Silo thinking

Aiming to improve the processes that affect you rather than aiming to improve the entire patient pathway or process; failing to recognise that decisions and changes you make can affect other departments and services outwith your control.

Statistical ProcessControl ( SPC)

A more sophisticated form of a run chart which measures the limits and the average of the process; has set rules to identify trends, for example, when a process is out of control (or a change has had an effect), and gives more information about the work that may be required to bring a process into more acceptable limits, i.e. to reduce the common cause variation. It can also provide evidence that changes made to a process have actually improved it.

Treat in turn
(first in, first out)

Where waiting times are long the temptation is to create urgent queues.

However, all research has shown that this simply leads to longer waits for the non-urgent patients. Treating patients in turn will reduce the churn and reduce the waiting time for all patients in the longer term.

Utilisation

A measure of how much capacity is used; 100% is often seen as the ideal, but actually the process should allow for some flexibility as the pressure to fully utilise resources will lower staff morale and incite behavioural change.

To calculate utilisation divide actual activity into actual capacity and multiply by 100 e.g. 480 minutes of activity out of 540 staffed minutes gives 480/540 * 100 = 89% utilisation.

Variation

There are two types of variation - natural (common cause) and artificial (special cause). Examples of natural variation include admission and discharge patterns, waiting list initiatives that impact on the demand on a downstream service, staff holidays. Examples of special cause variation include lost patient notes, a patient changed address, a piece of equipment broke down. Aim to reduce the effects of the natural variation so that all patients receive the same standard of care and the flow through the service is maintained. Try not to focus too much resource on identifying the special cause variation.

By using specific rules for SPC (see above) you can identify common and special cause variation and determine whether your process is in control.