Demand and Capacity - A Comprehensive Guide
Demand and Capacity - A Comprehensive Guide
What is it and how can it help me?
A great deal of analysis has indicated that most waiting lists or backlog of works within the NHS are relatively stable, suggesting that variation in capacity and demand are the cause. This comprehensive guide helps to get to the hub of the problem of why waiting lists and backlogs form and what you can do about it. If you are new to this subject, 'Demand and Capacity - Basic Concepts' is worth reading first.
In order to make the most of patient flow through a healthcare system, it is necessary to address the entire patient process. You need to analyse and understand the capacity, demand, backlog and activity issues wherever there are waiting lists or backlog of works.
What is demand, capacity, activity and backlog?
Demand: All the requests / referrals coming in from all sources and how many resources they need (equipment time, staff time, room time) to be dealt with.
Capacity: Resources available to do work. For example, the number of pieces of equipment available multiplied by the hours of staff time available to run it.
Activity: All the work done. This does not necessarily reflect capacity or demand on a day to day basis. The activity or the work done on a Monday may be result of some of Monday's demand (i.e. emergency) and the previous week's demand. The capacity is the capacity available on the Monday but activity is often less than available capacity (ideally 80 per cent of available capacity)
Backlog: Previous demand that has not yet been dealt with, showing itself as a backlog of work or a waiting list. It's logical: if you don't deal with today's demand today, there will be a backlog for tomorrow.
Why is measuring demand, capacity, activity and backlog important?
The mismatch or variation in capacity and demand is one of the main reasons why waiting lists or backlog of works develop and waiting lists and times increase. An everyday example may help to illustrate this.
Jim checks his email everyday. He has lots of friends who he keeps in touch with by email. He's pretty reliable, so gets back to most people within three days or so and tends to spend about 15 minutes a day on this. He goes away for a week so when he gets back home he has a week's worth of unread emails or 7 days x 15 minutes' worth of email writing (1 hour 45 minutes). This is due to variation. His capacity to read and write email was different while he was away.
A bottleneck determines the pace at which the whole process can work. In this straightforward example, replying to emails is the bottleneck. Jim is the constraint, or in other words, the bit of kit or resource that is the cause of the bottleneck.
If we need to speed up throughput there is no point in doing this without focusing on the bottleneck or the cause of the bottleneck. Buying Jim another computer won't help him to clear his backlog.
Why is it important and useful to compare the four measures of demand, capacity, activity and backlog?
If you have a backlog of work in your team or you find that your team is constantly in catch up mode you will find that by measuring demand, capacity and backlog, you are likely to find opportunities to make things better.
Your team is likely to be one stage in the patient's pathway or journey. It is possible that another team's work is the bottleneck. For example, in order to meet the four hour target a lot of hospitals focused effort on the accident and emergency department. When they looked at the patients' pathway they found that the bottleneck was the availability of beds in general wards. The causes included lack of planning around discharge and frequency of decision making for discharge. These patients represented another (less visible) backlog of patients waiting to go home.
Measuring demand, capacity, activity and backlog enables capacity problems to be resolved at the appropriate point of the system. By clearly understanding these four measures and identifying the bottleneck and its constraint you can:
- Manage and plan work in all teams
- Increase throughput by reducing variation and /or matching variations in capacity and demand at the bottleneck
- Focus improvement effort in the place (bottleneck) where throughput can be increased
- Shift capacity to the bottleneck or manage demand to the bottleneck
- Protect the bottleneck and constraint as the implications of reduced capacity or mismatches in capacity and demand effect throughput along the whole pathway
- Plan work around the bottleneck.
In all processes there is a bottleneck with a constraint.
Backlogs aren't necessarily always bad as they can help to manage and plan capacity to demand. For example, based on this week's demand for hip and knee operations it is possible to plan capacity at the right level so the right amount of activity (operating lists) can take place in say three weeks time. However, generally unnecessary backlogs and backlogs with long delays aren't good. For example, in one hospital administrative staff developed the motto 'what a difference a day makes' about practising batching or piling work which is poor practice as it causes unnecessary delay.
When should demand, capacity, activity and backlog be measured?
Measurement of demand, capacity, activity and backlog is fundamental to understanding how well services are performing and to identify if demand and capacity are in balance. It is good management practice and should be routinely and systematically carried out. It is an essential tool to avoid waiting lists or backlog of works occurring or growing. When demand exceeds capacity, waiting lists or backlog of works will form. Unless action is taken to address the capacity at the bottleneck or reduce demand, the backlog will continue to grow.
Once demand, capacity, activity and backlog have been measured the data and patterns that emerge can be used to start predicting demand and managing capacity, activity and backlog at the bottleneck.
So why do waiting lists and waiting lists or backlog of works form?
Waiting lists or backlog of works occur where demand has not been dealt with and results in a backlog. The main reasons why waiting lists or backlog of works develop is the mismatch between variation in demand and capacity at specific times, because the right people or equipment are not always available to deal with the demand in a timely manner.
Every time the demand exceeds the capacity, a waiting list or backlog of work is formed.
However, every time the capacity exceeds demand, the extra capacity is lost or filled from the waiting lists or backlog of work.
When does it work best?
Reducing waiting times represents a shift in backlog of patients at every stage of the journey. Looking at, and matching capacity and demand is a key approach to removing some of the visible and hidden backlogs along the patient pathway.
Central to this is identifying the bottleneck and its constraint as this determines the number of patients that can be managed in the patient pathway and the speed at which the total backlog can be reduced.
The true bottleneck (the thing that really limits the whole rate at which a procedure/process is done) is hard to find in complex systems. Often we work with the immediate constraint (a constraint is limiting but is not the thing that dictates how many patients you see).
This analogy may help - in a stream, the water will flow fast in some areas, slow in others e.g. where narrows (a constraint) but almost stop in other areas (a bottle neck). In complex systems like health care, it is often hard to find the real bottleneck and not think that a constraint is the bottleneck.
How to use it
What you need to do.
2. Identify the steps or parts of the process where there are the longest delays for patients
3. Then map this part of the process or the patient journey in more detail so you really understand what is going on: to the level of what one person does, in one place, with one piece of equipment, at one time. We also recommend that you look at the process templates tool at this stage, it is a great tool for bringing this to life and is not difficult to do
4. Look carefully for the true constraint. The constraint is often a lack of availability of a specific skill or piece of equipment. Waiting lists or backlog of works tend to occur before the bottleneck in the patient journey, and clear after the patient has gone past the stage with the constraint. Process templates easily identify the constraint visually.
Why? "the consultant does not have time to see all her patients in clinic.
Why? She has to see everyone who attends (including first visit assessments and follow-up patients)
Why? It is what she has always done.
Now you have a good idea about where the constraint is. Building on your detailed map, the next stage is around measurement. This means measuring demand, capacity, backlog and activity.
Demand, capacity, activity and backlog need to be measured in the same units for the same period of time i.e. hourly, over a 24 hour period, weekly or monthly. It is not possible to compare two or more items unless they are measured in the same unit of time. It is important to compare the four measures on a single graph and the same measures must be used for each.
This excel template has examples of how to measure demand, capacity, activity and backlogs. The examples contained within the spreadsheet can be used to understand the variation in demand and capacity.
There should be a focus on the constraint is. So for example if the constraint is consultant's time then this is a priority. If it is room time is a constraint then that is your focus.
It is possible that the constraint is not stable though. So for example if you know that processing 10 patients = 7 hours of consultant time and 12 hours of nurse time and you have 2 nurses then you are ok but if you have one nurse and two consultants your constraint will shift.
So measure your capacity (availability of resources identified in the process template) over time.
How to measure demand:
Demand is all the requests and referrals coming in from all sources to the bottleneck step. To measure demand at the bottleneck step, multiply the number of patients referred by the time in minutes it takes to process a patient. Predicting demand can be difficult as historical activity data is frequently used to predict demand. Often activity data shows the number of patients we saw or the number of procedures done on a specific day. This tends to reflect the supply of services at that time rather then the true demand. Methods are required to measure the true demand. This must be undertaken from all demand sources. For instance in orthopaedics, physiotherapy referrals, pain clinic referrals, etc. must also be included as well as direct orthopaedic referrals in order to produce a more complete picture of the true demand.
4 referrals x a consultation time of 45 minutes each = 180 minutes (3 hours) of demand each day.
This illustration above shows how to measure demand, the excel template has an example that can be tailored to your needs.
How to measure capacity:
To measure capacity, multiply the number of pieces of equipment by the time in minutes available to the people with the necessary skills to use it.
Capacity is the resource available to do the work. This includes all equipment and the staff hours available to care for patients. For example:
2 treatment machines x 480 minutes (8 hours) of session time = 960 minutes (16 hours) of capacity each day. The following example shows how capacity can be measured, which can be found in the excel template.
Capacity can then be converted into the number of patients that could be seen.
So if a patient takes 20 minutes to process, then the capacity is 960/20 which equates to 48 patients.
Ensure that you measure all available capacity. Many people do lots of different things, so you need to make sure you measure any hidden capacity. For example, if you wanted to calculate the capacity for a pharmacist dispensing or preparing chemical substances, you would need to know all the activities they currently do and understand the proportion of their time devoted to each task.
Determining the true capacity of a system is often easier than predicting true demand. There are four steps:
1. Determine the overall supply of the service
How much capacity is available, for example minutes in outpatient clinic time?
2. Consider how supply changes over differing weeks and months (e.g. Staff leave)
It is important to understand actual capacity as opposed to potential capacity and then to look at ways at bringing the two closer together (e.g. co-ordinating consultant leave may result in fewer clinic cancellations)
3. Identify how the supply is provided over shorter time periods.
It should be deployed evenly against predicted demand because the closer that demand and supply can be matched, the better the system will run
4. Is the service that is provided what is really required to meet the patient's needs?
For instance the provision of radiology services for the management of DVT, when this type of service provision could be provided in a more efficient way
How to measure activity:
Activity is the actual work carried out by staff including the time spent with patients, carers and liaising with colleagues. To measure activity multiply the number of patients processed through the bottleneck by the time in minutes it took to process each patient.
100 patients processed x 20 minutes each = 2,000 minutes (33.5 hours) of work done each day.
Warning: Measures of activity numbers are misleading as this does not necessarily reflect demand or capacity:
- the activity in the month of June may well include demand carried over from May, April or even March
- staff may have not been fully utilised. They may have been kept waiting for the patient, specialised pieces of equipment or test results. The following example shows how activity can be measured, this can also be found in the excel template.
How to measure the backlog:
Backlog is the previous demand that has not yet been dealt with, showing itself as a waiting lists or backlog of work or waiting list. In a community setting, this includes all patients waiting to be assessed by the occupational therapist. This includes those on waiting lists and new patients in the system.
To measure the backlog multiply the number of patients waiting by the time in minutes it will take to process a patient through the bottleneck step.
100 patients on the waiting list x 20 minute treatment time each = 2,000 minutes (33.5 hours) backlog.
Take care when measuring the backlog and ensure that you don't count the same patient more than once. There may be patients who have been put on waiting lists at different parts of the same process, e.g. patients requiring radiotherapy treatment can be on waiting lists or backlog of works for their pre-treatment, planning, and simulation at the same time. Only count them in the earliest waiting lists or backlog of work to avoid recounting them at a later stage in the process. In the radiotherapy example given, it will be at the planning stage.
The following examples demonstrate how backlog can be calculated, for a working example, see the excel template.
Match capacity and demand on a daily basis
Once the backlog has been eliminated, the next aim is to ensure that demand and supply are in equilibrium. This requires that supply and demand are matched on a daily basis.
Matching of (service) capacity to (patient) demand is critical to ensure access to every stage of the patient journey. The capacity of the system needs to be flexible enough to cope with small changes in daily demand.
There are two key strategies:
look for ways of gaining capacity within the system
look for ways of increasing the flexibility of the capacity
Although patient demand can be predicted and care needs anticipated, unexpected situations may occur. Clinical teams can develop contingency plans to meet overflow demand. This might include adding more appointment slots as needed, adding clinicians as needed or making temporary changes in clinic flow and staff responsibilities.
Clinical teams should consider patient demand issues and plan capacity to meet them on a daily basis. This way, the team can avoid backlogs of work building up and patient access being restricted.
A further critical aspect is the requirement to schedule today's work today at every stage in the patient journey. When backlogs exist, we are typically matching last month's or yesterday's demand with today's supply. When the backlog is reduced we can match demand and supply in closer proximity to the origin of demand. This presents the opportunity to eliminate or reduce the work of triage and all the expediting of the work process.
Supply will be matched daily but for a period two weeks ahead (as the expected and acceptable delay for a service is two weeks). When an expected and acceptable delay is created, there is a requirement to preserve capacity to prepare for that expected demand.
For example, if we want to reduce the waiting time for an exercise treadmill test to three weeks, then all the capacity past the three week boundary must be held/preserved or frozen to protect it for that expected, anticipated and predicted demand. When our systems are filled or frozen, we lose the flexibility we need to respond quickly. Generally, what we see in that instance is the tendency to freeze even more of the capacity in order to be prepared for even more predicted and unpredicted contingencies. This uses up even more capacity.
What we often see in healthcare systems is the tendency that, in order to protect today's limited capacity, we push demand and work off into the future. In so doing, we perpetuate this approach. To match demand and capacity on a daily basis, in order to protect tomorrow's capacity, we need to pull demand into today.
An example of how measuring demand and capacity leads to improved patient access:
A GP practice situated in an area with some challenging health had an average waiting time to see a General Practitioner of 4.79 days. The practice reviewed its demand: the number of appointments requested on a daily basis and its capacity: the number of appointments available on a daily basis. This information allowed the practice to change the appointment system to match the demand and the introduction of different ways of accessing care eg telephone consultations and access to repeat prescriptions. Based on the demand and capacity information skill mix was introduced and a phlebotomist appointed thus ensuring that patients were seeing the most appropriate member of the health care team. The practice now has a waiting time of 0.32 days. This is an improvement of 93 per cent.
Case study / scenario / good practice / journal articles about practice
How does the case study link? E.g. this illustrates how to project manage
Applying service improvement methodologies to reduce waiting in radiography
Reduction of waiting times in diagnostic stage of patient's journey from 19 weeks to 2 weeks over a 5 month period by applying principles of capacity and demand management.
Re-designing the Vascular Surgery One Stop Clinic at Good Hope
Care Pathway Simulator was used to test proposed clinic booking schedules that were designed to minimise patient waiting, reduce the number of clinic overruns and maximise clinic capacity.
Matching demand and capacity in cardiology
A classic capacity and demand study. Straightforward approach to releasing capacity and matching available capacity to demand.
Think about - reducing demand:
Should we see all these patients? - implement protocols
Who is appropriate to see them? - provide alternatives
Can the patient pathway or the process at the bottleneck be streamlined? (Do we need to do all these steps?)
Reduce waiting lists - reduce the demands they create
Prevention, patient education
Think about - increasing capacity:
Use scheduling to find and ease the constraint
Reduce the number of appointment types to reduce complexity and carve out.
Work differently - flexible hours, weekends, pre-plan and cover annual leave, extended roles, etc
Bids for resources only when constraint is equipment or staff and working differently will not help
Other useful tools and techniques on the website: