NHS gooroo blog

Archive for June, 2010

When modelling and planning to achieve referral-to-treatment waiting time targets, should we model the entire wait in one lump, or split out the outpatient, diagnostic and inpatient/daycase elements?

Modelling in one lump has its attractions. It matches the waiting time target, and it greatly reduces the amount of modelling we need to do. So under what circumstances could we do it?

Firstly, the same referral-to-treatment pathway would need to be followed by nearly all the patients. If some need a diagnostic and some don’t, or if some attend outpatients at a different hospital, or if some go round a follow-up loop before moving on to the next stage, then modelling in one lump won’t work at all well.

Secondly, we must be running a pure modelling exercise that is divorced from management decision-making. The results of modelling in one lump will tell us nothing about admitted patient capacity, or tariff costs, or the waiting times that must be achieved in outpatients. So modelling in one lump greatly reduces the usefulness of the results.

Thirdly, we would need to model correctly the different reasons why some referrals do not end up as admissions. Some miss their appointments, some are discharged at outpatients or after their diagnostic, some are removed while waiting for their admission. Each of these dropouts will have a different impact on scheduling and waiting times.

So if modelling in one lump is such a bad idea, how do we resolve the fact that our referral-to-treatment target is one number, but we have two or three stages to model along the pathway?

It is certainly a good idea to split the overall target up, and apply each component as the target for each stage along the pathway. For instance, if we have an 18-week target for an outpatient-diagnostic-inpatient pathway, we might allocate 6 weeks for each stage. Then we can plan the capacity and booking templates for each stage separately. (If we tried to leave the target undefined for each stage, and manage each individual patient ad hoc within the overall target instead, there is the danger that outpatient waits would drift upwards in an unplanned way, leaving diagnostics and inpatients with all the burden of achieving the target.)

Sometimes we may lack accurate data at the diagnostic stage, in which case we can allow a reasonable time for it (say, 6 weeks), and then divide the remaining 12 weeks between outpatients and inpatients. Gooroo Planner can work out the optimum split, but usually it is better to allow more time for the outpatient stage because reducing outpatient lists causes an immediate knock-on to the inpatient list.

All in all, it is usually better to model each step of the pathway separately.

In the first part of this posting we looked at austerity and waiting times from the hospital’s point of view. Here, we will look at it from the commissioner’s perspective (which is similar to that of senior hospital managers who are deciding how to allocate resources).

Waiting times – the commissioner’s perspective

As commissioners, we are often asked for money; we have to know when to say “yes” and when to say “no”. But this can be very difficult, especially when we cannot get hold of all the information we need to make a proper informed judgement.

Bids for extra resources to prevent waiting time breaches are a case in point. It is easy for a provider to show us that they are breaching the target. What is harder to establish is whether they could solve the problem themselves without extra resources.

For instance, imagine that our local hospital has a service that is struggling with the waiting time target, and they send in a request to increase contracted activity. The number of patients waiting longer than the target has been rising. So has the total number of patients waiting. They say they have done everything they can to pool activity, reduce demand, reduce inappropriate referrals, and so on, and actually we believe them. They say the only option left is to do extra activity to reduce the size of the waiting list. What they don’t tell us, though, is the detail about how they are scheduling patients.

We know that good scheduling can often solve a waiting time problem without any extra resources being needed. How can we tell whether this request for extra activity is justified, or whether better scheduling would do the trick?

There are two ways of doing this: examine the data, or examine the processes.

Let’s look at the data approach first. We want to work out whether this service would have a waiting time problem if it were well-managed. This turns out to be quite a complicated calculation, and in practice this would be performed by a good planning tool*. The result of the calculation will be the achievable waiting time target, and if this is less than the target in force then better scheduling should solve the problem.

As an alternative, we could use a good simulator to investigate whether good booking practices would bring the service back within target.

Or we could use a process-based approach. We want to assure ourselves that capacity is being correctly partitioned between urgent and non-urgent bookings, and that best practice is being followed when issuing patient appointments.

If the service is using a good automated booking assistant then they can demonstrate good practice using the system’s compliance log. Otherwise we will need to examine their processes in detail and check that casemix urgency has been properly evaluated, that the booking template is correctly designed, and that the rules used to allocate appointments are correct and are being followed accurately.

It all sounds quite involved, so is it worth it? When a “waiting list initiative” for inpatients can easily run to six figures, the answer must surely be “yes”. If that kind of money is spent in the wrong place, then it isn’t available to spend where it is really needed. That could make the difference between success and failure when austerity clashes with waiting time targets.

*e.g. the rule of thumb for a list with fully-booked, well-managed scheduling, and a target that 90% of activity must be admitted within target, is:

Achievable waiting time target in weeks =  1.25 * ( number waiting / average additions per week – ½ * maximum weeks waited by any urgent * proportion of urgents in casemix ) / (1 – proportion of urgents in casemix)+ 16 * (proportion of removals) – 3.5

Different formulae apply for partial booking, snapshot based targets, and very high proportions required to meet the target; a ready-reckoner is included in Gooroo SimActive’s Booking Rules calculator. Training packages, scheduling assistance, booking simulation, planning software, and research papers can all be found on the main Gooroo website.

Austerity, cuts, efficiency savings… whatever you call them, they’re coming. Life is going to be tough. Keep your head down, and you might just make it through.

Is that really the best approach, though?

When an age of austerity is galloping over the horizon, you want to be ready for it. What is austerity going to do in your part of the NHS? How can you make sure that if things start to fail under pressure, they “fail safe”?

In this two-part posting we will look at the effects of austerity on waiting times, and outline a response to it both from the hospital’s point of view and from the commissioner’s. Let’s take the hospital’s perspective first.

Waiting times – the hospital’s view

In an age of austerity, are GPs going to refer fewer patients? Perhaps they should, if commissioners can’t afford to pay for the extra activity. But past experience is not encouraging: referrals tend to rise year on year, regardless of the ability of commissioners to pay, and indeed commissioners have no powers to prevent it. So we need to be prepared for referrals to continue rising.

What about activity, will that carry on rising too? It might, modestly. But when it comes to waiting times, the absolute level of activity doesn’t matter: the question is whether activity can keep up with demand or not. In a time of austerity, we must assume not.

If activity falls behind demand, then waiting lists get longer. What then? Is it inevitable that waiting times will increase? If the shortfall is at all significant, then yes.

So are we all doomed? Will demand outstrip activity, waiting lists and waiting times go up, waiting time targets be breached, and everybody get the sack?

Not necessarily. That simple tale of woe conceals a lot of assumptions, and once we unpick those assumptions we can start to find a way through.

Let’s start with waiting lists and waiting times. Yes, if our waiting list goes up significantly then the average waiting time will go up too. But our target has nothing to do with average waiting times, it’s about the longer-waiting patients on the list, and long waits are affected by the way we schedule patients as well as the number waiting.

So step one is to make sure we are scheduling patients in the best way, so that long waits are minimised for the size of waiting list we have. In most cases this should provide quite a lot of headroom for our waiting list to grow, without breaching our waiting time targets.

The next thing we need to do is make sure other parts of our hospital are scheduling correctly too. Why? Because if we don’t, then other services might develop waiting time problems that should be solved by better scheduling, blame them instead on the size of the waiting list, and then consume scarce resources by laying on a waiting list intiative. Those resources might have been better spent elsewhere, perhaps on our own service, and we need to prevent this waste from happening.

What if we have done all that, and are scheduling correctly, but we still develop a waiting time problem? Now the problem must be the size of the waiting list, and there are several things we can try:

1. look for ways of “pooling” work between different parts of our service – if we are about to book a patient into an appointment slot where they will breach the waiting time target, see if an earlier slot is available with a suitable alternative clinician?
2. check that our data about the list is “clean”: e.g. no duplicates, no appointments still booked after the patient was removed, no follow-ups booked as first appointments;
3. look for cost-neutral ways to increase activity: e.g. reduce follow-up appointments to free up capacity, ensure patients see the right consultant first time;
4. look for ways to reduce demand: e.g. helping GPs to manage patients in primary care settings, working with GPs to reduce inappropriate referrals;
5. finally, seek extra resources to lay on extra activity.

If we do end up seeking extra resources, then we will have to make our case well. We need to show that our scheduling is good, either by demonstrating that we are following the right booking processes or by showing that the shape of our waiting list is appropriate to our casemix. We also need to show that we have exhausted possibilities 1-4 in the list above. Ideally, we should also show that some other service in our hospital has slack that we could take advantage of, in the form of poor scheduling or poorly-used capacity.

What if we have done all of these things, but we still have a waiting time problem? What’s the worst that can happen: a breach of our waiting time targets?

Actually, there is something worse than breaching the target when our waiting list is too big, and that is not breaching it. That would mean we were prioritising long-waiting routine patients at the expense of urgent patients, with the result that urgent patients waited too long and were put at clinical risk. Unfortunately this does happen in real life, especially when managers cannot distinguish between services that have genuine pressures and those that could solve their own problems.

Our understanding of waiting times and scheduling is now advanced enough that there should be no excuse for this any more. Resources should not be thrown at services whose problem is poor scheduling, nor withheld because of a misperception that the pressures are not genuine. This is the same issue that commissioners face, and in the second part of this post we will look at the problem from their perspective.

It isn’t a trick question. Surely any major public service should have high-level objectives? Especially one that has been around for half a century and spends £100 billion a year.

But does it?

If you go to the NHS website you will find “core principles”, not objectives as such. When the NHS was founded in 1948 they were:

• That it meet the needs of everyone
• That it be free at the point of delivery
• That it be based on clinical need, not ability to pay

In 2000 these were extended by the Labour Government of the day to:

• The NHS will provide a comprehensive range of services
• The NHS will shape its services around the needs and preferences of individual patients, their families and their carers
• The NHS will respond to the different needs of different populations
• The NHS will improve the quality of services and minimise errors
• The NHS will support and value its staff
• Public funds for healthcare will be devoted solely to NHS patients
• The NHS will work with others to ensure a seamless service for patients
• The NHS will help to keep people healthy and reduce health inequalities
• The NHS will respect the confidentiality of individual patients and provide open access to information about services, treatment and performance

We can tell that these are principles, rather than objectives, by asking ourselves a simple question: could we tell if the NHS failed to achieve them? The answer is: not easily.

In 2010 the same Government published the NHS Constitution, which contains “Seven key principles”, “underpinned by core NHS values”. Although there are overlaps, they are different from the core principles listed above. The Constitution principles were:

1. The NHS provides a comprehensive service, available to all
2. Access to NHS services is based on clinical need, not an individual’s ability to pay
3. The NHS aspires to the highest standards of excellence and professionalism
4. NHS services must reflect the needs and preferences of patients, their families and their carers
5. The NHS works across organisational boundaries and in partnership with other organisations in the interest of patients, local communities and the wider population
6. The NHS is committed to providing best value for taxpayers’ money and the most effective, fair and sustainable use of finite resources
7. The NHS is accountable to the public, communities and patients that it serves

… and the core NHS values were:

• Respect and dignity
• Commitment to quality of care
• Compassion
• Improving lives
• Working together for patients
• Everyone counts

Hmm. We aren’t any closer to something crunchy. What about the more operational guidance produced by the NHS? Does the annual operating framework contain any objectives?

Indeed it does. The current operating framework contains 18 “existing commitments”, for example:

A maximum wait of one month from diagnosis to treatment for all cancers”

and 63 national “vital signs” in three tiers of priority, for example:

NHS Breast Cancer Screening Programme will be extended to all women aged 47–73 by 2012

Much crunchier, but suddenly we’re deep into the detail. These are really sub-objectives created on the basis that, if you achieve them, then they will contribute to your overall objectives. The operating framework does direct us to the 5-year plan for the NHS which “set out a five year vision for the NHS and should be read in conjunction with this NHS Operating Framework which operationalises the first year of that vision”… and so we are straight back to the “vision” again without finding any high-level objectives in between.

Is there anywhere else we could look? How about the legislation that governs the NHS, now consolidated into the NHS Act 2006. Its opening clauses say:

The Secretary of State must continue the promotion in England of a comprehensive health service designed to secure improvement—
(a) in the physical and mental health of the people of England, and
(b) in the prevention, diagnosis and treatment of illness.

…which has several qualifications in it, but could become an objective by changing the first line to: “The objective of the NHS is to secure improvement—”. Add an objective to spend within budget, and to secure local political agreement before making changes to services, and we are starting to build something like high-level objectives for the NHS.

Other areas of life have objectives. For a business it might be “maximising owner value over the long term by selling goods or services” (Sternberg); carefully worded, requiring plenty of sub-objectives to make it happen, but a high-level objective nonetheless.

It seems the NHS has chosen not to define its high-level objectives, preferring to leave a wide gap between vision and detail. Which leads us to the next question. Why?

You are in the Post Office, queueing to send a parcel. A Post Office queue is ruthlessly fair: there is only one of it, and when you get to the front it’s your turn. If the queue is long then you have a long wait ahead of you, and so does everybody else who comes in. If the queue is short, then you and the others have a short wait. In the Post Office, your waiting time is all about the length of the queue, the imbalance between capacity and demand.

A cheeky young man comes in, glances at the long queue snaking around the barriers, and walks straight up to a counter. “Can I just get a passport application form? I’m in a hurry”, he butts in. The clerk sniffs, but reaches around for the form anyway. “How much is the Check & Send service?”, the young man asks. He’s pushing his luck now. People in the queue start to look annoyed.

Why are they annoyed? Because the cheeky young man has pushed in, and so everybody else has to wait longer as a result. If his friend comes in next and does the same thing, there will be complaints. If lots of people push in, the whole system will start to break down; the pushiest will wait no time at all, and the meekest will wait for ages. How long will the meekest wait exactly? That depends on the length of the queue. But it also depends on the amount of queue-jumping.

Queue-jumping doesn’t feature much in real-life Post Offices. But it does feature very strongly in the NHS, and rightly so: clinical priorities such as cancer patients have an excellent reason to jump the queue and they do. So do other patients with deteriorating or dangerous conditions. Nobody minds, because the benefit to the urgent patients (who jump the queue) far outweighs the detriment to the routine patients (who wait longer as a direct result).

Other things happen in the NHS as well. Perhaps a clinic is chock-a-block for the next three weeks, and a cancer referral comes in. Or a patient is booked in for an operation next month but changes her mind, leaving a gap in the session. Or a patient comes in for day surgery and confesses he had a slice of toast for breakfast, so he is sent home and rebooked for another day. It doesn’t happen in the Post Office, but it does in the NHS, and it all jumbles up the order.

What does all this have to do with yin and yang? In Chinese philosophy yin and yang are contrasts, like warm and cool, soft and hard, feminine and masculine. In the NHS, the number of patients waiting is easily measured and simply stated, yang-style. But the order in which they come in is complex, difficult to define, and subtle to justify, more yin.

But yin and yang are also two sides of the same coin, two aspects of the same truth, two parts that form a whole. NHS waiting times depend on the yin as well as the yang: the order in which patients come in, as well as the number of patients waiting. To achieve your waiting time targets, you must manage both. Let either one go, and you may fail.

Waiting lists are a longstanding NHS preoccupation. But scheduling patients in the right order? Not so much. It may be complex, difficult, and subtle, but that does not mean we can leave it to chance. Waiting time targets, and more importantly patient safety, depend on it.

That’s why I thought it would be worth taking a few years to study it properly, to work out the best ways to schedule patients, and to come up with rules of thumb that quantify the effects of each kind of disorder. Now that the work is done, you can learn all about it in just a few hours here.