Maximising SDEC: reducing variation

Written by Spencer Humphrys, Managing Director - Nexus Consulting

In my last blog on Maximising SDEC (same day emergency care; you can read it here), I talked about how getting the right patients into SDEC, even if that means SDEC isn't 'full', can have a far more positive impact on flow than just sending everyone that 'could' go. Hopefully, my message was clear:

SDEC works best when it treats the right patients. And when SDEC works well, flow improves, cost reduces, and quality gets better

I got a few comments and direct messages on the importance of the 'right patient, right place' philosophy, which was nice (it let me know I'm not just talking to myself at least!). But one message really caught my attention. Part of the message was, 'Does the wrong patient really matter, if the outcome is the same?'

In short, yes. It matters a huge amount. But the comment got stuck in my head. I wondered whether this might be a common misconception or an area where people understand the concept but may not have seen the science behind it. So this blog will go a little deeper into the science of the 'right patient, right place' philosophy and queuing theory, and it will touch on the short vs long-term thinking that I will write about another time.

---

So - let's start by busting some myths:

1. Queues are not random

2. Queues are not solely built by increasing demand

3. Queues will grow even when you have enough capacity to meet your demand

4. Queues are dangerous. As queues lengthen, whether they are 'virtual' (e.g. waiting lists; customer orders) or physical (e.g. A&E overcrowding; WIP build-ups), risk increases

5. Queues for admission avoidance services result in avoidable admissions, which, in turn, increase queues for inpatient beds, which then increase the need for admission avoidance services. It's a never-ending circle

---

Why queues appear

Waiting times are driven by three interacting factors:

• Utilisation – how close the service is to full capacity

• Variation – how predictable arrivals and service times are

• Processing time – how long it takes to process one item (this could be a patient, a referral, an item, anything really (and yes, I know I shouldn't use the same words in a definition!))

As any of the three factors above increase, waiting times will increase as well. But most importantly, as a service approaches full capacity, waiting times increase exponentially.

All service providers will have felt this. A service looks fine on paper, the shift (or week, depending on your cycle time) is ok, then something small changes. Maybe one or two more referrals than normal. Maybe one complex patient that takes longer than planned, and all hell breaks loose. Waiting times go through the roof. The tipping point has been crossed. Staff are knackered and stressed; quality reduces; nerves fray. Queues continue to build, simply because of the final small change in utilisation, variability, or processing time.

I won't go into the maths of queues here, but essentially:

Queues develop from utilisation x variation x processing time.

Utilisation is primarily driven by demand and processing time, and it's the variation in these factors that really impacts service performance.

Put simply, if you're expecting three referrals an hour, and they take four to six hours each to process, you can plan and manage your service. But if you have three referrals an hour and they take anywhere from one to eight hours to process, things get tricky. Tricky to know how to staff the service, how to manage peaks, and when to accept the last referral. And this variation makes it hard for referring departments too. Ever heard someone say, "When they close depends on the day of the week or who's in charge"? That's always a symptom of variation, but not always a symptom of person-to-person variation.

Focussing on SDEC....

Hopefully, the penny is beginning to drop now.

Every patient referred to SDEC who doesn't need it disrupts process control by increasing variation in processing time, and if you're referring people just because A&E is full, you're probably introducing more variation into demand too. And every delay in referral or pulling to SDEC adds even more variation and problems in demand management.

Operationally, these issues usually show up as symptoms of high SDEC utilisation during peak periods and amplify uneven arrival patterns ("We would be fine, but we've got 5 patients waiting for SDEC").

But, isn't SDEC more effective and efficient if they see some simple cases too?

No. Effectiveness of SDEC services should be measured only by outcomes achieved (e.g., a reduction in short-stay admissions or admissions by condition-specific HRGs). Utilisation will increase by seeing these simple non-SDEC patients, but increasing utilisation isn't helpful beyond ~85%. But more fundamentally, processing time in SDEC is influenced by many factors, including:

• access to diagnostics

• decision-making seniority

• pathway design for common conditions

• clarity of ambulatory protocols

• operational coordination across departments

When processes are streamlined to meet the high-volume SDEC demand, the average time spent in SDEC decreases. This does more than simply shorten individual patient journeys. It increases effective capacity. But by adding in non-standard referrals, it's likely that usual processes will have to be flexed. Maybe the carved-out diagnostic capacity for the 'true' SDEC patients is consumed by patients who didn't really need it. Maybe the physical space to take the patients who would really benefit becomes maxxed out. Whatever the reason, utilisation by patients who genuinely could benefit declines, and effectiveness declines with it. And as queues build, those who could have been seen and discharged inevitably end up staying overnight. Whether that's in an inpatient bed, in A&E, or in a corridor. They're still with you when they could have been at home.

If your SDEC isn't delivering its potential. Here are three things you can consider

1. Start simply. Ask the SDEC clinical team how much of their work genuinely avoids an admission. Then ask the A&E teams what value they get most from SDEC. I would bet money the answers won't match. And this misalignment will be driving variation up and effectiveness down.

2. Understand your variation. How much of your current service variation is due to uncontrollable factors (e.g. arrival times), and how much is influenced by A&E function, decision-making times, and four-hour pressures?

3. Understand your processes. Do you actually know how much time it takes to process one SDEC patient? Not just hours in the unit, but set-up time, clinical contact time, and follow-up activity (if essential).

   
   

These are simple questions, but the answers are the fundamentals. Improving them isn't likely to require large structural changes, but improving them together can transform your SDEC's effectiveness and staff morale.

---

If you this blog has struck a chord with you, please let us know, share with a colleague, or leave a comment on LinkedIn - we're really appreciate it.

And if you're really struggling with getting the best from your SDEC unit and team and you need help, reach out for a chat. We know what it takes to develop and run acute services. And we know how hard it is in practice.

Contact us