Flow_ology: the recipe metaphor

You need all the ingredients. And a method, too.

Imagine you've been given the job of preparing a great big three-course Sunday lunch for the full extended family. Three complicated dishes you've never cooked before. But all you've got in front of you is an incomplete, fragmented list of ingredients. No step-by-step instructions. And even the list of ingredients has got vital things missing from it.

The way we currently report patient flow in acute hospitals is a bit like this. We give managers an incomplete list of ingredients and no method. We measure and describe selected random individual bits of the patient journey but we don't show how all the bits fit together. For example, we show managers how many A&E attendances per month, per week per day, per hour, and we show them how long patients spend in A&E before being either discharged or admitted (although we usually abbreviate this to a count of the ones who breach the four-hour target), we show them how many admissions from A&E there were, and so on. But it's selected bits of the patient flow journey that we're showing them, and our selection is too heavily biased towards stuff that's happening at—or close to—the front door, and we take too little account of what's happening further downstream.

But it's not just that we're being incomplete. There's an additional problem: we're being fragmented, too. If we are to understand patient flow we need to connect the bits of information together. We have to make people see that what happens in A&E is affected by what happens downstream of A&E (for example, in Assessment wards). And if we want to understand what's happening in the Assessment wards, we have to be able to see what's happening downstream of Assessment in the specialty wards. Managers need to understand the cause-and-effect relationships between the different bits of the system.

Flow_ology addresses these twin problems of incompleteness and fragmentation. Flow_ology is a three-step method for joining up the data that describes the different bits of the system. It doesn't necessarily provide anything startlingly new in terms of the individual bits of analysis, but what is new about it is that it places emergency care data in a framework (I call it the Flow_ology grid) that helps managers and clinicians make the connections, see the cause-and-effect relationships and then to start putting in place actions to improve patient flow.

So yes, Step One is the grid. And yes, there's an actual grid. A 3x3 grid. And it's colour coded. Yellow, green, blue. And I show you to populate the grid. But the grid itself is only a starting-point. In Flow_ology workshops we use an interactive hands-on Flowopoly set-piece to get people to an understanding of why we chose the three column headings (A&E, Assessment, Wards) and why we chose the three row headings (How many? How long? How full?). And we make sure everyone understands how the grid is held together by "arithmetic glue" because that is a vital part of the step-by-step process when it comes to identifying the "ought-to-be" numbers.

With our grid in place, we can now move on to Step Two and home in on the diagonal relationships in the grid. How long people spend in A&E is usually affected by how full the Assessment wards are. Which in turn are affected by how long people spend in the Assessment wards. Which in turn is affected by how full the downstream wards are. And so on until we arrive at length of stay in the downstream wards which, if we are restricting our focus to the four walls of the acute hospital itself, is where we need to focus our improvement efforts if we are to improve patient flow. This is where scatterplots and set-squares (literal and metaphorical) come in useful.

But it's not enough to just understand the cause-and-effect relationships, which in many cases are actually pretty obvious. We have to use this understanding to move on to Step Three and specify what the "ought-to-be numbers" are. This is the real biting point of Flow_ology. This is Flow_ology's punchline. We use the grid to get us to the cause-and-effect relationships. And we use the cause-and-effect relationships to get us to the ought-to-be numbers. We don't just tell managers what the numbers are that are associated with poor patient flow; we tell managers what their numbers ought to be if they want patient flow to be better. And we recognise that these ought-to-be numbers need to be specified right down to individual ward level.

So if you follow the three steps of the Flow_ology recipe (1. Identify the data you need to populate the grid; 2. Use data to describe and understand the cause-and-effect relationships within the grid; 3. Use that understanding of the relationships to identify the ought-to-be numbers) you reach a point where you've replaced an incomplete, fragmented list of random "ingredients" with a whole-system, joined-up narrative whose punchline is a new set of numbers that managers and clinicians can use to inform plans and actions that will improve patient flow.

There are two open course dates for Flow_ology in the diary: Friday 24th February is London; Wednesday 29th March in Manchester. Email or phone me if you want to book a place, or if you want to talk about how to organise an on-site workshop.

There are more details about the session-by-session content on the Flow_ology page. And you can download a two-sides-of-A4 pdf flyer here.

[10 February 2017]

© Kurtosis 2017. All Rights Reserved.