Home   About Kurtosis   Courses   Course Calendar   Booking Information   Ideas   Technique   Clients   Contact

Flow_ology: a Mid-term Report

Five things we've learned from the first four months of Flow_ology

(1 comment)

Flow_ology is four months old this week. It had its first outing in Edinburgh on Monday 23rd September last year, soófour months onóit seemed like it was time to take stock of what weíve learned from it so far, and how itís evolved over the last four months.

What is Flow_ology?

But first, letís just remind ourselves of what Flow_ology (see footnote) is.

Flow_ology addresses the problem of how to use data to make sense of unscheduled care. It recognises that unscheduled care is a stubborn problem (to say the least!), and that if data is to have a role in solving that problem, then weíre probably going to have to use data in a different way than we have done up to now.

When you ask NHS managers and clinicians about how they use data to make sense of unscheduled care you get one of two answers: they either get too much data or they get no data at all. (I find that this latter response is more typical, the closer you get to the coalface. It's a constant source of anxiety for me that, once you leave the confines of the board room and you start talking to ďordinaryĒ consultants and charge nurses, you realise that these people arenít being told about how many patients they are admitting, how long those patients are in hospital for, or how full their wards are.)

Either way (too much data or no data at all), Flow_ology proceeds by providing a method for analysing and presenting unscheduled care data in a structured way. Itís a way of ensuring that the organisation publishes a specific set of indicators, that this set of indicators is customizable, that everybody gets to see all of the indicators (in other words, donít just show the specialty ward data to the specialty ward staff; show it to everybody), and that by showing everything together in one place will help people make connections between different parts of the whole system.

The objective of Flow_ology is to get more unscheduled care data "out there", but to get it out there in a structured way, in a way that helps people make the connections and identify the indicators that really matter in terms of understanding the cause-and-effect relationships.

Flow_ology comes in two flavours: thereís a one-day course for analysts (Flow_ology: the analystís cut); and thereís the tailored six-day consultancy project thatóas well as helping analysts do the number-crunching and dialoguing with coalface workersóinvolves organising plenary sessions to present the data to multi-disciplinary meetings of managers and clinicians.

So: what are the five things weíve learned over the last four months?

1 / People like the grid

The grid is more useful / has more resonance than I thought it would be. People like the idea that the data is now in some kind of order. I hadnít expected the grid to be as popular as it seems to be. Itís an antidote to the random haphazard-ness of how data is often published. People seem to like the idea that there is an underlying logic and order to the indicators.

2 / People prefer narratives to data

Although the order that the grid imposes is logicalóA1 is A&E attendances, A2 is A&E length of stay, A3 is A&E occupancy, B1 is Assessment admissions, and so onóthe order that you present the data in (as a narrative) will most likely be quite different to that order. For example, the slide sequence I used on Wednesday this week when presenting to a Director of Operations and a Medical Divisional Manager was: Slide 4, Slide 5, Slide 8, Slide 12 Slide 21, Slide 22, Slide 4, Slide 30, well, you get the idea.

So what weíve learned is that although as analysts it makes sense to populate the grid in a structured way, when we are constructing a narrative that explains why things are the way they are, weíll probably need to put the data in a different order. When youíre telling peopleó"real" people, that is, not analystsóabout Flow_ology, you shouldnít spend too much time on how itís a way of ordering the data. "Real" people arenít interested in that; you should focus instead on telling them what the data is saying and then, perhaps, in passing, at the end, if you must, tell them that itís this new of looking at the data that has enabled these insights.

3 / Little things make a big difference

I think we knew this all along, but Flow_ology helps you see that little things can make a big difference. When weíre looking for the reasons why performance has suddenly improved or suddenly deteriorated, itís often tempting to go in search of indicators that have moved similarly suddenly. But it doesnít always work like that. One example of this might be that a seemingly small increase in the average age of people attending A&E can have a disproportionate impact on the hospital. Another example has been that you donít have to have a week of extreme high-volume admissions to put the system out of kilter; instead it can be several weeks of unrelenting just-slightly-higher-than-average admissions that can have the most serious effect.

4 / The space between the notes

Itís not just whatís happening within the staging posts; itís whatís happening between the staging posts. Weíre increasingly finding that the conversion rates between the staging posts are the indicators that people donít look at routinely. But these are the indicators that cause the dysfunction. A&E attendances can be flat but if a higher proportion of them than normal convert to Assessment admissions, you will encounter problems. Similarly, in Assessment, you need to watch carefully how many of these patients convert into specialty ward transfers, and, in turn, look at how many of these go on to need complexóas opposed to simpleódischarge.

5 / Itís good to talk

Flow_ology is a great vehicle for getting analysts to realise that sharing work-in-progress informally with clinicians and managers is not only good in terms of moving the debate forwards and enhancing understanding of how the system works; itís also a rewarding thing to do in its own right. Dialogue is good for analyst job satisfaction, and dialogue flows naturally out of Flow_ology.

In Summary

Flow_ology works. It provides order. It has resonance. It helps people make connections between the different parts of the system. And it enables narratives.

Footnote: Flow_ology in six steps

The best way of summarising what Flow_ology does is to think of it as a six-step sequence that people need to understand and follow if they are to use data to get to grips with urgent care, and the sequence goes as follows:

  1. The first step in understanding flow is that you have to know what the indicators are that tell you that flow isn't happening properly. This means you have to be able to identify the symptoms of dysfunction.

  2. Secondly, you have to describe the processes in a way that has resonance for clinicians. This means identifying activity in the different staging posts. (a) A&E; (b) Assessment; (c) Specialty wards; (d) Post-acute. The key thing here is to separate Assessment inpatient activity from downstream ward inpatient activity. If you don't do this, the data doesn't mean anything to the clinicians.

  3. Thirdly, you have to combine all the data (from the four different staging posts) together into a grid. It's a 4 x 3 grid, because for each staging post there are measures of (a) activity, (b) length of stay and (c) occupancy (I actually prefer the word "full-ness"). The grid means that the data can be tamed and organised. Managers no longer feel as if they're being bombarded randomly by haphazard chunks of data. Also - the grid forces people to see the whole system in one eye-sweep, as it were, so that the connections between the different bits of the system are more apparent.

  4. The Grid allows you to both drill down into the data and to visualize the data in different ways. For example, to use the example you mentioned, you need to look at the length of stay in A&E of admitted patients separately from non-admitted patients. Also, you will want to look at the data sometimes as a time series and at other times as a distribution.

  5. The Grid also enables people to explore the connections between different parts of the system in a systematic way. You can put under the microscope the relationship between - for example - bed occupancy in Assessment and length of stay in A&E. This is a crucial stage because it helps people understand the cause-and-effect relationships in the system.

  6. Finally - all of this helps the organisation construct a story - a narrative explanation - of why things are as bad as they are. Why on some days they are not quite so bad. And why on some days they are apocalyptically bad.

[24 January 2014]


Make a comment on this blogpost:


Job Title:


Email address:


© Kurtosis 2014. All Rights Reserved.    

Comments on this article

24 January 2014:

This feels different, exciting, science, flow and narrative, people getting curious about how things work, relating experience to data. Brilliant.

Harry Longman

Chief Executive, Patient Access Ltd