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

The Wrong Kind of Bad

Yesterday, we needed a typically bad day; not an unusually bad day.

We chose the wrong bad day. So it's a good job it was only a rehearsal. We've got time to change it before the main event in May. And—most importantly—we got some clear pointers from the participants on how to choose a better bad day.

The wrong kind of bad day? What am I talking about?

I’m talking about yesterday's rehearsal of Flow_opoly.

For those of you who've missed me talking about this latest Kurtosis venture (actually a collaboration between Kurtosis and Intandum), Flow_opoly is a game-like table top exercise scenario re-enactment thing where people move cards around a table in order to get a better understanding of how patient flow works in an acute hospital and beyond. We first of all play out a “good flow” day, then we follow that with a re-enactment of a “bad flow” day. A lot of other things happen, too, but it’s this contrast between the good day and the bad day that I want to talk about here.

What we learned in yesterday’s rehearsal is just how important it is to choose a good “good day” and a good “bad day”. And what we mean by a good “good day” and a good “bad day” is that these days must be typically good and typically bad.

So: here are three observations about what the phrases “typically good” and “typically bad” mean.

First, I need to point out what criteria we actually used to choose our good and bad days. The good day was picked primarily because it was a day when nobody breached the four-hour target. But we were also conscious that we didn’t want to pick a day that was unusually quiet in terms of A&E attendances, or a day that was unusually blessed with plentiful empty beds. So we picked for our good day a Monday in October with a higher-than-average number of A&E attendances and where there were only five empty medical beds at the start of the day.

As for the bad day, we looked at the days with the highest number of four-hour breaches. We steered clear of a day that had such an unusually high number of breaches that we would risk being accused of picking an unusually bad day, and instead plumped for a day in mid-December when there were twelve breaches of the four hour target. Butimportantlywe didn’t check for anything else.

So, primarily we were using the four-hour target as our main criterion for deciding between good and bad. But although performance against the four-hour target is often a barometer of whole system health, it isn’t always a barometer.

Secondly, we were strongly reminded that the whole point of a re-enactment is that you don’t really want to see puzzled faces when you look around the room. Particularly when you’ve gone to the effort of making sure that it’s coalface workers and stakeholders that you’ve actually got in the room. In other words, the people playing the game are the very people who inhabit the whole system you are trying to describe using the tables, boards and cards.

The people who inhabit the system, they know that system, and they certainly know how their bit of the system works. You want them to be nodding their heads and saying things like “Yes, this is what happens. This is what it feels like on a good day. This is what it feels like on a bad day.” You don’t want them shaking their heads, saying things like “No, sorry, I don’t recognise this. If that was happening over there, then this would be happening over here. And it isn’t. And I’m confused.”

As part of my background reading for Flow_opoly, I’ve been reading a book called Moments of Impact: How to Design Strategic Conversations that Accelerate Change The authors reverently quote Pierre Wack (who was head of the legendary group planning team at Royal Dutch Shell in the 1970s) when he talked about the need to build on people’s existing knowledge rather than argue against it. You have to start by creating stories and visuals that that resonate with managers’ existing mental models and tap into the emotional and pattern-recognition parts of their brains – not just their analytic circuits.

This is what we are trying to do with Flow_opoly. Except that instead of stories and visuals we’re using a great big table with boards and cards.


Pierre Wack’s scenario planning sessions always started with a “conventional wisdom scenario” – a story that best represented managers’ current baseline assumptions about how the world works. Wack would treat this baseline scenario with due respect and show that many of its assumptions were valid. Then he’d gradually expand the managers’ field of vision by holding up a mirror to their perspectives, turning and twisting it from multiple angles. In doing so, Wack was bending and expanding the managers’ mental models one step at a time – rather than breaking or replacing them.


So that’s why we have to choose the good and bad days wisely. They have to be days that typify what good and bad days feel like, that best represent clinicians’ current baseline assumptions about how the world works. If you don’t get typical days, you can’t build from strong foundations.

Thirdly, the word “typical” has important repercussions for when we use data to describe and make sense of complex reality. What do I mean by that? Well, usually when we use data, we are making a generalisation. To take examples from acute unscheduled care systems, when we say that average length of stay is 6.9 days, we’re making a generalisation. When we say there were – on average – 21.4 admissions per day to the AMU last year, we’re making a generalisation. 96.2% average bed occupancy is a generalisation. And so on.

In other words, we know that there were individual lengths of stay that were a lot shorter or a lot longer than 6.9 days, and we know there were individual days when there were only 12 admissions in the day (as well as days when there were more than 40). But we are still making a generalisation.

The opposite of general is particular. Particular is what we talk about when we recount anecdotes. This particular patientJohn Smitharrived at 14:03, stayed in A&E until 17:45, when he moved to the AMU. He stayed there until 10:23 the following morning, at which point he was transferred to the Respiratory ward. And so on.

When we play Flow_opoly we are effectively re-enacting a series of anecdotes. In yesterday’s rehearsal there were 288 good day anecdotes and 305 bad day anecdotes. If you want to treat each individual day as an anecdote, then what we were doing yesterday was re-enacting two anecdotes: the good day and the bad day.

But if anecdotes are going to be useful, if they are going to help us make sense of the world, they have to be typical anecdotes. They have to be representative stories. It’s no use choosing the exceptional ones. Exceptional stories do have their uses, but they are less useful when it comes to working out what the patterns are.

So yes – in yesterday’s rehearsal, our good day was fine. It was typical. But our bad day was atypical. It was the wrong kind of bad. We need to choose a better bad day.

[19 March 2014]

Note: All identifying fields on the Flow_opoly cards are pseudonymized.


Make a comment on this blogpost:


Job Title:


Email address:


© Kurtosis 2014. All Rights Reserved.