Fri 11th December 2015 / 1:00pm to 4:15pm / Whitespace, Norloch House, 36 King's Stables Road, Edinburgh EH1 2EU

Obstacles to using data

Trying to reconcile different worldviews

To recap—just in case we needed to recap—the question we're addressing at Flow Data Symposium is "How can we bring data to bear on the patient flow problem?"

We all know it's important to do this, to use evidence (in the form of data) to inform our understanding and actions, but we somehow just don't seem to be able to do it—not the demand and capacity aspects of patient flow, at any rate. We're seemingly OK about using data to describe the "here-and-now-ness" of emergency care; but we're not very good at describing (or understanding) the general "big picture-ness" of it. What I mean by that is that we're reasonably comfortable using data to describe—for example—how many patients are currently breaching the four-hour target, or how many medical patients are currently boarding in surgical beds, or how many delayed discharges there currently are. But we're less skilled at using data to understand why these things are happening in the first place—what are the system dynamics that are causing the dysfunction?

And—in my opinion—the reason we aren't very good at it is that we haven't grasped the enormity of the obstacles in the way.

What obstacles?

The biggest obstacle is that the "data view" of the unscheduled healthcare world is—in my opinion—completely at odds with the "clinician/manager view". The "data view" is "patients plural" whereas the "clinician/manager view" is "patient singular". For as long as we use data to aggregate and average and summarise and calculate percentages, then we're going to find it very difficult to reconcile these two worldviews. (But of course the problem is that we do have to do all of these things to the data, so we have to find a way of making this aggregation—which in some ways is the negation of the individual patient experience—not just acceptable, but necessary.)

Secondly, beyond this basic, generic, abstract problem, there's another problem lying in wait for us. The purveyors of data haven't usually made much of an effort to find out how the clinicians and managers actually see the world. Data analysts tend to just assume that looking at tables and graphs is a natural activity for everybody to indulge in. Well, actually, it's a lot more complicated than that. Most NHS managers and clinicians don't see the world as a world that can easily be depicted—let alone understood—through the medium of tables and charts, so analysts who only use these as their means of communicating evidence are doomed. Pierre Wack used to emphasise that you have to somehow arrive at a shared mental model if you want to make headway with solving a difficult problem.

But even if you can get to a point where you've created a shared mental model, there's still the motivation obstacle to surmount. People need incentives to do something about the problems confronting them. And in the NHS this usually involves invoking individual patient experience. Most cliniciand and managers in the NHS are motivated to solve problems if they are doing it on behalf of patients. The care of Joe Smith would have been better if we'd have done x and y. As opposed to the analyst view of the universe that says: if only the numbers looked like this instead of like that.

Flowopoly was originally conceived and designed to try and cross the "species barrier" that separates the "clinician view" and the "data view". The central idea was that there'd be tables full of cards that were self-evidently patients but that were also—equally self-evidently—data. Everybody would see this. It was a way of helping people move from a patient-centred universe to a data-centred universe.

(Digression. There's a danger with universe metaphors. The temptation is to draw a parallel with the transition in 16th Century European thinking from a pre-Copernican geo-centric universe to a post-Copernican helio-centric solar system. The problem with this analogy is that the astronomy story is about replacing one paradigm with another. But in healthcare we need the two paradigms (the patient-centric and the data-centric) to exist side by side. It's as if F. Scott Fitzgerald's quote about the sign of a first-rate intelligence being "the ability to hold two opposed ideas in the mind at the same time, and still retain the ability to function" was made for this very situation.)

But although the theory was there, underpinning Flowopoly as a concept, it didn't necessarily work quite as smoothly as that in practice. As we've developed Flowopoly, we've realised that some things have worked, whilst others haven't.

Here are three things we've discovered about this along the way.

First, as a reflection of people's mental models of what the unscheduled care system looks like, Flowopoly seems to work pretty well. It's a basic model (although it's also unlike any other way that data is presented, so although it's basic, it's far from conventional), but it works. There is enough verisimilitude for people to recognise it straightaway. That doesn't always happen with conventional ways of presenting data, so we regard that as a step in the right direction.

Second, we've noticed a bias or tendency for people to ask for more specificity. As opposed to more generality. Nobody ever asks for more generality. This manifests itself in two ways. One is for more clinical detail about individual patients. The other is for patient stories. People always ask for patient stories. So we feel duty bound to give them patient stories.

Third, despite the fact that Flowopoly does indeed marry up the patient-centred universe with the data-centred universe, it doesn't on its own enable change or improvement to happen. Instead, it's a springboard for action. Or—perhaps better, in the context of Flow Data Symposium—a solid-yet-innovative platform on which to build a meaningful and fruitful conversation.

[13 November 2015]