Why a health service is like a bicycle OR the importance of deimplementation

Much of the emphasis in knowledge mobilisation is on getting new things into practice. The term “implementation science” conveys this too: we want to implement stuff. But just as important can be getting things out of practice: de-implementation. The rationale is straightforward. A health service is much like a bicycle, and the case of Mrs Armitage makes clear the problem.

Mrs Armitage on her bike - so far, so good

Mrs Armitage is a creation of English writer and illustrator Quentin Blake. She appears in three books: Mrs Armitage on Wheels, Mrs Armitage and the Big Wave, and Mrs Armitage, Queen of the Road. In both Wheels and Big Wave Mrs Armitage engages in similar behaviour. She takes something that works – in the first case a bicycle, in the second case a surfboard – and, perceiving various shortcomings, adds to it until disaster threatens. (I’ll come back to Queen of the Road, which is different.) On the bicycle Mrs Armitage is concerned that hedgehogs won’t hear her coming so she adds a selection of motorhorns; worried that she may need tools in case of breakdown so adds a toolbox; alarmed that her dog Breakspear is tiring so adds a seat for him; and so on: a snack box, a radio-cassette player, a mouth organ… etc.

So it is, all too often, with our health services. New technologies, new ways of working, new diagnostics come along and, if they seem to work and we can implement them, we add them to what we provide. But we often don’t or can’t remove or reduce the form of care or technology that the new one was intended to replace. In other situations, things are implemented on the basis of little or no evidence, never challenged, and persist indefinitely as established practice.¬†Like Mrs Armitage’s bike, the health service gets bigger, more expensive, more unwieldy.

Mrs Armitage's bike: overburdened, disaster beckons

And what happened to Mrs Armitage’s bike? Overburdened and out of control, it crashes and she and Breakspear find themselves sitting amongst the wreckage.

To avoid the looming possibility of expensive and unwieldy health services that risk crashing whole economies we need, rather than constantly thinking about implementing and adding, to think also about how to disimplement and take things away. Implementation Science published a useful short article on this by Vinay Prasad and John Ioannidis in which the authors set out a conceptual framework for evidence-based de-implementation and followed by a note from the editors stating they welcomed further contributions on de-implementation. De-implementation is not just the opposite of implementation and is likely to require different approaches and thoughtful ways to identify practices and technologies that should be de-implemented, then work to find strategies and techniques to de-implement and sustain the necessary changes.

Perhaps Quentin Blake thought about this too. In Queen of the Road Mrs Armitage begins with an antiquated car from which bits gradually fall off – hubcaps, roof, doors, and so on, always received by Mrs Armitage with a statement such as “Hubcaps? Who needs them?” – until she’s finally left with a stripped-down and efficient-looking roadster and annointed, by her uncle and his friendly biker friends, as Queen of the Road.

Hubcaps. Who needs them?

Getting to a stripped-down and efficient health service is the ultimate aim of both implementation and de-implementation. We just have to make sure we don’t forget the second part.

 

References

Prasad V, Ioannidis JPA. Evidence-based de-implementation for contradicted, unproven, and aspiring healthcare practices. Implement Sci. 2014; 9: 1. http://www.implementationscience.com/content/9/1/1

Cake-making and implementation

Becky wrote¬†an interesting post recently about KM and cakes. She made the very good point that knowledge mobilisation isn’t something that happens at the end of the research process, like adding the icing to a cake, but that it is part and parcel of research. I thought this was a great metaphor and want to take it a little further.

portal-cake

Let’s stick with the idea that research is like making a cake. A certain understanding of research is that cake-makers (clever researchers) make a fantastic new cake (do some research), present the cake (do some dissemination, probably through the standard academic routes of peer-reviewed publication and conference presentations) to a room full of hungry people (an imagined audience of practitioners or clinicians or policy-makers or members of the public, etc.) and then the people consume it (it is taken up and becomes part of routine practice). Or at least that’s how some people seem to imagine it works.

More commonly, in my experience, the cake is made and then one of a number of things happen, including:

– lots of people hear about the cake and eat it (this doesn’t happen often)
– some people see the cake but think it’s for someone else so they don’t eat any
– some people hear about the cake and one or two people nibble the cake but they think it looks funny or smells funny or just plain don’t like the way it tastes
– a few people hear about the cake but there are so many cakes to choose from that they are distracted elsewhere
– a few people hear about the cake but already have a cake so don’t pay any more attention
– the cake sits in a cupboard for a while. Eventually it gets so mouldy nobody would ever eat it. Perhaps it’s sitting there still. (This happens a lot.)

So sometimes this approach works but very often it does not. A lot of the time this is because it’s simply the wrong cake.

It’s your birthday and someone brings along a wedding cake: wrong cake.
It’s breakfast time and someone brings along a rich chocolate cake: wrong cake.
Everyone’s asked to bring along a salad and you turn up with your delicious pineapple upside-down cake: wrong cake – in fact, wrong food altogether.

The cake need not be totally wrong – maybe you’re happy to eat wedding cake on your birthday or have chocolate cake for breakfast – but for a lot of people it will be. And so it goes with research, at least according to the model outlined above: something is prepared with limited or not understanding of the context in which it’s going to be used.

Wouldn’t it be better if we could avoid this problem? How about, instead of turning up with a cake and hoping the people in the room will like it, we speak to them beforehand and find out what kind of cake they would like? A birthday? Great – I can make you a birthday cake! Going even further, we can keep speaking to the people who will eat the cake throughout the process to find out what they need and have them contribute to the cake-making: How many people is the cake for? A little less sugar? How thick do you like your icing? You’re not going to produce the perfect cake but you’re going to come a heck of a lot closer to producing the kind of thing people want to eat than if you just turn up with your random cake.

Cake making. It’s an implementation thing. (The cake is not a lie.)