The Influence of Context in Successful Implementation
Posted on November 15, 2011
John A. Øvretveit points out that the effectiveness of an intervention may depend on contextual variables.
Quite a bit of research I've been involved in is research in collaborative breakthroughs, and this is the spread method of choice in the health sector. And basically what it is is you have implementation teams from up to 30 or 40 services coming together three times over a year in for two days in things called learning sessions.
And they come to the first learning session, and they hear about evidence-based practices for achieving— solving that particular problem. It might be getting easy access, or it might be reducing infection rates or some other change, and they also get taught about the Plan, Do, Study, Act cycle and other change methods.
Then they go back and try out changes or plan it. Then they come back in three months and share their experience with their peers and get more input about, "You really should be measuring and collecting data. Here's how you collect data to test your changes," and get feedback. And then they go away for three months and then come back again, saying where they've got up to, and what they're doing.
So these what are called collaboratives in effect are temporary organizations set up to support these teams and their organizations to help them make these changes and learn the skills. And that's the method of choice for spreading evidence-based practices in the health sector and reasonably successful. I mean, a third will do quite well and make significant, proven changes. A third will sort of, "We're nearly there, but we need really another year for it to show." And a third, "Well, we can't collect any data. We've got all these problems and struggles," and a few will kind of drop out.
What's really interesting is the five or three of them—teams in different places that got spectacular improvements—like, for example, reduce their infection rates fantastically. One of the most well-known of these programs is the central line bloodstream infection program that was done in Michigan and elsewhere in intensive care units. If you get a bloodstream infection from a central line, which will put in certain medications, it can be very dangerous and very costly as well. You'll need a lot of aggressive treatment to stay alive, and you stay in the ICU a lot longer.
So with very simple procedures, you can reduce those bloodstream infections, but you have to follow these very carefully, and there are ways of doing this. And these have had very, very good success, and they're given as an example of what you can do to improve healthcare quality quite quickly.
And I think we can learn a lot from research that looks at implementational spread programs and to contrast the really successful sites and services with the unsuccessful ones and try and explain. And either we do retrospectively or we can actually set up an implementation study to say, "We predict that some will do very well, some will do very badly," and part of the study is to follow and understand those differences because we know this is going to happen in that way.
And I think much more research should be looking at those extremes and explaining that rather than simply reporting, "Oh, well this is the total average. I think it's the variations that are more interesting because it relates back to the point about it's not so much the intervention, it's the context that is as important as the intervention, and what aspects of the context are really important to allowing that change to happen and take root. And which aspects of context are hostile or toxic to that sort of change and learning more about that, so that we can then say, "This intervention will get fantastic results in this situation or under these conditions," rather than pretending that it will get those good results everywhere.
Excerpted from interview with Dr. Øvretveit at the 2011 Global Implementation Conference in Washington, DC. Dr. Øvretveit is a professor of implementation, improvement, and evaluation at the Karolinska Institute in Sweden.
Please note that the feedback is viewed only by 4researchers staff and is not intended for communication with individual contributors.
Use the form below to submit feedback about this article. If you would like a response, please be sure to include your e-mail address.