When I heard that the #HCLDR tweetchat this week was on the subject of “positive deviance” I knew immediately I wanted to write about it. It’s not that I think I’m positively deviant when it comes to interest in healthcare workflow, it’s that I’ve used the principles behind positive deviance to find examples of great workflow in healthcare organizations.
Now, as presented in the HCLDR blog, positive deviance is relevant to finding and duplicating healthy behaviors in individuals. Many of the same ideas I’ve had about positively deviant workflows can also be applied to life-flows as well.
May be of interest > Discovery of Personal Processes from Labeled Sensor Data
— Charles Webster MD ()
Plus, organizations can be viewed as having positive behaviors, and workflow is, I believe, the appropriate level of abstraction for documenting, generalizing from, and duplicating healthy healthcare organizational behaviors and workflows.
Way back in 2009 I wrote a paper about process-aware information systems in healthcare and submitted it to MedInfo in Cape Town South Africa to occur in 2010. Here was my “algorithm” for improving healthcare organizations’ processes (AKA workflows).
1. Generating process models of existing practices.
2. Comparing measures of productivity (throughput and throughput time).
3. Explaining differences in productivity in terms of differences in processes.
4. Suggesting process improvements for low productivity practices.
Compare above with
1. Identify positive deviants or teams of positive deviants
2. Study their behavior to generate hypotheses about practices that allow organizations to achieve top performance
3. Test hypotheses statistically in larger, representative samples of organizations
4. Work in partnership with key stakeholders to disseminate the evidence about newly characterized best practices
http://qualitysafety.bmj.com/content/early/2014/07/21/bmjqs-2014-003115.full
And here is my slide from 2012, at the annual Healthcare Systems Process Improvement Conference, in Las Vegas.
Here is pecific example where we rank best and worst practices re throughput.
Then we drill down into process mined process maps to explain why the good practices were good and the bad practices were bad.
The EHR trainers and configurers would “diffuse” workflow innovations (the positive deviance) from good practices to bad practices.
Anyway, great questions!
T1 What examples have you seen of Positive Deviance in healthcare?
See above.
T2 What challenges do you see in trying to scale what works on one team to the entire organization or from one region to the entire community?
Coming up with representations of the behavior that are perspicuous and transportable. In our case, we used models of workflow, and then we tweaked workflows according to these models.
T3 Which do you believe is better – studying errors & why they happen or successes & how we can replicate it?
As you can see, we did both. Often the biggest improvements came when we turned errors into successes.
T4 As healthcare leaders, what can we do to encourage/highlight positive deviant behavior?
Strictly workflow speaking, we need time-stamped event data, aligned with both workflows and means to easily change workflow, without re-writing a lot of code or putting users through training ringer.
P.S. The following is from an blog post titled A Conversation About EMRs, Workflow, Usability, and Productivity. It’s about finding steps in workflow that are more frequent, more costly, and take longer than average. But the same approach could be used to find workflow tasks that are less frequent, less costly, or shorter in duration than average.
“Chuck: One of the great promises of EHR workflow management systems or healthcare BPM in general, and specialty-specific EMR workflow systems in particular, is the pairing of activity-based costs with process definitions. Since each step in a process definition is time stamped as to when it is available to be accomplished, when it starts to be accomplished, and when it is actually accomplished and who (cost per minute) and where (rent per minute) is the resource used during each task, the total cost of each patient encounter can be calculated. In conjunction with the revenue per patient encounter that is available from the practice management system, the profit per each encounter can be calculated.
By comparing encounter profitability across similar medical practices, specific reasons for decreased profitability can be located: (1) a step is more expensive per minute than it should be (that is, it is accomplished by less expensive resources at other practices), (2) a step takes longer to accomplish than it should (compared to other practices), and (3) a step is executed more frequently than it should (compared to other practices). The win-win-win analytic result is to find those too expensive and too long steps that are being executed too frequently and change the workflow so as to increase encounter profitability.”
The following is from a presentation I gave in 2012.
“Suppose you know some Key Performance Indicators (KPIs) for these facilities, such as patient throughput and cycle time, cost per encounter or encounter type, or perhaps even measures of user or patient satisfaction. Process mining can generate process models that you can compare to explain differences between KPIs [6]. Traditional clinical business intelligence report and dashboard software may tell you what the KPIs are and help benchmark them. However, to understand the likely causes of flagged KPIs, you’ll need evidence-based process models such as process mining can provide.”