It’s the Workflow, Stupid

When nurses complain that the technology-based tools (especially EHRs) that they are tasked to use slow them down, usually what they mean is that these tools don’t model their workflow.  We talk about workflow a lot!  But what is it?  Why is it so important?  And why is it so hard to model with computerized/electronic/mobile tools?


One answer is the “throw it over the wall” model of hardware and software development.  There are a couple of issues here.  One is the linear, non-iterative product development process.  That is, traditionally, technology-based product development has started with requirements definition and product specification — what it should do and how it should look.  If end-users (nurses!) are involved at all in the process, it is here.  Later comes development, testing, and deployment.  By then, the product  functionality and design are frozen and the product is “thrown over the wall” —“Here is what you ordered, like it or not.  Our job is done.  Live with it.”


The second aspect of this dysfunctional yet common product development process is the assumptions of the developers.  Nurses will be the users but the developers are engineers.  Engineers often lack deep understanding of the content or process of nursing work yet they make statements such as, “Oh, nursing, I know all about it.  My wife is a nurse.”


What that engineer doesn’t know that he (or she) doesn’t know is about the workflow of nursing.  And they are not alone in this information deficit.  Even the magazine, in an otherwise excellent article entitled “Interoperability: Better Care Through Better Information Sharing” (Jan/Feb 2016) describes as obstacles to usability the simple user interface design errors such as poor screen design, cluttered screens, and too many clicks needed to document a task, but none of these are actual workflow issues.  The article does observe that “Clinical workflow [is] the end-all, be-all” but fails to say what that means. 


More subtle yet realistic issues are those such as the one I described in my healthcare technology blog about my friend Pat’s experience in which an allergic patient was documented as having no allergies because the nurse who was entering data about something unrelated was forced to click through the allergies screen (leaving it blank) when — at that point in her workflow — she didn’t have this information and it was irrelevant to the task she was doing.  This is a more subtle workflow issue and, for the engineers to have built a system that served the nurses (and patients!) well, would have required a deeper understanding of what nurses do and know at particular points in their activities, and when they do and know it.


Analysis of the content of any complex job — not just nursing — to this level of detail is subtle and elusive.  As the article reveals, even nurses know it’s important but are at a loss to explain what to do about it.  Until we learn to analyze and then model the “deep processes” of nursing work, the technology-based tools developed by engineering firms will always come up short.


Blowing our own horn for a moment here, it’s worth observing that at Nurse Tech, Inc., we have staff with rare combinations of skills — both nursing and software engineering — that enable us to build tools that satisfy these subtle and challenging demands.