Expect the exceptions with workflow engines

Workflow engines have been deployed in a number of different industries and professions, from streamlining manufacturing to helping large law firms stay on task. Could workflow engines solve efficiency problems in radiology? Bradley J. Erickson, MD, PhD, suspects they can.

Erickson is a radiologist and associate chair for research at the Mayo Clinic, and has spent recent years evangelizing the benefits that workflow engines can bring to radiology after seeing what such a system brought to his own department.

A workflow engine is a piece of software that helps to automatically execute and monitor a business process. There are a number of factors that separate a true workflow engine from current strategies in healthcare, according to Erickson. The first is that workflow engines can handle exceptions to normal processes that would normally cause other systems to fail.

“Patients don’t read the textbook. They don’t behave like they’re supposed to,” Erickson told Health Imaging. Database-driven workflow systems that are more commonly seen in healthcare use basic scripting language that’s fairly linear. If something unexpected comes up, which can be quite often when you’re talking about patient flow, a workflow engine can handle that exception.

Even a total crash shouldn’t knock out a workflow engine. “A commercial grade workflow engine can recover from the middle [of a process],” said Erickson. “They know what the states of the various steps are and can recover in the middle, even if the computer system goes down.”

Another key component to a workflow engine is a graphical interface where a process can literally be drawn out and adjusted on the fly. Steps are listed within a “time box” displayed on screen, with the system instructed to complete them within a given amount of time, otherwise it reverts to planned out steps to handle an exception in process.

Erickson said this type of construct is hard to do with traditional computer languages, and thus a true workflow engine will use specific business process languages that can deal with the unique challenges involved.

This vision of radiology driven by a workflow engine was realized at Mayo, when Erickson and colleagues deployed DEWEY, the DICOM-Enabled Workflow Engine System. This system is based on an open-source workflow engine that was extended to ‘understand’ medical imaging events. It monitors scanner output and helps automate as many processes in the department as possible.

For example, Mayo uses an algorithm to look for changes in brain tumors, but previously had to rely on people to send both a current exam and old studies from the film room for comparison. If human unpredictability delays a step in the process by not quickly retrieving priors, the workflow chain would be broken.

Erickson says the workflow engine now automatically handles this retrieval, increasing the reliability of the process and reducing grunt work for staff who can then focus on other tasks.

On the same page

Before workflows can be standardized, people have to be on the same page with the terms they use to describe it.

“If you’re going to do metrics, and especially do metrics across an enterprise or between enterprises, be sure you’re talking about the same thing when you discussing performance and optimal care pathways,” said Erickson.

To this end, he’s participated in an initiative from the Society of Imaging Informatics in Medicine to define workflows using a standardized lexicon. It’s been adopted as part of the overarching RadLex lexicon promoted by the Radiological Society of North America, and the next challenge is vendor adoption, according to Erickson.

Standardizing terms and standardizing workflows forces stakeholders to come together and agree on a process. Erickson says this process can have the side benefit of revealing previously unrecognized disagreement over processes, and begins the work of truly coordinating.

“It forces everybody to agree on workflow and that agreement is one critical step.”

Evan Godt
Evan Godt, Writer

Evan joined TriMed in 2011, writing primarily for Health Imaging. Prior to diving into medical journalism, Evan worked for the Nine Network of Public Media in St. Louis. He also has worked in public relations and education. Evan studied journalism at the University of Missouri, with an emphasis on broadcast media.