Study Shows How Workflow Visualization Can Enable Big Data Transformation
Big data tools show enormous promise to modify operations in businesses and organizations, boosting efficiency and profits. But as study after study shows, those benefits can be difficult to realize without the proper implementation, expertise, and resources. A new study from Gloria Phillips-Wren (a professor of information systems at Loyola University) and Sueanne McKniff (a senior clinical informaticist for WellSpan Health System) examined one particular technique for convincing organizations to engage in big data transformation: visualization.
“Research has shown that the use of big data can modify operational processes in organizations,” the authors wrote in the paper, titled Overcoming Resistance to Big Data and Operational Changes Through Interactive Data Visualization. “However, little research has been conducted on overcoming resistance to the process changes needed for adoption of big data technologies. … Our goal [was] to demonstrate how the choice of visualization of workflow and operational processes impacts decisions to embrace real-time, big data technology.”
To that end, the researchers built a case study: a partnership between Loyola and WellSpan, a health system serving central Pennsylvania and northern Maryland. Healthcare, the authors explain, is a particularly change-resistant field, owing to the large number of data stakeholders, stringent privacy requirements, and entrenched (if antiquated) systems of information management.
WellSpan’s orthopedic group, the authors said, was using an open-platform electronic health record program called Allscripts with features like prescriptions, mobile access, care guidelines, patient flows, and decision support tools. “In actuality, however,” they wrote, “the orthopedic office had developed workflows to circumvent the EHR platform, including the use of dictaphones requiring post-transcription and later data entry with the belief that these interventions were more efficient for physicians and patients.”
The researchers worked with a professional process flow specialist to observe and record how WellSpan’s orthopedic group navigated its daily workflow. Along the way, the researchers encountered staff modifying their behavior in response to this observation, contaminating the data and – because the performative attempts were largely unsuccessful – reinforcing the organization’s distaste for more data-intensive tools. Eventually the researchers embedded trained clinical staff in the workforce, gaining access to uncontaminated insights.
Those insights revealed that physicians remained reticent to use the technology due to an impression that it reduced face-to-face time with patients. The researchers attempted to communicate the value of the big data tools by presenting the physicians with a graphic of their current workflow versus a streamlined, big data-enabled workflow. The physicians were unconvinced, citing the “acceptability” of the current workflow and the need for more assistance than the revised workflow provided.
So the researchers tried a new technique: interactive visualization. The researchers allowed the staff and physicians to walk them through the workflow as it unfolded on the screen, allowing a greater sense of ownership over the resulting flow chart – which, of course, ended up resembling the revised workflow initially presented to (and rejected by) the physicians. The staff responded positively to the interactive visualization and the resulting workflow, and some months later, the researchers found that implementation was progressing smoothly.
“The study demonstrates that interactive data visualization of operational processes can be an enabler in overcoming organizational resistance to big data technologies in a change-resistant organization,” the authors concluded. “The concomitant benefit is that big data analytics is placed directly into the hands of primary decision makers.”