Guest Column | October 26, 2017

Mining Data For The Future Of Radiology

Combining Human-Centered Design And Big Data In Pharma

By Kent Thomas, Vice President of Solutions and Business Development, vRad

In 2012, vRad embarked on a data analytics initiative to better understand and manage our practice.

The good news was we had access to a massive amount of data—about 18,000 daily reads across a geographically and demographically diverse footprint from 500 radiologists in all 50 states.

The bad news was the data was raw and dirty. For example, within our platform, there were more than 15 ways in which our client sites identified a CT of the head.

After an extensive effort to build an analytics engine that could standardize and normalize the data, we had the preliminary tools necessary to carefully examine our practice. Since that initial exercise, we’ve augmented our capabilities, adding meaningful attributes and even Natural Language Processing (NLP) to turn that data into useful information.

We’ve made operational changes and decisions aimed at making our practice—and those of our clients—more accurate, efficient and profitable.

No longer shooting from the hip, using anecdotal information, we proactively manage our practice with a firm grasp of the issues and forces affecting it.

It’s a model that, now more than ever, is the future of radiology.

The Value Of Insightful Questions

The greatest value of a robust data analytics program lies in the questions it spurs a radiology practice to ask—and answer.

Do we have inefficiencies?

The most common “a-ha!” moments for practices and health systems who initiate a data analytics program are related to how inefficient their practice is.

Are all of our techs taking lunch at the same time? Why are we fully staffed until 4:30 when our imaging trails off by 2:30 every day? Do we need another outpatient slot on Saturday mornings?

Is an investment in technology warranted?

Decisions regarding expensive technology can be even more costly if they are made without actionable data.

Are we sure it is the right time to add a modality? Is it the right modality for our current needs? Are we conducting the right kinds of studies on each of our existing CT scanners? Are there any peaks and valleys in our scanning activity that can be leveled out through more efficient scheduling?

Are we aware of our opportunities and vulnerabilities?

Using only anecdotal information, crises have a way of sneaking up on an imaging service line, while opportunities for growth can be easily overlooked.

Why is the smallest hospital in our system doing a higher volume than our second largest? Healthy women’s imaging tends to be 8 to 12 percent of the volume of a typical radiology service line. So why does our large health system only garner a 5-percent volume?

How effective is our professional staff?

Your professional staff is the engine that runs your practice. Have you looked under the hood to know how effectively and efficiently it is running?

How does Relative Value Units (RVU) generation differ across our radiologist team and why does that matter? Can we share the best practices of our most productive radiologists? What outside influences affect radiologist productivity, and where?

Are we making intelligent staffing and scheduling decisions?

Embedded in the data of your practice’s reads is a wealth of information that holds the key to better staffing and scheduling decisions.

What kind of imaging are we doing and what are the results of the studies? Was the imaging valuable and how might the lessons we learn from those studies shape the future ordering habits of referring physicians? Can we use historical data in our triage activities and in staffing and scheduling to promote better clinical outcomes?

Who Is Managing Your Practice?

The beauty of numbers—especially in the context of questions like those above—is that they don’t lie. But their exceptional value is muted when radiology service lines don’t have someone to truly manage their practice.

Do you have someone whose primary responsibility is to take a detailed look at your radiology service line? Just as important, how far out is that person looking? Are they simply planning tomorrow’s schedule or are they looking toward the modalities, patient populations and initiatives of 2019?

It may be tempting to answer that you have a department manager or a director of imaging that manages your practice. But we all know how busy those people tend to be with the day-to-day department activities.

The vast majority spend time putting out fires—making sure that consumables are ordered, that all modalities are up and running, that vendor relationships are strong and that staffing issues are addressed.

Take The First Step

Anyone hoping to develop a useful data analytics program, therefore, should start with a person devoted to managing your radiology service line. If that’s not possible, seek a partner to help make it possible.

At vRad, for instance, we provide data analytics to the practices, facilities and health systems we serve as a value-added component of our radiology services.

Once that person or partner is in place, try to determine what information you have today. What kind of analytics are you able to generate and what does it tell you? As you dig in to this rich information, begin to ask those insightful questions, recognizing that the answers can be identified within.

And know that with those answers come the building blocks to a better, more successful radiology service line.