Sponsored By

Unlocking the Hidden Value in Imaging DataUnlocking the Hidden Value in Imaging Data

Locked within millions of pieces of imaging data is valuable information for guiding and optimizing care, but it must become accessible and actionable for its value to be realized.

December 10, 2014

8 Min Read
Unlocking the Hidden Value in Imaging Data

Tim Bosch

Like the U.S. healthcare system as a whole, radiology is undergoing fundamental changes. Competition is growing, while reimbursement is declining. Meanwhile, healthcare economic models are shifting toward a value-based system that compensates providers based on outcomes instead of procedures. All of these factors are forcing radiologists—and those who design and manufacture the devices and systems they rely on—to redefine their value proposition for a new healthcare era.

Transformation is nothing new for radiology, a specialty born of technological innovation with the first x-rays in the 1890s. A century later, another technology revolution occurred: the move from conventional film-based imaging to digital imaging technologies. More recently, cloud and mobile computing have entered the arena, transforming the way radiology images and other data are shared. This shift is enabling new clinical collaboration possibilities and new opportunities for medical device makers.

New Models, New Challenges

Progress in the radiology sector presents new challenges. Increasing emphasis on value-based, accountable-care models is forcing device manufacturers and information system vendors to help radiology professionals connect the dots between the services they perform and their contribution to patient outcomes. It is also forcing them to justify costs. Devices, systems, and healthcare providers must deliver and demonstrate value like never before.

Where is this value to be found? Think of the millions of pieces of data generated every day—from digital images and measurements to clinical notes. Locked within those data is valuable information for guiding and optimizing care, but it must be made accessible and actionable for its value to be realized. That’s the next big challenge: How to unlock the full potential of imaging data to reveal new health insights and provide new sources of revenue.

Three Dimensions of Value

Value can be derived from imaging data by striking a balance among the analytics, data management, and interoperability dimensions.

Overcoming this challenge begins with an understanding of what drives value when it comes to data. There are three dimensions to this value equation.

The Analytics Dimension. Data analytics is about gleaning useful insights by modeling data. Gartner has described an analytical maturity model that advances from descriptive (telling you what happened) to diagnostic (why it happened) to predictive (what will happen) and, finally, to prescriptive analytics (how to make something happen). As you move up the maturity scale, the value increases—but so does the difficulty and complexity, increasing costs and time.

The Data Management Dimension. Data management is about making data, including Big Data, readily accessible and sharable. Here, too, there is a maturity curve that ranges from limited peer exchange to centralized processing of structured data to rules-based management of such unstructured data as images and handwritten notes. Ultimately, the curve includes converged data ecosystems that deliver the greatest value. The cloud plays a key role here, centralizing data access while dramatically reducing data storage and management costs.

The Interoperability Dimension. Interoperability involves connecting data sources to enable the kinds of connections that drive new insights. The maturity curve starts with interfacing proprietary datasets at one extreme and advances to data that use a consistent vocabulary and metadata that make structured and unstructured data truly interoperable. While it is more difficult to achieve, moving higher up the scale reduces human involvement and minimizes errors and costs.

Value can be derived from data by advancing in any one of these dimensions. However, the greatest value comes from a combination of all three dimensions. The key is striking the right balance for a given application or objective, as illustrated in Figure 1.

Combining analytics, data management and interoperability can allow manufacturers and vendors to produce systems that enable radiologists to leverage their expertise in new ways. While allowing system manufacturers and vendors to extend their value chain beyond the traditional ‘box and consumables’ or ‘standard reports’ models, this combination allows radiologists to combine patient data with imaging results and observations, facilitating direct clinical intervention. The value of these capabilities is clear—but realizing them is not easy.

Five Questions You Must Answer

To successfully tap into the full value of imaging data, radiology system vendors and manufacturers must address five key questions.

1. Do I know the questions for which I need answers? What insights do I want to draw from these answers?

Start by defining what you’re trying to achieve. Are you trying to generate new clinical insights? Streamline workflows? Improve reporting quality? The answers will dictate what data you need to access. For example, combining imaging with historical patient information provides greater clinical context.

2. What level of data detail do I require? What implication does this have for my data governance and compliance capabilities?

If the required data include personal health information, appropriate safeguards to ensure compliance with the Health Insurance Portability and Accountability Act (HIPAA) and other privacy regulations must be considered. For example, if deidentifying data are required, awareness of FDA’s deidentification guidelines is critical. Will the system be cloud-based? If so, the cloud system or service provider must be party to a HIPAA business associate agreement and conform to all reporting and audit requirements.

3. What are the data sources, and what data transformations are needed? How often must the data be updated?

Combining data from disparate sources enables a more complete view of the patient’s health. However, because these data may exist in a range of formats, they may need to be transformed to render them usable. Access to such unstructured data as handwritten notes from an admitting nurse may require specialized text-search capabilities or natural language processing. Data ‘freshness’ is another critical consideration. Do you need near-real-time data, or is daily or weekly updating sufficient? And what volume of data must be updated? The answers to these questions have profound implications for system design, complexity, and cost.

4. How will I manage the data provenance to address trust and integrity needs?

Data from different sources may have varying degrees of reliability and integrity. For example, a blood pressure reading reported by a patient using a home monitor may not be as reliable as a lab test. Or different labs may deliver test results with varying degrees of accuracy. Knowing what information they can trust is invaluable for clinicians making treatment decisions. But gaining access to trustworthy data may require the use of systems that can attribute the source of each piece of data and even assign it a value ranking based on its provenance.

5. What will it take to maintain the infrastructure to drive the application?

Healthcare institutions and device manufacturers alike are focused on purchasing or developing systems with the lowest possible total cost of ownership. Indeed, this is a key driver behind the move from on-premise systems to cloud applications, which dramatically reduce the cost of maintaining applications and adding new features. Does the system provide the interoperability necessary for the radiology or other departments to connect with multiple devices, provide additional datasets, or share data with those who need them? Both clinicians and medical device makers need to focus on solutions that deliver the maximum benefit with the minimum infrastructure.

All of these questions flow from the analytics, data management, and interoperability dimensions addressed above. By answering these questions, radiology system vendors and manufacturers can determine the right balance among these dimensions to deliver the greatest clinical value with the least difficulty. Getting the balance right can result in exciting new capabilities that add value to radiology and offer a competitive advantage for those manufacturers and vendors that can crack the code.

Move Thoughtfully but Rapidly

Whether you are a healthcare provider or a provider of service and system solutions, realizing the value locked within imaging data requires a thoughtful approach. It starts with understanding where you are today, clearly articulating what you’re trying to achieve, and identifying the necessary steps and transition points to getting there.

Along the way, it’s crucial to make choices that advance your capabilities without getting mired in complexity. It’s not just about how much you improve but also about how rapidly you can get there that counts. Relying on the right architects, data scientists, and informaticists can make all the difference in the world. If these skills are not your core competency, consider engaging an outside partner to fill the gaps and keep you on track.

The data revolution is upon us. Information system vendors and medical device manufacturers must become service and solution providers. Those that recognize today’s historic shift and make the right moves to unlock the power of imaging data will gain a distinct advantage as the healthcare landscape continues to evolve.

Tim Bosch is vice president and chief architect, medical and life sciences of Burlington, MA–based Foliage Inc. Part of the company since 1999, he directs client engagements for a variety of complex medical device and healthcare information technology products and systems. With 30 years of experience focused on system architecture, interoperability, data management, and analytics, he leads consulting and system architecture engagements. He also has extensive software experience with C#, Java, C, and C++. Prior to joining the company, he held technical and management positions at the Codman Research Group, where he was director of new product development. Bosch received a bachelor’s degree from Carnegie-Mellon University, and he has completed graduate coursework in computer science and business at Harvard University’s Extension Division. Reach him at [email protected].

Sign up for the QMED & MD+DI Daily newsletter.

You May Also Like