Radiology’s new data plan

Walking through the exhibit halls of the Radiological Society of North America’s annual scientific show may provide the ultimate technological shopping experience for diagnostic imaging and information system clinicians and administrators.

Aside from the bells and whistles, device convergence and equipment modularity options that have progressively emerged through the years, providers should be able to spot one convenient, over-arching theme woven through the educational sessions and product demonstrations: “Too much is never enough.” In fact, contemporary imaging technology collects more data on patient conditions and care process workflows than ever before.

But is sheer volume enough to make healthcare delivery more effective and efficient in the long run? Or is it merely a short-term stalling procedure until technology catches up to questionably unrealistic patient care demands? Some contend that improvements in data acquisition must partner with meaningful post-processing analysis and measurements to help convert data to relevant and useful information for patient treatment.

It remains to be seen whether diagnostic imaging equipment suppliers, healthcare IT professionals and radiologists are heading in that direction.

To gauge industry interest, Health Management Technology reached out to several imaging technology experts for their insights.

HMT: Do you believe that healthcare facilities collect enough clinical, as in imaging, data on patients to make as accurate and effective a diagnosis and treatment plan as possible?

Patti: No. Individual healthcare facilities do not collect information about the patient’s complete radiologic imaging history. Typically, institutions keep records pertaining to the patient’s prior exams at their own institution. Given that patients are highly likely to receive care from multiple institutions, being blinded to imaging data performed outside the institution can lead to sub-optimal diagnosis and treatment planning.

Celebi: In terms of collecting enough clinical data, my answer is “yes.” The bigger question is how to complement clinical data with other types of non-clinical, patient and population data in order to make more in-depth determinations about personalized diagnoses and treatment plans.

Tolle: In general, the industry is still very transactional rather than episodic when it comes to patients. While there are some instances that a complete picture of a patient’s clinical data is available to drive better decisions, there’s still a gap for many in the industry. For example, many imaging referral and order processes are still very manual. Sometimes a patient walks in with a paper order or the imaging center receives a fax, and sometimes the order is sent via EHR. Typically with first-time referrals, there’s no transparency to an existing medication list or of the patient’s medical history. This ends up putting a lot of work on the radiology technicians. They fill out a form when working with the patient, then scan it into the document management system so the radiologist has a more complete picture of the patient. The information is there in the end, but it takes a lot to get there.

Kolb: Yes. That being said, it is often a challenge to get all of the historical clinical information from the patient to the radiologist. What is often missed is prior imaging and report data. The complete medical record is critical to maximize the accuracy of the radiologic interpretation. Our staff is often fighting to get that information from our clients so our radiologists can read the imaging study thoroughly. The imaging providers have good processes, but the patients often miss the importance of this aspect of their treatment. Having good systems in place to easily upload and transfer this critical information is paramount in maximizing accurate outcomes. The aforementioned state-of-the-art equipment is plentiful and ever improving, but radiology interpretation is often telling the medical story and developments of the patient. Without critical prior data, chapters are missing and the story is incomplete.

Aharon: No. Although new technologies and integrations are enabling facilities to collect and have more clinical data, there is still more that can and should be done to enhance the precision and effectiveness of diagnosis and treatment.

First, enough attention is not given to the types of non-clinical data that can be collected. Such data is still very complementary to clinical data and instrumental to improving the quality of care. Specifically to the realm of CT, radiation dose is an issue that is surging in awareness among facilities and patients alike. Imagine the higher quality of care possible if facilities across the country leveraged dose-monitoring solutions to collect radiation dose data and reported to the Dose Index Registry (DIR), a data registry where they can compare their CT dose numbers to regional and national values. Once the data is accurately captured and documented, it can be used to ensure that low-dose imaging protocols are used for increased patient safety.

The benefit of this can be seen from facilities that have already adopted this as a best practice. The industry as a whole could collectively manage CT dose for patients and deliver much safer care with confidence. Several healthcare facilities across the country are showing how effectively this can work, taking advantage of the DIR to establish the standards of low-dose imaging they must achieve and working with vendors to integrate the appropriate solutions.  The results: These facilities quickly saw themselves become among the top-ranked low-dose providers within their respective regions for delivering the lowest cumulative dose levels. Now it’s time for others to follow suit and for new policies to eventually mandate such data reporting to provide further support beyond the information provided by clinical data and truly bolster the effectiveness of patient care.

Sinclair: Let me start by saying that I’m not a doctor, and I have no hands-on expertise when it comes to making diagnoses or creating treatment plans.

What I can tell you, however, is that there’s a larger body of health data that goes beyond the specific tests a physician might order or the questions they might ask a patient about their symptoms and history. The health records at the physician’s fingertips may be incomplete, not because the information hasn’t been recorded, but because it’s scattered or siloed elsewhere. Multiple records of care often exist – both inside and outside the healthcare organization – and too often they aren’t all available to the physician or team making the diagnosis and creating the treatment plan.

Even when plenty of data is being collected, issues remain. How is that data being aggregated, normalized, analyzed and delivered to the people who need it, when they need it, in forms that are understandable and actionable? How do you dive into a mountain of data and come back with the specific insights needed to create a treatment plan that addresses the long-term as well as the short-term needs of the patient – and long-term as well as short-term costs?

That’s where the right technology can make a big difference. Sophisticated, fully interoperable population health management solutions can aggregate and analyze health data from diverse sources, including patient demographics, EHR systems, pharmacies, labs, claims records and more. Most importantly, they can then deliver actionable information and insights derived from that data to the physicians, specialists and administrators who need it, when and where they need it.

Being able to do that is becoming more critical every day, not only to improve patient outcomes, but [also] to ensure the financial health of provider organizations as value- and outcome-based reimbursement models continue to displace traditional fee-for-service payments.

Kao: Exciting new 3-D imaging modalities provide enhanced visualization for a variety of diseases and conditions. In addition, new tools are available that can, for example, provide more precise lesion measurement to help enhance diagnostic and follow-up activities.

The challenge – and it’s a big one – is to deliver patient data and previous imaging studies quickly and easily to radiologists and physicians so they can use that information when making diagnostic decisions.

Embedding PACS into an EMR can make patient data and previous studies immediately available to radiologists and physicians. In addition, Carestream’s Clinical Data Collaboration platform offers a new voice-recognition tool that enables radiologists to insert hyperlinks into the final report that can contain key images and quantitative comparisons such as vessel analysis and measurements to aid clinicians in treatment planning. Embedding a zero-footprint viewer in the EMR allows remote physicians to conveniently access 3-D/MPR images, multi-media reports and non-DICOM videos and photos from their mobile devices or workstations. Simplifying and speeding access to patients’ medical records has the potential to significantly enhance diagnosis and treatment.

Morgan: The total amount of clinical imaging procedures is correct. The use of the most effective imaging procedure and access to all prior clinical imaging could be improved. [Health Information Exchanges] are attempting to be the broker for prior clinical history but are still in the early stages for imaging.

Primo: As we’ve learned in the news about the domestic Ebola case in Dallas, all patient information is important, including travel history. The good news is that in North America medical imaging data are acquired sufficiently, and diagnostic reports are available. At the time of the diagnostic reporting, however, there are questions.

Access to prior examinations can be challenging for the radiologist if prior exams were performed by another hospital or [diagnostic imaging center]. CDs containing imaging studies for importing these data can be lost or broken, or not available at all. Is all clinical information of the patient accessible for use by the radiologist? Knowing the anamnesis or hetero-anamnesis is indeed important. Further, family history may point to the patient’s predispositions for developing certain diseases, which can cause the radiologist to look for certain radiological signs that are typical for these illnesses.

Having all the information above could lead to more insightful and personalized diagnoses.

Further, the value of the anatomic data in the images can be increased by the use of advanced quantitative imaging techniques, such as extracting numerical data from the images. The [American College of Radiology’s] Imaging 3.0 initiative recommends quantification techniques to become part of the standard diagnostic toolkit. Quantitative imaging can lead to a deeper insight into the disease, including the response to treatment over time, which in turn could lead to more personalized treatments

HMT: What might they be missing, and what more do they need to collect?

Patti: Institutions are beginning to recognize the value of having a complete and accurate radiological history of their patients. In payment models where compensation is based solely on volume, repeated or unnecessary examinations can actually be a source of revenue. In payment models where compensation is based on quality of service or outcomes, unnecessary or inappropriate examinations are a cost to the institution. 

Celebi: Although the facilities collect a sufficient amount of clinical data, we believe there is an opportunity to collect more data regarding the care delivery process, including health system engagement and non-clinical characteristics of patients such as patient engagement information, care-seeking behavior, behavioral patterns and consumer information. With advancements in genomics, therapeutics and targeting technologies, we are quickly moving toward personalized diagnostic and treatment plans, and these types of data are critical for these types of efforts.

Tolle: We’re not too far away from a time where referrals can come from a physician’s EHR with patient history, including their current medication list, allergies and other problems, which would make it easier for radiologists to leverage data to its fullest extent and provide insight for better diagnosis and treatment plans.

Physicians are relying on patients to tell their own story and bring their own information. Often times, patients aren’t prepared, especially when they’re possibly dealing with a life-threatening illness. There are likely, pertinent details missing from the overarching story.

From an imaging perspective, most organizations find it hard to longitudinally manage and look at a patient’s imaging history because the data is in a PACS or [vendor-neutral archive] – it’s too disparate. When we’re at a point where all patient data is accessible in the EHR, there will be one centralized place to find pertinent information, including radiation dose details and patient imaging history, which creates an easy way to alert providers of an issue and gives providers more data to make more informed decisions.

Sinclair: There may well be additional information that needs to be collected, but the real question is how you figure out what’s missing.

Doctors may be able to answer that in some cases, but often they won’t be able to because they don’t know about everything that is being collected. Healthcare organizations are already collecting vast amounts of raw data on patients, and much of that information is inaccessible or inadequately used. As a result, physicians may request the collection of information that already exists, or order tests or procedures that aren’t necessary.

That makes it crucial to start by using technology – and specifically, sophisticated population health management solutions –
to aggregate and analyze the data that already exists. This not only enables clinicians to access and use the information that does exist, it also makes gaps in the data – places where more information needs to be collected – more apparent.

Kolb: Prior patient data is often incomplete and/or missing. Systems [that are] able to easily gather and transmit this information are a key factor to ensure thorough radiologic evaluations. Use the simple example of a mass. Regardless of where the mass is located in the body, if there were previous images and radiologic interpretation of the mass, seeing the development of it is critical for the doctor to analyze its current state. Understanding patient history contributes to the radiologic assessment as well.

HMT: Do you believe the next technological advances should focus on collecting additional and different kinds of data, or should they concentrate on diagnostic measuring and planning tools to analyze and interpret the amount of data currently collected?

Patti: If we need to collect and store a complete and accurate radiologic history for patients, there need to be robust interoperable technologies supporting HIE frameworks so that the ever-increasing mass of data is stored efficiently, and access to that data is readily available to radiologists at the time of service.

Collecting the data and making it accessible should not be considered success. Imagine if one collected a phone book worth of data but did not spend the time to categorize the data. While the data might be present, complete and accurate, the usability/impact of the data would be low. End-user tools should be produced to semantically parse that mounting data. This can be accomplished both by traditional natural language processing and a corollary process applying logic to and detecting consistent attributes of 2-D and 3-D imaging data sets.  

Celebi: We believe that the next set of advances will focus on making better sense of the clinical data currently collected and combining that with patient information collected from emerging sources, such as wearables and remote monitoring devices, healthcare social networks and consumer data, etc., to create a better context for the clinical information.

Tolle: Both collecting additional and different types of data as well as diagnostic measuring and planning tools are equally important, but it depends on the market and the specialty you’re looking at.

For instance, take eye care. In eye care imaging, advances in computer-assisted detection of diabetic retinopathy or glaucoma would bode well for early detection and better patient care, plus it could have a massive effect on the costs.

In orthopedics, there are a lot of advances happening right now around case planning technology in the cloud. Technology like this would help to make it as easy as possible for a surgeon to know all the latest and greatest implants, stay up to date on the science and plan well to ultimately improve patient care.

Overall, we should be more focused on the nuts and bolts of the transactions between the different players. We can’t pretend there’s unlimited money in healthcare, and I would love to see vendors focus on solving the problems around moving data more effectively. If you can do that, you can really get down to comparative effectiveness measurements: What is the value of a lower spine MRI? How do we know it’s worth it? We don’t know until we have a full patient history on several patients to understand which one was more cost effective and had better outcomes.

On the payer side, there are many pre-authorization rules that determine whether or not procedures will be reimbursed, and most of the time imaging providers have to rely on very labor-intensive, manual processes to ensure compliance with these rules. 

There needs to be an effective standard for transmitting referrals and orders from one point to another, and also a standard for how to document and deliver clinical findings. This way, creating effective advancements around diagnostic tools will be that much more targeted to leverage data and use it for improved care.

Sinclair: The primary focus now, and for the foreseeable future, needs to be on solutions that aggregate, analyze and interpret the data that’s already being collected. And from the standpoint of healthcare organizations, the up-front focus needs to be on identifying the right technological solutions to do that. Some of the population health management software being sold right now has been developed by EHR vendors, and understandably it’s closely coupled to their own EHR software. That might seem fine if it’s your own EHR vendor, but it really isn’t.

Your population health management solutions need to be fully interoperable with all the systems you need to draw data from, now and in the future – including other EHR systems, lab and pharmacy networks, and more.

In addition to interoperability, look for modern, cloud-based solutions from a vendor with a strong record of client support and service. Cloud-based technology minimizes up-front costs and implementation time, and virtually eliminates maintenance costs.

Kolb: I would suggest the latter with the thought that it’s not about collecting more data but all the data. NRAD often received prior images and reports after the initial read occurred. When this happens, there are often diagnostic changes to the report based on the additional data. There are two aspects to this:

Get all the data. As simple as this sounds, all radiology firms are forced to interpret studies “with limitations” because all the data is not readily available. Patient education will help, as well as systems that enable a smooth flow of this information. Often, patients have received studies at multiple locations, which in and of itself poses challenges.

Once the data is collected, analytical tools enhancing radiologists would certainly be helpful. The radiologists are trained to make the final diagnosis, but analytical tools, especially for patients with a long history of imaging, would enhance the process. Imaging advancement should never be slowed. However, I seldom hear radiologists complain about image quality. More often, they comment on the lack of patient history. 

Kao: Both patients and physicians are best served by continuing to develop new and better ways to visualize diagnostic information as well as provide new measuring and planning tools that can enhance data interpretation. Both of these efforts are equally important and can have a dramatic impact on patient treatment decisions and outcomes.

Morgan: Clinical decision support is being implemented to assist in ordering the proper imaging procedure. While EMRs are collecting large amounts of data for all areas (labs, pathology, radiology, cardiology, oncology, etc.), a more user-friendly interface and graphical representation of this data based on the user and reason for access is needed.

Primo: Collecting more data, relevant to the patient’s actual condition, like anamnesis, family history and genomics, in an automated fashion will be important. Smart indexing techniques will have to be implemented to automate data collection from many disparate information sources in the ACO.

Further, access to national databases and registries with relevant information on a specific disease will have to be part of the information collection and presentation policies and techniques. Many databases are already available, but consulting these databases is often a manual process. The content is mostly unstructured. Integrating automated data collection from these databases in the radiology workflow could add value to the accuracy of the diagnosis, but a prerequisite will be that these databases’ contents need to be converted to a structured and machine-readable format.