ubmslateDI-logo-ubm

DI Mobile Menu

Search form

Topics:

Strategy Behind Watson Health’s Medical Imaging Collaborative

Strategy Behind Watson Health’s Medical Imaging Collaborative

Last month, IBM announced the latest addition to its suite of Watson artificial intelligence technologies – the Watson Health medical imaging collaborative. This global partnership brings together 16 vendor, health system, and academic partners with a shared goal of improving cognitive imaging for radiologists and referring providers in a myriad of specialties.

When fully deployed, the collaborative will have enhanced imaging in heart health, cancer, diabetes, eye health, brain disease, and many other conditions. Diagnostic Imaging spoke with representatives from two collaborative partners – Nancy Koenig from Merge Healthcare and David Lehr from Anne Arundel Medical Center – to discuss the impetus behind the burgeoning partnership and what its impact might be.

 

Nancy Koenig, Merge Healthcare, General Manager

Diagnostic Imaging: Why launch the Watson medical imaging collaborative?Nancy Koenig, General Manager, Merge HealthcareNancy Koenig, General Manager, Merge Healthcare

Koenig: As with any breakthrough in technology, it requires a series of talented individuals and forward-looking organizations to come together to bring the breakthroughs to market. That’s the purpose of the collaborative -- to deliver cognitive solutions to market. You need to focus intently on identifying proper data sources that can be used to train the Watson data health engines. Involving and driving together clinical and user designs can help the collaborative bring the new breakthrough to market. That’s the purpose and mission behind it all.

DI: What are the intended benefits?

It will be a benefit to bring together the best and brightest in the industry to have the situation evolve that everything comes to market and that we can focus on health care and how to deliver it with a cognitive solution that will help with the diagnosis of cancers and eye health. The benefits are working with providers with data sources to train Watson’s engines in these use cases and to bring together preeminent clinicians and provider staff to help design the solutions properly.

DI: What will each of the partners be required to do?

As we focus our efforts, our first area of concentration will be on heart health and, then, addressing the diagnostic processes of certain cancers. As we get into the working group of the collaborative, people will identify which use cases are most germane to their patient populations and service lines. They’ll want to focus on those, but that’s not determined yet.

The members of the collaborative have been identified and have entered into an agreement. Project agreements on the next steps are in progress right now. There’s no timeline for when they’ll be completed. They’ll be evolving and will continue to do so. We plan to have project agreements in working place this summer.  

DI: What are the goals at this point? Benchmarks for how to reach them?

At least in the initial phases, approaching health care is a huge topic. We want to be clear that we’re not saying that everything will be wrapped up and solidified this summer, but negotiating and working out project details are paramount. There are several ideas that are being discussed among the collaborative members to date. They’re working on potential initiatives to break down the barriers in interpreting information that will be loaded into patient systems, such as lab systems. In our industry, radiologists themselves are faced with an overwhelming amount of data on their patients that they’re asked to interpret it at increasing speed. To aid their efforts, we need to come to market with solutions that help them digest that information and assist in the diagnostic process.

DI: What will the impact be on radiologists? Patients? Referring providers?

At Merge, we’re focused on physician efficacy. We think about what’s going on in the industry today. The number of images ordered is increasing and reimbursement is declining. We’re asking our clinical professionals to do more with less time. Our focus is on efficacy, and that’s not just efficiency. It’s also helping them with the diagnostic process through products. That’s the intent behind our statement of bringing cognitive products to market.

Right now, the collaborative itself is very focused on the physician. The benefit to the referring physician is that the process is delivering results to them faster and improving the quality of the studies they get. That’s the intent behind the initiative.

The early application of cognitive solutions will help in the interpretation process by digesting large amounts of data. Cognition at its heart and soul is the ability to consume large amounts of structured and unstructured data. Merge and its collaborators will deliver the ability to digest, absorb, and help with the process of imaging. By digesting and absorbing that information, we’re hoping we can positively impact the early stages of diagnoses. Post-diagnosis, the collaborative will help transfer some of those strategic findings into the report itself. The intent is to alleviate the redundant processes that physicians and clinicians go through and help them focus on getting the best possible diagnosis and the best possible outcomes for the patients. Everything that we’re doing in terms for bringing cognition to market is a breakthrough. All the applications that are being introduced don’t exist on the market today.

DI: What will success look like within a year? Five years?

If we think about the next 12 months first, we have two products slated to be delivered -- heart health and valve disease, and the second will follow shortly thereafter. It will focus on cancer with our priorities being breast and lung. In five years, we’ll see multiple applications on the market and extend the platform to address different diseases to improve patient lives and reduce the cost of health care worldwide.

Everything is a work in progress. For example, one in process right now is with a melanoma application. You take a camera like the one on your iPhone and take a picture of a mole. The Watson engine can look at the picture and suggest, based on its training and imaging data, whether that skin mark is likely to be a melanoma. The application isn’t diagnosing, but it’s indicating to the physician that the mark is something they might need to take a look at and that the probability of the mole being X or Y is high. The physician will then examine the mole accordingly.

 

David Lehr, Executive Director of Analytics and Report, Anne Arundel Medical CenterDavid Lehr, Executive Director of Analytics and Report, Anne Arundel Medical Center

DI: Why did Anne Arundel Medical Center get involved with the Watson medical imaging collaborative?

Lehr: This has been an ongoing initiative for Anne Arundel Medical Center. We’ve made a strong investment in technology and analytics simply because, for us, it’s the strategy for providing the best possible care to our community. If we want to provide the best care to our community today, this technology is important – and it’s even more important if we want to be in the position to provide the best, most efficient care for our community five or 10 years from now. We need to be an active participant in shaping the future of how health care will be delivered. This is one more step in an ongoing initiative to use analytics, data science, and technology to advance and position ourselves to help move health care to where it needs to go.

DI: What will Anne Arundel’s involvement be? How do you intend to participate and implement?

We’re participating in a bunch of different ways to provide expertise. Especially for us, we’re focusing a lot on making sure Watson is able to be used for real-world health care problems in community care. We don’t see this as a one-off research project. We see it as something we’d like to use with every patient to improve the quality and efficiency of care that we can provide. Expertise from our community-focused physicians and radiologists – to help bring that perspective to the IBM Center of Excellence with our data. Machine learning has become particularly important, and the data must be well-rounded. It can’t just be from specialty patients for specific initiatives. It’s important that our entire community be represented in the data algorithms that the technology comes up with.

We have lots of repositories that we use, and we’ll be de-identifying the data to protect the privacy of our patients. We’ll train our staff with these algorithms so when they see new patients, they’ll be able to understand their histories. For example, just think of the medical record and what it looks like for a complex patient. You might be imagining binders full of data. Since the advent of the electronic medical record, there’s even more data added to the medical record. Devices are streaming information minute-by-minute during patient encounters – time stamps can be collected anytime there’s a clinical interaction with a patient. If a radiologist jumps into the middle of the patient episode of care and hospitalization and they’re being shown a film of a patient that’s having pain, how are they supposed to pull out the most relevant specific things in the patient’s chart? We see Watson as being able to understand the context under which the imaging order is placed and present the radiologist with the most useful information to guide him or her during reading the film.

DI: What benefits does a collaborative like this provide?

This initiative – and those like it – we envision, will lead to better and more affordable care. We see technology being used to identify quality improvement opportunities and performance improvement opportunities. But, at the end of the day, success will be a healthier and happier community.

We haven’t yet gotten to the specific level of how the collaboration will work, but it will involve a discussion between us, IBM, and a bunch of other organizations. There will be agreed upon parameters and responsibilities that will be broken up into sub-projects. We’ll have the opportunity to be involved in one or more sub-projects.

DI: In what ways will the initiative be implemented?

It really comes down to reducing barriers in communication and bringing community focus to the group. A lot of folks will bring insight from research and specialty care perspectives. Anne Arundel Medical Center’s role – we hope – will be to bring strong community perspective for a wide population of patients. We want to focus on population health and what we can do to improve the treatment of our patients.

I think one of the major things we’ll focus on is breaking down the communication barriers through the patient’s chart. During the course of treatment for a patient with a complex disease, there are quite a few different physicians and nurse practitioners who are documenting in the chart every day. There are some things they need to have communicated up front so everyone on the care team can see it and benefit from it. And, there are things that don’t need to be read every time. I think bringing the expertise of all these different contexts to facilities will improve the flow of information among the care team. It’s one of the key things we want to focus on.

DI: What type of impact do you see the collaborative having on patient care or a provider’s ability to offer care?

It can all lead to fewer errors and lower costs. Fewer duplications. Less waste. Fewer things that are just being ordered even though they already exist. Every time you introduce a new intervention or medication to a patient or put a new chemical in their body or perform another test, there’s a risk. The extent to which we can cut down on unnecessary testing and medications, we will. We’ll look at things that have already been tried, but not communicated, things that need to be re-done, the extent to which we can cut down, how we can improve patient safety, the overall cost of care and what will help the patient.

Certainly, every time you improve the flow of information for physicians, you’re increasing their efficiency. Things are less frustrating, and there’s time to search through a patient’s chart to find things they were looking for. It’s not always easy to find things in a mountain of information that’s being stored at an increasingly rapid rate. So, facilities that can create that flow will make providers’ lives easier, as well.

DI: How will this collaborative impact the use of imaging?

The more information that can be communicated in context-specific ways, the less likely there will be any duplication of imaging. That duplication reduction focuses on ordering, and it’s going to affect radiology ordering patterns the same as anything else, like a lab test. For the radiologist, beyond just the natural language processing, there are things that IBM Watson can do to help assist them, such as making measurements and identifying things the radiologist should be paying attention to.

DI: Has there been any feedback from within Anne Arundel Medical Center?

Dr. David Todd, our medical director of radiology, is the other key person involved in participating in the IBM Center of Excellence. The radiology department is leading the charge on this. It’s not a technology initiative as much as it is a quality initiative.

In the back of our minds, we have ideas. We haven’t vetted them yet internally to the point of being able to share them, but once we get together with all the participants, we will set goals for each project.

DI: Why is it so important to have something like this collaborative from a medical industry perspective?

If we really want to zoom out on the health care landscape as a whole, we see an industry that’s growing in terms of total cost much faster than the GDP. That’s not sustainable. Clearly, you quickly realize that, if you just talk to people, that it’s important for communities to be healthy. It’s important for care to also be affordable. Using the technology at our disposal to take better care and offer more affordable care isn’t just a good idea – it’s our responsibility as an organization. So, using machine learning algorithms like IBM’s Watson for identifying quality improvement and performance improvement opportunities, as well as waste reduction opportunities, is a key strategy that Anne Arundel Medical Center is employing. If we’re successful, we’re going to have a healthier community.

 
Loading comments...

By clicking Accept, you agree to become a member of the UBM Medica Community.