HIStalk Interviews Paul Roscoe, CEO, VisionWare

Paul Roscoe is CEO of VisionWare of Newton, MA.

Tell me about yourself and the company.

I’m the CEO of VisionWare. Before VisionWare, I was the CEO of Crimson. I’ve been involved in healthcare technology for the past 25 years in both Europe and the US.

VisionWare was a company I’d known for many, many years, founded by Gordon Cooper, a friend of mine. While I was tracking the company, I also got a chance to see VisionWare from a customer’s perspective because while I was at Crimson, the technology team decided to deploy VisionWare’s master data management solutions to help the Crimson platform.

What’s the definition of master data management?

Master data management is a well-understood genre of technology tracked on a horizontal basis. Gartner has a magic quadrant for master data management, for example. From a healthcare perspective, people may have looked to master data management in terms of technology like EMPIs, or enterprise master data indexes.

Master data management as we define it is the ability for VisionWare particularly to provide an effective and a single perspective on integrating the various different disparate data sets that exist from a healthcare organization — matching, verifying, governing, visualizing that data across these different data silos to provide a 360-degree view of the healthcare data.

The most obvious one of that is patient data, but it could be a 360-degree view of a provider, a facility, or an entity of any description. Patient is the most obvious one.

If you look at the competitive landscape of analytics, where would you position VisionWare?

VisionWare’s technology enables a lot of the analytics solutions that are out there in the healthcare domain at the moment. I know that coming from Crimson that one of the challenges for a lot of the analytics, population health, and care management solutions that are out there is accurately identifying the patient and accurately identifying that patient across the various different care venues in which those types of solutions are being deployed. They are very sophisticated. They have great insight. But only as far as the lowest common denominator, which is accurate patient information or accurate physician information.

We don’t see ourselves as competitors to analytic solutions per se. We have a lot of those analytics and population health vendors that approached VisonWare recently and are looking to integrate our master data management technology to enable a more effective view of the patient information within their solutions.

The obvious example is to look at situations where there are multiple systems. If you look into any health system — never mind an accountable care organization — you will find lots of disparate clinical and financial systems. Organizations are increasingly looking to link those two domains together, so the recipe for mismatched or inaccurate patient data is there.

Now you expand that as you look at the complexity of a health system, not just an inpatient setting, but also inpatient ambulatory. Then you expand that even further to affiliates, employed practices, long-term care, and skilled nursing facilities. You’ve got a very complex picture where the patient’s information is being held. At every one of those venues, there’s opportunity for that patient information to be inaccurate. When I want to lift up and look holistically and longitudinally at that patient, it’s very difficult unless I’ve got accurate patient information.

Clearly disparate systems and the disparate nature of healthcare delivery is promoting this challenge. But even in situations where you’ve got a single EMR system … there was some research done not that long ago relative to the Epic deployment at Kaiser, where it talked about a single deployment of an EMR, but different instances of Epic across different regions. It reported that just within the Epic domain that the rate of patient identity matching fell to somewhere around 50 or 60 percent when they were sharing information across different regions, even within the Epic world. Clearly a single system doesn’t always mean that you’ve also got a handle on effective patient matching.

What’s the cause of mismatched patients within a single system?

You’ve got a number of challenges. The data that’s being collected at these various different registration points is not necessarily conforming to a standard of data governance. How information is collected at Point A on a patient may be very different than the way it’s collected at Point B. How we might use a simple thing as a surname field may be very different from system to system.

There’s really for many of our clients not a lot of data governance standards in place. That’s promoting the challenges of dirty data coming in. You can have the most sophisticated matching algorithms, but if you haven’t sourced the issue at the point at which the data’s being entered, then you’ll always have challenges.

We believe that master data management can be solved to a degree with technology, but it should be part of an overall information governance strategy that health systems are starting to embrace. We are realizing that in this post-EMR era, they’ve got amazing digital assets, amazing data that is locked up in these systems. But without being able to accurately identify that data and to be able to normalize and harmonize it, it starts to lose its value.

When people think about interoperability being the Holy Grail, sharing an IHE profile, HL7 document, or CCDA in itself will not solve this problem because there are still challenges where technology can help probabilistically and deterministically matching these patients together. That’s what we do at VisionWare.

What customer base do you have or seek?

The company was based historically out of the UK. Over the last couple years from the UK, we have been focused on selling to two primary constituents, the HIE landscape and also with technology companies who are looking to provide a master data solution within their own product portfolios. We’ve been successful in both of those areas. We have a large number of HIEs and a number of different technology companies.

Increasingly over the last year since I joined, we’ve now started to focus our efforts on the provider marketplace, ACO marketplace, and the payers. What we’re finding is that a lot of those organizations have the first-generation EMPI technologies. They’re finding that those are somewhat monolithic. They were developed in an era where you only needed to look at inpatient data. That was the key driver.

In today’s world of healthcare and care coordination across this continuum, those first-generation technologies aren’t really fit for purpose. That’s why we started to see quite a lot of traction in the last six to nine months with a solution that was more designed to operate in this more collaborative environment.

Not to keep going on about this, but one of the things that’s quite unique about VisionWare and was appealing to me when I looked at the company is this notion of what we call a collaborative data model. The ability for us to not say, “This is the definition of a patient or a provider, take it or leave it” as some companies in this space do. It’s more, “You give us the data as you see fit and it’s our responsibility to make sure that we can take that data in whatever format, match it, merge it, and send it back to you in the format that you want.” It’s much more collaborative as opposed to predefined.

Analytics companies that are new to healthcare might have missed the concept of patients coming from different venues and different systems without a single identifier. Do you think they are just starting to see the nightmare of what seems simple in identifying a patient?

I think there’s definitely some aha moments for a lot of those vendors, where they realize that they’re taking the data in from those various customers and that they’re responsible for making sure that they can create meaningful value from it. One of those challenges is being able to accurately identify that patient. Yes, we’ve seen quite a lot of traction there.

What we’ve also seen is organizations that have gone through acquisitions. One of our clients is a very large electronic medical record vendor who went through an acquisition of another vendor in their space and wanted to provide a way of quickly having a single view of the patient across these two assets now instead of a single asset. We see that in a hospital setting, as organizations are increasingly looking to either employ practices or merge with other hospitals. That in itself presents large challenges in being able to identify accurate patient data or provide the data across those various assets. So M&A activities tend to be a big driver for us as well.

People may also miss the need for a master provider index and what that means in terms of credentialing or doing any kind of quality work. Is that something that’s also not very state of the art from other vendors?

The first stage of the work that we did with our friends at Crimson was around providing a single provider registry. For any level of quality reporting or performance analytics on a physician, you need to make sure that you’ve got an accurate representation of all of your physician’s activity. Without having a provider directory, that’s challenging. That’s a big area.

What’s the interest in your geospatial capabilities and how that might be used in a public health context?

When you think about data on a patient, we understand data that’s been captured in a hospital or an ambulatory setting. Particularly around patient engagement, there’s a lot of information that is presenting itself on patients – and it will continue to get larger and larger — that might be interesting for a care manager.

The problem you’ve got is that data may be patient supplied or it might be sourced from non-hospital-based systems. Therein lies the challenge. How do you take some of the information from these other areas that a patient’s interacting with that historically hospitals don’t really care too much about? But now as we’re trying to engage the patient or trying to understand how the patient is managing their healthcare, we may take more notice of. There is a challenge there of how you link that information that’s being provided to the hospital information systems.

We have a solution specifically aimed at allowing us to enrich hospital data with third-party data that we’re obtaining or is being obtained by the health system from a variety of different sources. A simple example would be how do you look at an increasing number of self-pay patients? The ability to do effective credit scoring might be important for our health system. How do you link that patient with data that might be in Experian or other credit-scoring system? That’s a challenge. It might seem very simple, but it’s actually quite a big challenge for a lot of healthcare organizations to match that Paul Roscoe with that Paul Roscoe in the credit scoring system without a solution that allows that to happen.

Hospitals have to become more interested in what happens to patients who aren’t having an encounter using more of a CRM-type system instead of just waiting for them to show up. Are organizations interested in using your tools to do outreach for at least targeted groups of patients?

Yes. Not only those cases we talked about, but we’ve also created within the VisionWare portfolio a visualization layer that allows us to visually represent a patient in ways that might be interesting to look at, but you couldn’t get from a flat analytics view that you might get through the dashboard, etc.

If you think about it, we’ve mastered all of the data that’s flowed through the health system. We know the patient. We know the relationship with that potential patient’s family. We know the relationship with the physician. We’re in a great position to be able to then provide a visualization layer that allows you to explore the data in meaningful ways.

You might put this in the hands of a care manager who’s looking at a particular small panel of patients and wants to understand as much as they can about their interactions with the health system regardless of where they are. That’s particularly relevant in an HIE way. You might have access to data now across this broader network. This visualization layer allows you to visually explore the data, potentially on a patient-by-patient basis, and see correlations and data that might not have been obvious to you before.

With ACOs or acquisitions, hospitals are suddenly getting access to data from other systems. Do they have to figure out how a given patient fits into the new grand scheme?

Absolutely. You’ve got situations where you might have a small fragment of the patient record, but the patient is being seen in another facility. Without knowing the connection between that sliver of Paul Roscoe and the broader Paul Roscoe that might be in a medical record that’s being held somewhere else, you may be missing an opportunity from an engagement perspective.

It may be more fundamental than that, maybe patient safety issues. I’m treating this Paul and I don’t really have the longitudinal view of Paul because I don’t have that complete medical record because it’s been duplicated or mismatched. There’s significant impacts to that.

I believe it was the CHIME survey not that long ago in which a fifth of respondents said that there were adverse events happening from mismatched patient information. This is fundamental, not nice to have. There are patient safety concerns that can be addressed by having a more effective handle on your patient and integrity of your patient data.

Where do you think the company’s future lies?

What we are focused on at the moment is building out a larger install base in the US. We think there is a lot of difference between what we do and what the incumbent vendors are doing.

Our job at the moment is to get our name out there. Doing the work that we’re doing with your organization helps. And help health systems understand how our approach is different than the incumbents that are in the marketplace — speed to deploy, the price point that we can offer to our customers in the US, and also just the sophistication of the solution.

Our goal at the moment is to build a strong base in the US. We have a strong UK organization already behind us. That platform allows us to build out our US organization and continue to deliver value for our US healthcare customers.

One of the other areas that we can do is innovate. You’ll see us shortly coming out with a solution which allows us to look at, for example, biometric data on a patient and link back to a patient’s identity. This is a potential Holy Grail of patient identity, which is the linkage of a patient’s biometric signature with the information that’s being stored in the health system. We think we’ve got a really effective way of doing that.