Ecosystem Master Data Management for Healthcare QA with Gaine CEO Martin Dunn
Why is failure rampant in healthcare data management?
Healthcare is a technology diverse and complex market with an inherent reliance upon interoperability. We know that the key to unlock interoperability lies in good Master Data Management and quite frankly, traditional MDM platforms have failed to deliver the necessary level of support to solve the interoperability riddle.
Managing master data within just a single organization is difficult enough. Multiple internal stakeholders with different data requirements, priorities, opinions, budgets and regulatory boundaries demand a level of sophistication that is often overlooked in the initial design. In fact, most traditional MDM deployments within healthcare are limited to a small subset of data in a single domain and rarely scale to an enterprise level because of limitations in the fundamental design of the solution.
Now consider that the inherent diversity of the US healthcare system requires the coordinated management of data across multiple, interdependent organizations with differing levels of data sophistication – the complexity explodes, and traditional MDM just can’t keep up.
Without effective master data management, it is impossible to reliably exchange data between different systems. This is why the healthcare industry is so challenged and this is why Gaine has been focused on solving MDM across the ecosystem.
What is Ecosystem MDM (E-MDM)?
E-MDM is the new industry gold-standard by which all others should, and will, be measured. It is master data management that extends beyond the four walls of a single organization to integrate the other participants within the ecosystem in which it exists.
E-MDM is an exponential leap over the last twenty years of healthcare master data management. It solves the problems of interoperability by acting as an intelligent clearing house that recognizes – and accommodates – each entity’s view of the data.
Let’s unpack that last statement a little further.
In traditional MDM, the central goal is to align each connected system to the “golden record”. Experience tells us how difficult it is to get multiple stakeholders to agree on a central standard, and even more difficult to get system owners to align their systems to this centralized view. If this alignment is challenging within the confines of a single organization, its exponentially more demanding to do so across multiple independent organizations where each organization may have a different view of the data.
What do you mean by ‘view’ of the data?
Generally, and historically, healthcare organizations approach the exchange of information between systems via point-to-point interfaces enabled through APIs and a common standard, like FHIR. However, this strategy requires each endpoint to adhere to a common set of data standards and to interpret each field in precisely the same way.
This is a flawed assumption.
Consider a basic data field that represents a Providers Primary Phone Number. One organization may use the front desk number of the provider’s administrative office; another may provide the provider’s assistant’s direct line; another may use the provider’s mobile number. All may be considered “correct”, but they are in conflict and someone is going to be disappointed with the result. Most of the time, the meta-data that more fully describes the data is tribal knowledge rather than electronic data in a database. This particular example can be solved with an extended data model that more fully describes the phone number but, and this is a big but, even if every healthcare organization in the country could agree on the data model and standards, they would still need to adjust all their internal systems, processes and nomenclature to reflect an agreed-upon standard.
The idea is completely impractical across all organizations. It’s usually impractical even within the silos of a single, large entity!
The answer is to introduce an integrated, intelligent data exchange within and between healthcare organizations that preserves everyone’s unique internal standards and translates them through a common data model to whatever the other party expects.
How does E-MDM support these conflicting opinions?
For starters, we recognize that not all systems within an organizational ecosystem will be aligned across all elements all the time. Not all systems can, want, or need to update its data, even when some of the data is inaccurate.
For example, if one system only sends payment via electronic transfer, they do not need the PO Box of the billing addresses. Traditional MDM solutions will flag the incorrect P.O boxes and expect all systems to align to the “correct” data. But the owners of said system are not going to spend the time and trouble of fixing the errors because they don’t care about P.O. boxes. And why should they? They don’t use the data and probably have more pressing priorities.
The reasons that systems remain out of alignment are many and varied: competing priorities, budget, technology restrictions, even internal politics. Whatever the reason, over hundreds of data elements and millions of records there will always be a substantial gap between operational systems and the “golden record”.
A second challenge is competing data standards. Whereas one organization uses the providers name from their state license as their preferred name, another may use the “name on the door” which may be different. When these companies exchange data, they will appear to disagree, but this is just two differing standards for the same data.
E-MDM deals elegantly with both situations.
To use our first example. We recognize that a system that doesn’t care about PO boxes shouldn’t be trusted for PO Box data and should not be expected to align with the “golden record” PO Box information. Coperor E-MDM understands this nuance and uses it to make informed decisions on notifications, survivorship and propagation of data.
How can you deliver E-MDM at scale?
The fundamental dead-end many data management projects run into is a lack of scalability. Projects tend to start small with a limited scope and then stall-out when adding new systems, extending the data model, or dealing with more complex workflows.
We are particularly proud of the successful application of Coperor E-MDM for the Symphony state-wide provider directory solution in California.
California is the biggest single healthcare market in the world. In 2015, the state regulators envisioned a single cross-industry platform to enable every provider, health plan and regulator to share and reconcile information in a way that supports complex contractual relationships in a cost-effective way.
Together with Integrated Healthcare Associates, a 501(c) industry association, we built a state-wide Health Data Exchange on Coperor E-MDM. The service is called Symphony Provider Directory Utility and, as of the end of 2020, Symphony delivers a centralized data exchange for the exchange of more than 240 provider data attributes between 14 health plans, tens of thousands of provider organizations, regulators and marketplaces.
What do you mean by a Health Data eXchange?
HDX is a term we use at Gaine to describe a collaborative data management platform like Symphony. HDX applies the principles of E-MDM across a multi-stakeholder community for some defined scope of healthcare data. An HDX is typically not owned by any single subscriber but may be sponsored by a central agency to connect multiple healthcare organizations.
Does this mean you are competing with HIE’s or Blockchain?
That’s a great question. Let me deal with these separately.
HIE’s would certainly benefit from the capabilities of E-MDM, but the deficiencies of Master Data Management within HIE are only a small part of the challenges faced by the HIE movement. Fundamentally we believe that consumers are more likely to control their health records via mobile devices and personal cloud storage – which brings different data management challenges than the ones getting in the way of HIE’s today. At Gaine we are working on solutions that address the consumer-centric control of medical data rather than fixing the current HIE paradigm.
Now for blockchain. As with most shiny-new-objects, most people don’t sufficiently understand the underlying tech to differentiate between practical and non-practical applications. A blockchain is essentially a peer-to-peer network with a unique security protocol that makes interference with the data practically impossible. It’s no wonder that banking is the poster-child of blockchain, where the motivation to avoid central banks is strong, and the data passed between nodes is unambiguous and static. If I send you 1 Bitcoin at 1:00 UTC on 1st April then these data are not subject to interpretation, they are facts that are set-in-stone on the blockchain network. It works great for that.
Now rewind to earlier in this conversation where we spoke about the subjective interpretation of provider data between parties. By avoiding a central server, blockchains also exclude the important role of the central server to normalize the data for its intended audience. That’s strike one.
Another attractive value proposition of blockchains is that they ensure that the data can’t be tampered with between sender and recipient. For applications like provider directories, which are often touted as a potential use case for blockchain, this is solving a problem that does not exist. I’ve never heard of a case where someone hacked a provider roster file as it moved between a provider group and health plan, that’s not the underlying problem – and strike two.
Finally, blockchain only works when everyone shares a blockchain. Look at our experience with HIE and EMR, we’re a decade into the journey of both and there is no line-of-sight to a national platform or unifying single technology. Blockchain does not address the commercial, political and regulatory hurdles that make healthcare data interoperability difficult- and that’s strike three.
Having said all that, we do believe that small blockchains will exist between closely related parties that share a common set of data standards and integrate with other networks (even other blockchains) via a central server. To prove this point, we seamlessly integrated a provider data blockchain into the California Symphony network to show how this hybrid approach can work but that’s material for another time.
As you can tell, there are many subtleties to achieving Ecosystem-wide MDM and paving the path for healthcare data integration. I invite to you read more detail about the Coperor by Gaine Platform and Gaine HDX on our web site and reach out to us via our contact form.
Opt-in with Gaine for More Insight
Keep ahead of the rest with critical insight into Healthcare and Life Sciences MDM and interoperability technique, best practices, and the latest solutions.