Why the DMIX Needs MarkLogic

Blood sugar or blood glucose?
Acute mental confusion or delirium?

Language discrepancies like these are but two examples of how clinical notes may differ from practitioner to practitioner – let alone facility to facility. Further, one facility may identify patients through a number system – while another system may use a combination of letters and numbers.

These are just two examples of why sharing patient records is a healthcare nightmare in general – but it is far more pressing for the United States military. Its global health care system includes 1,200 military medical facilities and a population of about 10 million active duty service members, their families, reservists, civilian defense employees and others.

Compounding the complexity, military families often are treated in non-military care facilities. That means a lot more locations, clinicians, patients – and patient notes, tests and records that need to travel from practitioner to practitioner. During war time, the situation can be even more nightmarish – the confusion of casualties and hostilities, the movement of patients from battlefield, to base, to transport to stateside hospitals – all provide plenty of opportunity for information to be lost – with catastrophic results.

Exchanging data between different facilities and multiple types of systems and data models is even more difficult when using legacy platforms. At a joint hearing of the House Armed Services and Veterans Affairs Committees, Rep. Jeff Miller, R-Fla., chairman of the House Veterans Affairs Committee, proposed that all entities use the same software system. And while that may seem to be a logical solution — software is not the issue. It’s the data model. Or, in this case, the myriad of data models (schemas). It would take years – if not decades to mandate a single data model, and even more years to have all the data conform to that schema. And of course there would be all those legacy systems that would need to be ripped out and replaced – which would of course impact workflow and productivity.

So if software won’t be replaced, and we can’t get doctors to all use the same terminology, and we can’t get the facilities to all conform to a single schema – what can be done?

The Defense Healthcare Management System Modernization (DHMSM) program is choosing to build a Defense Medical Information Exchange (DMIX) that allows the seamless exchange of health data between Department of Veterans Affairs, third party hospitals and legacy DoD healthcare systems. The DMIX will provide needed exchange of data without interrupting current processes.

My colleague James Clippinger concurs that the chances of DMIX fulfilling its promise to provide seamless data interoperability will be significantly improved with the use of MarkLogic as its core. James has a long history of trying to get Federal agencies to play nicely together in information-sharing efforts. He told me bluntly, “There really aren’t too many choices for the DMIX. You have to use a schema-agnostic yet schema-aware data management platform that can take in all those different values without disrupting workflow. That platform also needs to have the enterprise features of a traditional RDBMS.”

“But that is only part of the equation,” James added. “This is a massive, distributed, essentially zero-downtime effort that would challenge even the most flexible traditional technologies. One of the biggest challenges is what DoD calls disconnected, interrupted, and low-bandwidth (DIL) networks. You will have disconnected facilities in the combat theater that need to be able to automatically synchronize with the central DMIX when they can get online with no data loss, often over low-bandwidth or interruption-prone connections. Of the few products with proven DIL capabilities, none can approach the flexibility or stability of MarkLogic.”

The same type of flexible schema, enterprise data management platform that powers the exchange for Healthcare.gov.

That massive initiative relies on a data exchange hub centered on MarkLogic. That exchange took just 18 months to build, serving all 50 states, and exchanging data between federal and state-based marketplaces, multiple federal agencies, and every insurance company in the country. Healthcare.gov is the most complex eligibility and enrollment system in the United States and embodies one of the most complex data challenges in recent history.

How MarkLogic Will Help DMIX Succeed

For the DHA, MarkLogic will help create more complete health records, integrate DoD, VA, and third party data, and modernize the software supporting DoD and VA clinicians – at a fraction of the cost of any other solution.

By using MarkLogic at its core, the DMIX will be able to take all those doctors’ notes, emails, lab reports, medical device data, patient history – and any type of images, video, and meta-data from these rich data sources – and make it all readily available to those who need it. Instead of guessing, not knowing, or reproducing costly tests, practitioners can have the most complete view of every patient – to create the best treatment plans.

MarkLogic can consume data in any format and coding scheme in a natural way given its schema-agnostic nature. The data can be easily mapped between all the medical terminologies currently used by the DHA and their partners through a seamless integration with the 3M Health Data Dictionary (HDD). For example, these controlled vocabularies will allow the DMIX to easily determine blood glucose is a synonym of blood sugar, and so on. Doing this transformation and disambiguation at scale – and at high speed – using the 3M terminology mappings is a unique capability that only MarkLogic can provide.

To my colleague James’ other point, MarkLogic technology has also been a key player in the Distributed Common Ground System (for the DoD), providing secure information exchange and supporting disconnected environments with no data loss and automatic synchronization of data when communications systems come back online. Third-party tests have proven it to perform 50 times faster than the legacy versions supported by relational database technology.

MarkLogic can also provide capabilities beyond those offered by today’s Health Information Exchanges (HIEs). Traditional HIEs are focused entirely on patient care and cannot be easily modified to support information exchange with Decision Support Systems (DSS) powering force readiness initiatives. The fundamental challenge of fusing force readiness models with healthcare models to accurately mission-plan is a critical capability required by the DoD. MarkLogic is the only platform with a proven track record that supports the fusion of multiple domain models for decision support across disparate systems.

Let’s take the case where immunization data for soldiers is stored in a myriad of systems. There is an immunization requirement for all deployed troops. Today, because these records are located in so many places, current immunization records for each soldier may be incomplete at the facility where they are being checked for their battle readiness. As a result, there are many cases where soldiers are immunized again because of gaps in the available data. MarkLogic will solve this problem by allowing access to systems that normally would not have access to an HIE.

I’ve only been able to scratch the surface here in terms of the capabilities MarkLogic brings to the table – in future posts, we’ll explore some of these features in more detail, including the MarkLogic capabilities that have made healthcare.gov and DCGS so successful.