Solving Issues Around Big Data

As Health Information Exchanges (HIE) become more common there is a greater focus on technology as a differentiating factor for success. A technology advantage is required to solve issues around Big Data, which entails the class of data that combines volume with velocity, variety, and complexity. The challenges with volume and velocity are clear. The number of requests for patient data and the sheer volume of that data are growing exponentially, in communities large and small, whether serving an integrated delivery network (IDN), hospital, or a large physician practice.

Advances in data exchange also have increased the volume of data and clinical communications to the point where physicians are becoming accustomed to having access to vital patient data when and where they need it across their entire community. This volume also creates a greater need for a technology that is able to sift through data quickly—in real time—to deliver what is specifically required and to be able to communicate that data to the provider.[1]

Beyond data access and exchange, another major factor driving growth is the increasing number of providers who are contributing data to HIEs. Even when using data exchange standards, each new provider adds more information variety and complexity. In the past, hospital systems and IDNs produced vast amounts of this complex data and got little in the way of return. As physicians increasingly adopt electronic health records (EHR) and seek to achieve Meaningful Use, the number of smaller practices capable of contributing data to exchanges and giving back to IDNs is set to explode. And this explosion of clinically-produced data will place an extraordinary burden on already stressed technology to transport, analyze, and store critical information.[2]

Not All HIEs are Created Equal

However, not all HIE solutions adequately address Big Data challenges. Insufficient solutions miss key requirements, resulting in inefficient HIEs that put a heavy operational burden on the exchange community. The end result is an exchange that is not widely adopted by the entire community, thus compromising the overall effectiveness of the exchange.

So with the number of potential source systems for patient data growing fast, it is vital to ensure that an HIE be able to deliver the clinical data a provider needs in a timely fashion as a critical component in creating a platform for success both in data exchange and value added services. Physicians participating in HIEs want more options for getting at data outside the four walls of their offices. Such demand for seamless data access also puts pressure on HIE vendors to go beyond scaling for performance and Big Data volume to support new standards for interoperability and content. Moving that data around demands compliance with long time standards such as HL7 messaging for tracking, demographics, appointments, orders, and clinical results; but also places a great deal of emphasis on emerging standards designed to bring rich summary data straight into local EHR systems. Making this happen is a problem for traditional HIE offerings. Rather than creating purpose-built interfaces, vendors are required to provide standardized web services interfaces such as those specified by the IHE or NwHIN Direct as well as even newer mechanisms such as those described by the ONC’s Direct project.

This new set of challenges for HIE technology also includes wrestling with content standards. Bringing clinical information directly into the EHR means agreement on what clinical content looks like. That agreement means that HIE vendors must provide compliance for rich formats such as Continuity of Care Documents (CCDs) that seldom play nicely with older technologies.

As HIE solution vendors over-allocate resources to meet performance and scalability requirements, tradeoffs become evident. Requirements around interoperability, security, and usability are put at risk. Interoperability is particularly challenging for traditional technologies, since the types of information formats in HIEs are often varying, which directly conflicts with the rigid schemas that traditional technologies expect. These tradeoffs also result in slower time to delivery and higher total cost of ownership.

The Challenge with Open Source Big Data Technologies

As mentioned earlier, there is a growing trend of open source Big Data technologies that attempt to address some shortcomings of traditional technologies. The problem now is that while performance and scalability are sufficiently handled, other challenges still exist.

For example, new technologies such as “key value stores” give up granular access to data elements. This means that extracting key points out of reports is difficult. Considering that HIE information consists of a wide range of atomic information (e.g., lab results, medication orders, doctors’ progress notes) and composite information (e.g., documents leveraging interchange standards), the ability to read, search, and report on all levels of information is critical. Also, other technologies like “document-oriented databases” provide no enterprise-class full-text search and transformation capabilities. As with traditional technologies, tradeoffs arise that ultimately lead to inefficient exchanges.

The Right Technology for the Job

Consider an HIE solution built with a technology that was designed for the types of Big Data challenges that HIEs face today. These challenges include aggregating diverse information types, finding specific information in a large dataset, complying with changing formats and standards, adding new sources, and maintaining high performance and security, all while keeping costs under control.
Conclusion

So Big Data then represents one of the big challenges of HIEs now and in the future. For an HIE to succeed, whether in an IDN or a state, it must have the technology to be able to handle vast amounts of data at increasing volumes and increasing speeds. As physicians and patients become more reliant on having clinical information available when and where they need it, the challenges of managing and transporting data will only become more critical. Interruptions to the flow of data will damage the not only the credibility of information systems and client providers, but can put patient health at risk.

___________________

[1] Frost & Sullivan, “Drowning in Big Data: Reducing Information Technology Complexities and Costs for Healthcare Organizations”, July 14, 2011

[2] The Healthcare Blog, “2011 EHR Adoption Rates”, Margalit Gur-Arie, December 2, 2011

John Smith is Director of Communications at ICA. This blog post was first published on ICA’s HITme Blog. John has over 20 years of experience in healthcare communications with a focus on health information technology, having served as Senior Vice President and Healthcare Practice Leader at several communications firms, including Fleishmann Hillard, Manning Selvage and Lee and Brodeur Worldwide.