In order to deliver the right patient data to the right providers at the point of care, health information exchange (HIE) organizations must ensure that their patient matching process is as efficient and accurate as possible. To this end, New York City-based Healthix, Inc., the largest public HIE in the country, is taking steps to improve accuracy and efficiency in identifying and linking patient records across its platform.
Healthix matches patient records from 300 disparate sources for more than 16 million patients and delivers health information to hundreds of participating organizations spanning nearly 1,400 facilities across New York City and Long Island within its HIE platform. With the patient’s consent, Healthix enables clinicians to view their patient’s composite record, receive notifications when an important event happens and support population health management.
Healthix recently announced it would deploy a cloud-based patient matching platform developed by health IT vendor Verato to boost the matching performance of its existing technology. Healthix currently has a state-of-the-art patient matching process using IBM’s master data management (MDM) application, according to Thomas Check, president and CEO of Healthix. However, patient identification resolution and records matching is an ongoing and complex challenge for HIEs and healthcare organizations given disparate sources of health information and the different ways that organizations record patient identification information.
“One challenge when linking patient identity from disparate sources is that two different sources could record the information from the patient somewhat differently, whether the name, date of birth, gender, address, phone number. If the information is recorded differently, then those need to be reconciled, is it the same person or not the same person?” Check says. “The second issue is that a person’s data changes over time. While the data was accurate for a particular individual when they visited one provider in January, it may not be the same data that they presented when they visited another provider in July. They could have gotten married and their name changed, and they could have changed their address or phone number. So both are accurate at the point of time, but when that information comes together, they don’t agree,” Check says.
Given the size and diversity of Healthix participating organizations, from large complex health systems to single provider practices, accurate and efficient matching and linking of patient records across the HIE is fundamental.
“Our mission is to help our participants provide better care by facilitating the secure exchange of information between them. To do this, we want to offer the best technical solutions available to match and link records of millions of patients, each of whom may visit dozens of healthcare institutions over the course of their lives, and many of whom have similar identifying information like name and birthdate,” says Check.
Check says Healthix’s current MDM software platform efficiently reconciles patient matching discrepancies with regard to common data discrepancies, such as a misspelled address or the use of a patient’s nickname in one record and a formal name in another record. “The software is smart enough to know different representations of the same data,” he says.
Healthix utilizes a stringent, conservative matching process, and, currently, if the software platform is not able to determine whether two records match, then a manual review of the data is required, which is a time-consuming process.
As patients visit healthcare providers over time, their data changes over time and that can create discrepancies that are challenging to resolve using an automated process. “It’s difficult for software alone to know that John Jones with this date of birth who lives at this address in Queens three months ago, is now the same person with the name John Jones with the same date of birth that lives in Brooklyn now,” Check says. “It’s especially problematic when you deal with common names and dates of birth, and especially when dealing with large numbers of people as we are. The chances that you’re going to have the same last name and first name for multiple people increases, and even the chances of having the same first name, last name and date of birth for more than one person increases.”
Check says, “That’s really one of the reasons that we found over time we’re accumulating a lot of instances that we haven’t been able to resolve data discrepancies in an automated way, and that’s where Verato came into the picture.”
After a competitive test with other technologies, Healthix decided to implement the cloud-based solution because of the quality of match results it provided, its ability to provide substantiated data for its match decisions and its real-time capabilities. Moving forward, Healthix will use the identity resolution services to automatically process millions of its most difficult match decisions.
Verato’s technology is based on a reference database, called Carbon, consisting of commercially available data sources, according to Mark LaRow, CEO of Verato. Referring to Carbon as a “self-learning database,” LaRow says, unlike traditional patient matching approaches, the technology isn’t limited by underlying patient data that is inconsistent, out-of-date, or incomplete. Essentially, he says, Verato’s patient matching technology automates the manual identity resolution process.
“At any given hospital system or HIE today, there are human beings sitting behind desks determining whether person A is the same as person B. They’ll go out and do research on the internet and then make a human judgment about whether it’s the same person. It’s the same thing here, but instead of doing research, the customer is doing an automated query to us and it’s the same basic questions,” LaRow says.
As an HIE that shares electronic health data, Healthix considers patient privacy and data security as its highest priorities, Check says. With the cloud-based identity resolution platform, Healthix can take the unresolved matches and can run a real-time query of each record. “We won’t have large amounts of data leaving our premises, giving us better control over the security of the data, which is extremely important to us,” he says.
More broadly, accurate patient identity is at the core of the kind of data sharing that is vital to population health efforts and coordinated care approaches, LaRow says.
“The electronic health record (EHR) systems are now gathering much more data than ever before and that’s a good start. We’re also seeing an avalanche of start-up technologies that hospitals are interested in, technologies that don’t come from EHRs vendors. From a standards perspective, there has been a lot of work done with Health Level Seven (HL7) to make healthcare technologies interoperable at a connection level, a communication level.”
He continues, “But the one big thing missing from all these interactions from EHR systems and this new breed of new applications is interoperability at the identity level. It’s a whole new idea. So you can have interoperability at the technical level, and everybody has seen that it has gotten the industry just so far and there has been major strides forward, but people feel limited in the inter-connectability of systems. And the big stumbling block right now is the lack of identity interoperability. All of these systems need to know that they are talking about the same person and to do that, everybody could implement their own matching system on premise, at the hospitals, and try to exchange identity linkages with each side matching for everybody else, but ultimately, that’s an end-to- end problem that falls apart at scale. Population health and the need to do analytics across many different databases requires identity interoperability, and coordinated care also requires identity interoperability, and I think that idea is going to become more common.”