Six years ago, San Diego Health Connect (SDHC) began as a local health information technology project and has since grown into a regional health information exchange (HIE) that connects San Diego’s three largest health systems and enables the exchange of 3.2 million patients’ electronic medical records (EMRs).
As the HIE has expanded, SDHC leaders have worked to improve the exchange and the quality of the data that’s available through it. That work has come with its share of challenges, and through the use of third-party data quality and data integration technology tools, SDHC has been able to overcome many of these hurdles.
SDHC, a not-for-profit HIE, officially evolved from the San Diego Beacon Community, the largest of the 17 Beacon Community projects that received a total of $250 million in three-year grants from the Health and Human Services Office of the National Coordinator for Health Information Technology (ONC). In April 2010, UC San Diego received a $15.3 million Beacon cooperative agreement from ONC. The San Diego Beacon Community was tasked with building and strengthening local health information technology infrastructure and implementing new approaches for making measurable improvements in the cost and quality of healthcare.
The regional health information exchange was one of the initial projects of that Beacon community, and in 2013, UC San Diego transferred operational and oversight responsibility for the HIE to SDHC. The Beacon grant gave San Diego healthcare organizations a jumpstart for implementing HIE to improve healthcare quality and efficiency, and according to SDHC executive director Dan Chavez, the HIE is further ahead than most HIEs in its development and technological sophistication.
The community-wide HIE now enables the exchange of electronic medical records for 3.2 million patients between 25 different clinics and hospital systems throughout San Diego and Imperial counties, including UC-San Diego Medical Center, Scripps Health, Sharp HealthCare, Kaiser Permanente, Rady Children’s Hospital-San Diego, the U.S. Department of Veteran Services and the County of San Diego Health and Human Services. San Diego is the fifth most populous county in the United States, with three military facilities, 18 federally recognized Indian reservations, 19 acute care hospitals, four non-acute, rehab hospitals, 115 clinics and 9,000 physicians.
SDHC provides healthcare providers with medical records exchange, direct secure messaging and alerts, public health reporting capabilities as well as the EMS Hub, which transmits pre-hospital data from EMS vehicles en route to the hospital.
Throughout the SDHC’s roll out and adoption, the organization’s leaders have faced a significant obstacle that is a common challenge for HIEs—accurate patient identity matching. SDHC is a federated model and does not house medical records, as the clinical data is housed at each of the participating facilities. The HIE uses a master patient index (MPI) as a record locator service to manage the identifying information for 3.2 million people across the San Diego region and to ensure that the records belong to the correct patients. The HIE facilitates 7.5 million transactions a month between the 25 provider institutions.
Using the matching capabilities in the MPI tool, when patient records do not match they end up in an “exception queue.” Due to stringent matching criteria, SDHC’s MPI matching algorithm had, by 2015, excluded 187,000 patient records because of unresolved patient identification, Chavez says. And, each new provider joining SDHC led to more backlog, due to data quality and differences in governance, he says. The HIE also was using a manual data stewardship process. So, the task of addressing the backlog of 187,000 records in the “exception queue” would take eight years for two full-time employees.
Accurate patient identification is not only a data management and data quality issue, it’s also a patient safety issue. As reported earlier this year by Healthcare Informatics, ECRI Institute ranked patient identification errors as second on its list of top 10 patient safety concerns. ECRI Institute analysts discovered that patient identification errors "were not only frequent, but serious.”
In order to develop a better patient matching progress, SDHC convened a working group of 41 HIE members from 13 organizations. As a result of that working group’s efforts, in May 2015, SDHC began a pilot project with health IT startup Verato, a Software-as-a-Service provider that developed a cloud-based patient matching solution.
Accurately identifying people, essentially matching the right record to the right record, is a pervasive problem in healthcare as well as other industries like retail, according to Brent Williams, founder and chief technology officer at Verato.
The problem, Williams says, is that identity data is a collection of attributes which often change over time. For instance, names, addresses, phone numbers, email and marital status can change over time. In addition, there’s also ambiguity with names, such as hyphenated names, nicknames and twins. There can also be spelling errors and homonym errors and data governance issues, such as formatting and data quality. So, one person can be represented by old, incorrect or incomplete data, resulting in different identities across provider systems. The matching capabilities in MPI tools typically result in a 70 percent match rate given typical error rates in identity data, Williams says, and the other 30 percent of unmatched identities must be manually resolved, as the case at SDHC. Or, the identities are falsely categorized as non-matches which prevents the information from reaching a patient’s care providers during the care visit.
Verato launched its technology in 2012 to try to solve this problem and the core of the technology is a specially programmed reference database, Carbon, that’s provided as a service to hospitals and government agencies to reconcile identities faster.
During the pilot project, SDHC employed Verato’s technology and was able to resolve 75 percent of the 187,000 mismatched records, thereby eliminating 75 percent of the manual effort that would have been required to match the 187,000 records in the exception queue. Of the 45,000 in the exception queue not matched, 95 percent were pediatrics patients. In addition, during the pilot project, SDHC was able to boost its patient match rates as the technology platform identified an additional 126,000 patient matches that the MPI algorithms had originally missed. In total, SDHC increased the number of patient record matches in its MPI by 110 percent, Williams says.
Of new matches that were identified during the pilot, 20 percent had at least one critical error, either conflicting birthdates, a difference in last name or a different address. According to Williams, the software utilizes a referential matching engine. “MPI matching, or probabilistic matching, can’t see through different or bad identity data. Referential matching works despite different or bad identify data,” he says.
Due to the success of the pilot program, SDHC has implemented the technology platform, and according to Chavez, it will be used to augment the MPI database through an automated data management feature aimed at detecting matches missed by the standard MPI process.
The combination of high quality data and stringent patient matching standards enables SDHC member providers to have access to a complete picture of a patient’s health information, Chavez says.
“SDHC’s mission is to facilitate the exchange of medical records throughout our region so doctors can provide better patient care, decrease the number of duplicate tests and procedures, and decrease costs,” he says. “Foundational to exchanging records is the ability to match various records that belong to the same patient, and this technology can increase our match rates without sacrificing SDHC’s stringent data governance standards.”
Chavez says the value of the HIE to San Diego health systems includes improving care coordination, improving Meaningful Use compliance and reducing preventable admissions. For physicians, the HIE helps to strengthen provider engagement and reduce costs, and for community clinics, it improves access to relevant patient information and speeds up the referral process. At the same time, the HIE also enables automated real-time public health reporting.
Moving forward, SDHC continues to expand its services and capabilities. California Healthcare Foundation recently awarded SDHC a grant to pilot, implement and maintain a Physician Orders for Life-Sustaining Treatment (POLST) eRegistry in San Diego. San Diego will be the pilot geography to work with a group of diverse local providers to make the POLST forms accessible by a wide variety of disparate organizations and functions, Chavez says. POLST is a standardized medical order form that indicates which specific treatments, such as a ventilator or feeding tube, a seriously ill patient does or does not want. Unlike a health care directive, a POLST form is signed by the patient and physician and is intended to serve as medical orders that move with the patient across settings of care.