Person Matching Supports Data Exchange
By Elizabeth S. Goar
For The Record
Vol. 33 No. 6 P. 24
A giant insurer has taken steps to validate, cleanse, and link member data across hundreds of its health plans.
The negative impacts of ineffective or inaccurate patient matching have long been the bane of health care. HIM professionals have been vocal about the patient safety, cost, and workflow issues that arise when patients cannot be identified quickly and accurately due to incomplete or incorrect data or ineffective matching algorithms.
But the challenges patient misidentification introduced into the nation’s pandemic response have created a new, greater sense of urgency around identifying a true solution to the problem, along with a flood of new solutions targeting the ability to recognize and match patients with their health care data.
Just like their health care provider counterparts, payer organizations struggle with patient misidentification. But they are further challenged by managing an even broader scope of data that makes up the full health care consumer experience including information on both care and claims that becomes part of a consumer’s health history. Dubbed “person matching,” it along with identity management is considered foundational to the broader function of health information exchange (HIE).
Says Karen Proffitt, MHIIM, RHIA, CHP, vice president of data integrity solutions with Just Associates: “The vast scope of person matching makes it foundational to effective and efficient HIE. We’re talking a range of data that is far beyond what we normally consider when we think about patient matching, which is typically limited to the patient’s demographic and clinical information collected within provider settings.”
Payers’ Person Matching Challenge
According to Proffitt, unlike patient matching, person matching encompasses clinical data “as well as a wide range of other information, including claims data, social determinants of heath, and even patient-generated wellness data.”
Adds Mariann Yeager, CEO of The Sequoia Project: “The transition from patient matching to person matching is a recognition that individuals are accessing health care services in a wide variety of settings. People may be sick patients or healthy individuals engaging in wellness activities. Person matching is more encompassing.”
For payers, she continues, person matching is particularly important because “individuals change health insurance plans frequently. Payers have the same problem of linking health care services provided to an individual over time.”
Michael Trader, cofounder of RightPatient, notes that the broad scope of data that goes into person matching makes it particularly difficult—and important—to maintain the integrity of the data being aggregated and exchanged. The efficacy of any kind of data exchanged is only as good as their quality, he says.
“When disparate health care providers share data, rates of duplicate and overlay medical records increase exponentially,” Trader says. “Accurate patient matching is especially important here since polluted data can directly impact health outcomes. Having access to a patient's complete medical history—information like allergies, preexisting conditions, medications, etc—enables clinicians to provide the highest quality of care.”
It's not just the frequency with which information changes that creates obstacles to accurate patching matching from the payer perspective. The sheer volume is also a problem—one that is critical that they overcome.
“Payers need to work with the cleanest data possible to optimize claims processing and population health analytics. Accurate person matching is also vital to avoid HIPAA breaches. For example, you don't want to send an [explanation of benefits] to the wrong person,” Trader says.
Some solutions try to stop problems before they happen. For example, biometric platforms are designed to accurately identify the patient or member on the front end during a patient visit where the entire health care cycle begins.
“There can be a lot of downstream fallout when mistakes happen there,” says Trader, who adds that solutions like front-end biometrics “can mitigate those issues by ensuring accurate patient identification. This is essential to achieving the highest degree of data integrity.”
A Blue Cross Blue Shield Priority
For Blue Cross Blue Shield Association (BCBSA), the solution to its person matching problem was a homegrown project that validated, cleansed, and linked member data across hundreds of its health plans.
Solving the seemingly intractable problem of person matching was a priority for BCBSA in order to align with its interoperability principles of ensuring patients own and have access to their health care data when and where they need them; standardizing how health data are exchanged in order to make them accessible, safe, and private; and removing barriers preventing the flow of data and to allow secure, efficient, and cost-effective access.
Achieving these goals was hampered by several realities, notably that the data in question were not only disconnected but also siloed within disparate systems across unrelated facilities. Further exacerbating the problem is BCBSA’s sheer size and volume of patient information; it manages the health care experiences of more than 175 million active and inactive members and processes $450 billion in annual claims from 95% of the nation’s hospitals and physicians in all 50 states. Its national health care data encompass 56 million provider records from 1.9 million active unique providers and treatment cost estimates from more than 6.6 billion covered procedures performed each year at 67,000 health care facilities and by 800,000 professional providers. And, thanks to the explosion in patient-generated information, these numbers are growing exponentially.
Ensuring that patients and their physicians and other providers have access to their complete health history at any time required BCBSA to get creative, which lead to the creation of its Member Matching Initiative (MMI) to harmonize data across its entire system.
The MMI, which was the focus of a HIMSS21 presentation by Yeager and Desla Mancilla, DHA, RHIA, BCBSA’s manager of health data interoperability, links data from BCBS companies where individuals were members of more than one BCBS health plan.
During “Person Matching for Interoperability: A Case Study for Payers,” Mancilla described the MMI as “foundational to develop person-centric health insights that can be used to support members’ journeys to well-being.” The initiative was described as the following:
• person-centric and unchanging, meaning it can relate a member across addresses, life changes, products, accounts, and plans;
• critical to informing BCBSA’s clinical studies, providing data for insights to support better health outcomes, and facilitating coordination of benefits, coordination of care, and wellness initiatives; and
• secure, because it uses field-level tokenization and multilevel security controls, and is SOC (System and Organization Controls) II, Type II, and HITRUST certified.
Mancilla noted that the MMI, while not intended for use on member ID cards, nonetheless supports linking member identities, activities, and events across time to other sources of data such as EHRs, care management, and population health. It also enables greater payer participation in HIE, as required by federal regulations.
Finally, the MMI enables the patient experience to be reimagined by “unlocking member engagement, care, and disease management data to create a seamless, nationwide experience.”
The MMI was implemented in two stages: a demonstration project to confirm efficacy followed by broad implementation across all BCBS plans.
The workflow starts when plans submit names, addresses, etc, to the BCBSA for data validation and rules-based cleansing. The MMI Recognition Manager then assigns a unique ID to each member record, which is validated before the data are sent back to the originating plan. That plan reconciles the MMI ID assignment for members where the BCBSA MMI assignment doesn’t align with the plan’s unique person identifier.
A third-party vendor was engaged to provide “the referential consumer data [and] also convert the addresses into standardized USPS format,” Mancilla noted in her presentation.
Thus far, 130 million unique members across the BCBS system have been identified through the MMI matching algorithm. Of those, more than 15 million have had coverage in more than one Blues plan, representing about 12.7% of unique members in multiple Blues plans’ commercial businesses.
A 99.5% match rate was achieved using the MMI solution, a figure validated by The Sequoia Project in the demonstration project that was designed to corroborate the MMI accuracy rate in comparison to the matching accuracy rate of an HIE organization.
Based on its early success, the MMI project is now entering the second phase: expanding to all BCBSA companies. Opportunities to enhance interoperability with standardization using Health Level 7 Fast Healthcare Interoperability Resources–based web services have also been identified. These are important advances that can help health care organizations realize the full potential of person matching.
“Person matching across organizations is a foundational building block, and a remaining barrier, to interoperability,” Yeager says.
— Elizabeth S. Goar is a freelance writer based in Wisconsin.
PERSON MATCHING PROVIDER-PAYER DIFFERENCES
In “Person Matching for Greater Interoperability: A Case Study for Payers,” a report published by The Sequoia Project that details the Blue Cross Blue Shield Association’s (BCBSA) Member Matching Initiative demonstration project, distinctions were drawn between “person matching” in payer and provider settings, calling it “one of the most basic differences among industry stakeholders.”
Linking disparate dates of service for an individual is typically managed through a master patient index (MPI) or Enterprise MPI in provider organizations and health information exchanges. In some cases, a master data management (MDM) system is used to create a unique identifier.
However, according to the report, “the nomenclature in health care payer organizations revolves around members, dependents, claimants, etc. Hence, the matching solution built and employed by BCBSA is referred to as the master Member Index Identifier with a member being defined as a unique person across all BCBS companies.”
Terminology referring to duplicates and overlays is also different between the two settings. In provider organizations, the term “duplicate” reflects a single individual with multiple identifiers, while “overlay” refers to two or more unique individuals represented as being the same person under a single MPI number. In payer settings, the focus is on unique matches, which is why BCBSA uses the terms over- and undermatches.
Furthermore, payers have contractual relationships between the subscriber and their dependents and can therefore implement algorithmic rules that don’t allow dependents on the same contract to have the same identifier. Providers do not typically have access to this type of data. Payer member data usually come from one source, such as the employer or from an insurance application, whereas provider member data originate from sources within the provider ecosystem, such as physician offices, labs, and hospital systems, and are obtained at the point of service.
“As a result, there may be more disparity in provider data and greater accuracy in payer data,” notes the whitepaper.