Why is Patient Matching So Important Today?
Patient matching is the ability to link all of a patient’s data across different applications to one correct person, using demographic data as the key, to form a comprehensive understanding of each patient. But patient matching is challenging because 30% of demographic data is inaccurate, incomplete, or inconsistent. This problem multiplies when systems attempt to exchange information with one another. The Pew Charitable Trusts recently reported that match rates are far below the desired level for effective data exchange across organizations.
As the number and variety of data sources continues to increase with M&A activity and care-coordination efforts, more healthcare initiatives—like population health, value-based care, and digital access—begin to depend on accurate health data. Legacy patient matching technologies already struggle to perform at scale and will not be able to sustain new initiatives and new data sources. This leads to decreased patient safety and increased cost, as the accuracy of analytics, billing, and clinical decision making are all tied to the accuracy of patient records.
How is patient matching done?
Patient matching has long been an issue that technology tries to address, most recently with master patient index (MPI) technology. So far, there have been two revolutions in this approach:
- From first generation MPI to second generation enterprise master patient index (EMPI).
- From second generation EMPI to the newest third generation Verato Universal Master Patient Index (UMPI).
This third-generation of patient matching is:
- Cloud-based—making it more accessible.
- Nimble—minimizing the manual effort required to respond to rapidly changing requirements.
- Scalable—remaining accurate and performant regardless of scale—even across organizational and regional health information exchange.
- Accurate—harnessing the power of Referential Matching technology to supplement inaccurate, incomplete, and out-of-date source demographic data with highly curated reference data for unprecedented matching accuracy.
This blog will examine how these revolutions occurred by discussing five defining characteristics of MPI technologies.
1. Deployment model—how the solution is deployed
2. Data scope—how many applications the solution is designed to (and capable of) working across to match patient data
3. Matching—the methods and algorithms the solution uses to identify patients and match their records
4. Data stewardship—how the solution manages records that it cannot match automatically
5. Configurability—how easy it is to customize the solution and integrate it with other technology
Read on to learn more about what each generation of patient matching looks like.
First Generation: What is an MPI?
An MPI is defined by:
- Deployment model: Embedded in an EHR or other IT system.
- Data scope: Single application.
- Matching: Deterministic algorithms.
- Data Stewardship: Rudimentary data stewardship—lets humans take over where the MPI struggles to match.
- Configurability: Minimal—includes integration of other systems, and algorithm tuning for unique names/populations.
However, MPIs face challenges:
- MPIs’ algorithms are often based on yes/no deterministic logic that struggles to overcome even simple data errors and inconsistencies.
- Matching across multiple data sources is not something they were designed to do—making them unsuitable for organizations that need to connect data from multiple sources.
With greater adoption of electronic health records came the need for an enterprise-wide matching solution that could handle multiple data sources, which led to the second generation of MPI technology, the EMPI.
Second Generation: What is an EMPI?
An EMPI is defined by:
- Deployment model: Standalone product from a specialized vendor, typically deployed on-premises, but recently vendors have begun offering EMPIs as hosted solutions in the cloud.
- Data scope: Multiple applications including EHRs, data warehouses, and analytics.
- Matching: Probabilistic algorithms.
- Data stewardship: Manual resolution process based solely on source system data.
- Configurability: Extensive, but complex.
Because of their advanced algorithms and data stewardship functionalities, EMPIs can scale to support many enterprise-wide use cases. However, they too face challenges:
- They are on-premise or cloud-hosted solutions and so they incur large upfront costs, take months to implement, need database administration efforts to remain accurate, and require expensive software upgrades every few years.
- Because their algorithms are sophisticated, they can overcome simple data errors and inconsistencies, with advanced manual algorithm tuning to ensure the EMPI is best suited to match unique patient populations. But their algorithms cannot overcome dramatic data errors and inconsistencies—they struggle when connecting patient data with out-of-date data (like old addresses or maiden names), ambiguous data (like using a middle name as a first name), incomplete data (like missing SSNs or birthdays), and default data (like 1/1/2000 for a birthday).
- Some EMPIs require algorithm re-tuningwith each additional data source, taking months of effort to maintain matching accuracy.
- EMPIs can match data against multiple patient repositories but struggle as the breadth of data sources begins to include sources with different levels of data “richness”—meaning data that contains different attributes, relationships, or entities.
So, while EMPIs can integrate and match with more data sources than an MPI, they are not equipped to handle the data challenges of modern health systems that face a rapidly changing regulatory environment, higher demand for data exchange, accelerated M&A, and increased demands for patient access. Verato identified this issue, knew there was a better way, and created the UMPI.
Third Generation: What is the UMPI?
UMPI is a Universal Master Person Index and it is only offered by Verato. The Verato UMPI uniquely offers these capabilities:
- Deployment model: SaaS—deployed in a HITRUST-certified cloud—person matching as a service. Go live in weeks.
- Data scope: Multiple traditional applications (EMRs) plus consumer-oriented systems, cloud-based analytics, care platforms, and emerging data sources like claims, PHR apps, patient portals, telehealth, genetics, biometrics, SDOH, wearables, Salesforce.com, CRM, and marketing data.
- Matching: Referential Matching—incorporates smart probabilistic algorithms with a reference database of demographic data spanning the US population with up to 30 years of historical data. Its probabilistic algorithms use this database as an answer key during matching to overcome dramatic demographic data errors and inconsistencies. Essentially, this reference data is used to augment existing data for matching—overcoming the 30% of demographic data that is out-of-date, incomplete, or inaccurate
- Data Stewardship: Automated and informed. UMPI uses its Referential Matching to automatically resolve identities that EMPIs could never resolve due to incorrect, incomplete, or inconsistent demographic data. In fact, UMPI typically generates 2x-4x fewer stewardship tasks requiring manual review. UMPI also enables data stewards to be informed by its reference data during task resolution so they can be over twice as efficient in resolving tasks. So, counter-intuitively UMPI is defined by the amount of data stewardship you have to do—which is significantly less.
- Configurability: Easy to integrate with RESTful APIs. Minimal configuration necessary for added data sources. No tuning required.
The UMPI thrives at supporting modern healthcare initiatives, like:
- Digital Front Door—Patients are becoming consumers and they expect efficient care from the comfort of their homes, through mobile applications, telehealth, and patient portals. Empower your patients with their medical records and meet them where they are to deliver convenient care.
- M&A Activity—Health organizations are rapidly scaling and expanding to remain competitive. Maintain data integrity as you migrate disparate sources into one EHR to uphold patient safety and patient experience as you grow.
- Consumer-centered analytics—Report with confidence by identifying all patients accurately and achieving a “single view” of patients to inform downstream analytics.
- Patient Access—Gain confidence that you are providing accurate, complete patient records whenever a patient logs into a portal or an application.
Verato built the UMPI to support rapidly evolving patient matching needs that legacy EMPIs do not address well. Supported by curated reference data, Verato confidently matches patient information of varying “richness” and from multiple data sources. Most importantly, the Verato Universal MPI is nimble: it’s an easy-to-implement SaaS solution that integrates through modern RESTful APIs, and requires no hardware, maintenance, or downtime. The Universal MPI was built with healthcare in mind and healthcare organizations are already realizing the benefits. Verato’s customers represent the nation's leading providers, payers, and HIEs including Northwell, Intermountain, Healthix, and Manifest MedEx.