IBM InfoSphere MDM - EMPI - Initiate - MDM - Master Data Management
Learning Centre: Ideas

Key Points from the ONC Patient Identification & Matching Working Group

On June 1st, IMT’s Corinne Blair (Healthcare Practice Lead) and Deanna Nole (Solution Offerings Director) attended the day long ONC Patient Matching and Identification Working Group session.  This was a marathon session of almost 30 presentations from healthcare providers, policy folks and vendors presented important considerations and recommendations for better patient matching from the perspective of their experience and domain.

The presentations covered many different perspectives/challenges/solutions and day moved along pretty quickly!  The slides and the recording will be posted on healthit.gov sometime in the future.

Here are the key points we took from the sessions: 

  1. Intermountain Healthcare CIO Mark Probst did a great job clearly stating the success criteria and goals for patient identity management and true interoperability: Safety, Quality, Cost Savings and Accuracy.  He also outlined challenges with facility identification vs enterprise identification and stressed the importance of a person-centric approach for digital identity across systems using common standards and a common set of credentials.
  2. Effectively creating a single Universal Patient Identifier (UPI), Patient Safety Identifier (PSI) or Patient Identity Passport, requires algorithms, policies, training, process and technology to assist in proper assignment to unique individuals and make sure duplicates are not created.  This was reinforced by presentations on universal identifier adoption from an international perspective.
  3. If a single ID can be reliably established, but is not used everywhere care is provided, it will not achieve its stated goals and promote true interoperability.
  4. To share digital identity across systems, a common standard and set of credentials for identity and proofing is needed that will likely require specialized technologies like smart phone apps and facial recognition. These methods are  certainly innovative but may not work for everyone and would take some time to roll out.
  5. Biometrics have their place in identification and matching, but reliable use and trust is limited.  Local identification can work well, but it is not suitable for widescale identification, nor is the industry great at securing such sensitive information.
  6. Referential matching may help, but it’s not clear how.  Very limited information was provided on the best sources of reference information and how it works for all persons including children, a common obstacle for suitability of reference matching technology.
  7. Identification is one thing, authentication is another.  The presentations on Digital ID and Trust frameworks for patient/consumer facing applications seems a bit beyond the focus of the session and would probably be best served with its own workshop as it is a complex, important topic.
  8. Enterprise Master Patient Index (EMPI) is still the most widely used technology to effectively match data across systems in the absence of a common IDs and other mentioned technologies rolled out at a large scale.

The ONC working group has a challenge ahead of them, to digest the various approaches and concepts in a coherent recommendation to Congress.  Obviously, proposing a single identifier would be the simplest recommendation to make, yet more complex and nuanced to implement correctly to achieve the goals of patient safety, privacy, cost savings and accuracy to ensure true interoperability.

IMT did not get a chance to present today, but we hope to get on the presentation roster for the summertime working session.  We want to share our experiences implementing standard processes, common credentials for patient identify matching and linking across a large set of healthcare providers to reinforce that there is no silver bullet.  No single identifier approach or technology exists to solve this problem.  The right set of existing matching technologies (i.e., EMPIs) with next generation tools and processes are needed to support a person centric approach to digital identity sharing.

  • Innovative Tools to prevent mis-identification and data entry errors​ through Active Integration
  • Tools for easy decision-making for human stewards based on stewardship best practices
  • Automated matching and linking using advanced machine learning approaches
  • Regular measurement and adjustment via an Intelligent Data Platform for analytics and insights that can be used to monitor data quality and completeness

Corrine and Deanna were glad to take a day out and hear what the rest of the industry is thinking and are glad to see how many organizations are passionate about patient identity and matching. They look forward to staying engaged in the conversation as it moves forward.