ONC PMAL Project: Creeping Forward on Patient Matching

0 Share

ONC PMAL Project: Creeping Forward on Patient Matching

Last week, the Office of the National Coordinator for Health Information Technology (ONC) released the final report from its Patient Matching, Aggregation, and Linking (PMAL) Proje ...

Last week, the Office of the National Coordinator for Health Information Technology (ONC) released the final report from its Patient Matching, Aggregation, and Linking (PMAL) Project, as well as an additional report describing a pilot project to test the Patient Demographic Data Quality Framework (PDDQ) to Support Patient Matching that was released several years ago. Funded from June 2015 through September 2018 by the HHS Office of the Assistant Secretary for Planning and Evaluation (ASPE) through the Patient-Centered Outcomes Research (PCOR) Trust Fund, PMAL was one of the activities I described in an earlier blog entry.

The Final Report reviews the four challenged of patient matching and linking that the PMAL project attempted to address:

  • Improvements to Matching Algorithms: This section primarily addresses work that was done to support the ONC Patient Matching Algorithm Challenge, including development of test data and an open source test harness. I have previously raised my own concerns about the challenge participants “tuning” their algorithms for the test data set, and ONC also seemed to indicate at the time that the “gold standard” data set was not actually used for the challenge. With regard to the test harness, I am not aware of anyone leveraging this tool for algorithm evaluation but it very well may be in use.
  • Improvements to Data Quality: The main activity here seems to be regarding PDDQ, an evaluation tool for organizations to use to assess the standardization of their policies, procedures, and practices with respect to patient matching. The more detailed report describes a pilot project using PDDQ done in conjunction with a set of community health centers affiliated with OCHIN. Once again, I am not aware of any organizations who have used this tool, and the results from the pilot study were mixed at best.
  • Expanded Data Sharing: In many ways, this section described projects I was least aware of and which I found most interesting. It describes a number of efforts within the standards development world to create specifications and tools related to security and trust which are prerequisites to effective matching and linking. The project built upon standards like OAuth and FHIR and worked with the OpenID Foundation to create the Health Relationship Trust (HEART) Working Group. They also experimented with innovative technologies like blockchain to explore its relevance to this domain.
  • Data Standardization: Activities in this arena focused on tools to evaluate record completion (necessary for good matching or linking), data provenance (necessary for better reliability, supported by an ONC Data Provenance challenge), and work on provider identity and provider directory (again, to help ensure better reliability and accuracy).

This project helped ONC and the PCORS community “go deep” on a number of important technical and policy issues. It is less clear how widely these tools are being used or even how useful they are given some of the limitations described in their own studies. We still do not have a unified approach to patient matching and linking in healthcare, but perhaps we might make more progress if Congress proceeds to lift its ban on a national patient identifier. As care becomes more patient-centered, and the need for data sharing and integration extends beyond just healthcare (see the Stewards of Change National Interoperability Collaborative’s Person Matching Workgroup), better person matching and linking become even more important.