When the Office of the National Coordinator for Health Information Technology (ONC) released its draft report on patient matching challenges and best practices in December 2013, it provided health information organizations (HIOs) with a blueprint for resolving one of the most significant challenges they face: accurate patient matching.
The report, which assessed current industry capabilities and best practices, delivered eight recommendations that will help the healthcare industry support a zero tolerance policy for patient matching. These ranged from requiring standardized patient identifying attributes to including patient matching functionality in certification to creating an open source algorithm for vendors to use to build/test patient matching algorithms.
Most importantly, the report makes it clear that HIOs must place a far higher priority on accurate patient matching than they have in the past. With limited awareness of their responsibility to assure accurate patient matching, many HIOs have implemented policies with the very real potential to damage data integrity and jeopardize patient safety.
For example, some expect participating organizations’ staffs to merge duplicates and fix overlays directly in the HIO’s system. This creates a serious liability for all stakeholders due to inconsistent validity decision making amongst data trading partners and non-HIO staff making changes to the HIO’s data. A better policy is for data issues to be corrected in the originating system and then updated the data in the HIO’s system via an electronic message.
Careful reading of the ONC’s report reveals other best practices that HIOs should implement, including:
- Conduct thorough evaluation of a system algorithm’s real record matching capabilities with threshold testing to eliminate auto-linking of false-positive matches and to optimize functionality with a goal of reducing false-alarm and false-negative matches
- Require HIO database users to enter several identifying data points (Last Name, First Name, DOB, etc.) or a combination thereof during front end record searches to help ensure the requestor gets the record they’re looking for while minimizing the number of records that are inadvertently disclosed to the requestor
- Store and utilize the last four digits of a patient’s Social Security Number, without revealing them to the HIO users
- Develop comprehensive data management policies and procedures including guidance on merging same-system confirmed duplicates, correcting demographic data errors and utilizing electronic messaging updates to communicate updates to the HIO
- Train HIO staff on correcting cross-system overlaps
- Audit record-matching algorithms’ effectiveness
- Establish a comprehensive data sharing agreement, compliance with which is mandatory
HIOs should ensure sufficient staff resources to review potential cross-facility matches and communicate same-system duplicates to the originating facility for correction. Data integrity errors should also be communicated, including records without enough data to verify potential overlays. This communication is another value-add the HIO can provide to its member organizations.
Finally, HIOs should establish a data governance sub-committee to craft a strong information/data governance program that defines required data fields and sets minimum percentages of those fields that must be populated with valid data. Policies should also determine a formula for calculating intra-facility duplicate record rates and establish the maximum allowed. As with the data sharing agreement, adherence to this program should be mandatory.
Healthcare—and HIOs in particular—must adopt a zero-tolerance policy for patient matching errors. The recommendations contained within ONC’s report, and the information that will ultimately come from the input currently being solicited by the ONC, will bring us much closer to achieving that life-saving goal.
Beth Haenke Just, MBA, RHIA, FAHIMA is founder and CEO of Just Associates (www.justassociates.com), a healthcare data integration consulting firm focused on patient matching, data integrity, data migration and health information exchange.