For today’s blog post, Dwight Blubaugh, IMT Solution Architect, offers his guidance based on years of dealing with less than perfect address data.

In January 2022, the Office of the National Coordinator for Health Information Technology (ONC) issued version 1.0 of a specification for standardizing patient addresses. Project US@ is an outcome of the ONC’s Patient Matching working group, which includes health IT developers, standards experts, state and federal agencies and public health organizations, amongst other stakeholders in health data advocacy.

 The goal: higher quality patient data matching and exchange.

Here at IMT, we have over 30 years of experience with patient demographic data and matching. We want to provide OUR take on the guidelines and let you know how we think they are best applied.

You can find the Project US@ specification here. Standardization encouraged, but not enforced at this time.  However voluntary compliance will yield benefits for cross organization patient data exchange, which IS mandated via the Cures Act.

Although ProjectUS@ (a clever way to say “USA,” by the way) does not obligate healthcare systems to modify or update existing data in any way, these recommendations strongly encourage healthcare providers, payers, public health organizations, and state and federal agencies to implement the guidelines for improved matching.

Historically, traditional patient and client registry implementations have not required standardization for accurate matching. At IMT, we have always focused on making it possible to match and link data as it was received from a source system, to minimize the risk of data loss or introducing errors due to misinterpretation of the data.

However, with increasing interoperability and data exchange mandates, along with advanced analytics initiatives across healthcare stakeholders, standardization does indeed have benefits. Standardization helps ensure that all systems can understand and use address information in the same way.

Why is address data still so bad?

Addresses have always been and continue to be an important demographic element used in matching. Identifiers are unreliable. And in the absence of universal identifiers either today or in the future, we must leverage what is consistently available. However differences in data collection practices, lack of data entry validation and contributions of useful, yet unstructured data from legacy systems bring inconsistencies that impact how the data can be semantically used, making the data less effective in some matching scenarios.

Transient and elderly populations often transition residences often with temporary stays across residences, moving from private residences to rehabilitation facilities, assisted living and long term care facilities. Lack of clarity on primary vs secondary or temporary addresses makes patient matching challenging.

A 2019 study by Indiana University indicated that using USPS address formatting can improve matching by up to 3%. However, this gold standard repository is not available to healthcare systems, despite many valiant efforts.

The USPS does validate 3^rd party systems called “Coding and Accuracy Support Systems” (CASS™ | PostalPro (usps.com). These systems require annual subscription licenses and must be integrated into your existing systems.  Overall, this can be cost-prohibitive while requiring some of your IT team’s time.

Let us look at what the Project US@ guidelines mean for your data. We will also share some best practices you can leverage by applying data profiling, quality, and enrichment solutions to your existing data.

Summary of the Project US@ Guidelines

The US@ specification includes these recommendations and requirements to standardize addresses for improved matching and data exchange:

1. Parse address data into individual fields to better standardize and validate. In a world of structured data, instances do still occur where addresses are captured in a single string. Here are reasons why that data should be parsed out:
   1. To verify that that the ZIP provided is valid for the given City and State
   2. To be sure that Street Name Spelling is correct and complete and not inadvertently abbreviated based on certain rules. (e.g., 123 N EAST AVE, where the STREET NAME is EAST. In this case EAST cannot be abbreviated nor can it be combined with NORTH to become NE. Street names with State names cannot be abbreviated, i.e., 123 NEBRASKA AVE. If both examples were changed to 123 NE AVE, you have a problem.
   3. Abbreviations are correctly applied per USPS Publication 28 and Project US@

2. Limit data loss in address transformation. Ideally the address as originally captured should be maintained as original source data records in your EMPI (Enterprise Mater Person Index), but the transformed address should not create conflicts in matching. The specifications recommend transformations such as:
   1. Removing extraneous characters
   2. Mapping Unicode with diacritical marks to reduced character sets (Ä maps to A)
   3. Characters should be uppercase
   4. Removing business line information and keeping separately (i.e., XYZ RETIREMENT VILLAGE)
   5. Abbreviating directionals (EAST, WEST, etc.) and suffixes (DRIVE, PARKWAY) unless they are part of the actual street name

3. Distinguish between types of addresses with effective dates. This can give insightinto a patient’s current living arrangements and considerations for post-release care, which can serve as an important social determinant of health.
   1. home, billing, long-term care
   2. current vs historical
   3. physical vs postal
   4. unknown and/or homeless

4. Verify the correct spelling of address components such as street name. This can help reduce issues occurring with edit distance of phonetic word matching.

5. Validate that house number is correct format and preferably to be in a known range. If a house number of 2090 BREEZY VIS is captured but the street number range is 100-300 then 2090 is incorrect.

For additional excellent guidance beyond the technical specifications, be sure to check out the comprehensive AHIMA Companion guide, which goes beyond data formatting and quality guidelines to offer guidance on improved data verification at the time the data is collected from from patients across the many ways they engage with your health delivery or plan systems.

IMT can prescribe the right approach for your data

IMT can evaluate the best approach to comply with these guidelines while helping with your organizational data exchange and analytics initiatives. We have experience with specialized utilities, verification, and transformation processes that can transition data prior to load (or before it is shared) without requiring a full reload of your current patient data repositories. We can help you profile and understand which data elements do not comply with the recommendation so you can focus your attention and resources on just those areas.

Your approach will vary based on your current platform and tools.  IMT is prepared to help you improve address data quality for your existing patient, client or provider registry solutions using our developed best practices in MDM data matching optimization, or with Health:iDM on the Semarchy xDM intelligent data platform which provides the flexible multi-vector architecture that allows you to retain source records while also persisting a standardized address in one platform.

Here are areas where Health:iDM can help with all-in-one tools in a single platform:

1. Profile your data to validate correct formatting of current addresses
   1. Test if Street Line 1 is the primary address, Street Line 2 is secondary, etc.
   2. Validate that a correct list of abbreviations is available

2. Differentiate addresses as private or business (e.g., retirement and rehab centers) through reconciliation with organizational or and facility reference data for more accurate population health analytics

3. Apply targeted validation and enrichment either through integrated third-party services such as Melissa or Google, or with internally managed reference data

4. Integrate your environment with external CASS systems as new data arrives or prior to outbound transmission from the EHR, EMPI or other system (United States Core Data for Interoperability (USCDI) | Interoperability Standards Advisory (ISA) (healthit.gov)) as part of your integration pipeline

5. Validate that externally presented addresses meet US@ recommendations and USCDI standards for effective dates in addresses

We can help you to understand your address data quality and help prescribe the right mix of analytics, tools, and processes to improve your data for matching and exchange. Contact your current IMT Engagement lead, Account Manager, or send a note here!

Stay tuned for our next topic – IMT’s recommendations on expanding patient data profiles to support the USCDI Patient Data Set!