The CoLoCode is a hyper-conformed version of U.S. practice location data, making it possible to link providers by their place of practice, despite variations in address data. That makes a lot of things possible, including the operations described below.
A Way to De-duplicate Healthcare Provider Records
There are three common issues in healthcare provider data that cause de-duplication of records to be challenging.
- In most cases involving federal healthcare provider data, the providers themselves supply the information. As with all data entered by people who are not primarily focused on data entry tasks, the result will be inconsistent information. For example, a practice address entered as "2345 West 39th" may appear on another record as 2345 W. 39 St," or a number of other common representations of the address, resisting deduplication efforts.
- In some provider datasets, an individual provider (Type 1 providers in the National Plan and Provider Enumeration System, known colloquially as the "NPI Registry") may also have a record for their personal business entity (a Type 2 record in the NPPES). For example, a physician will have his one and only one personal Type 1 record as required by law, and also a Type 2 record tied to their corporate or single-partner LLC business entity. And this is all within the same dataset.
- Also within the same dataset, a physician practice organization may have many Type 2 records, all belonging to the same business entity. In fact, a business entity is entitled to have as many Type 2 NPI registry records as they deem desirable for their operations.
The CarePrecise CoLoCode system hyper-conforms location data, going well beyond USPS Pub 28 standardization, to provide the key missing element in de-duplicating similar records.
A Way to Link Practitioners With Their Employers/Companies
The exclusive CarePrecise CoLoCode is used to identify individuals and organizations providing medical services at the same location. CarePrecise computes a CoLoCode for the full database of 7.2 million U.S. healthcare provider records. It can help in determining practice size, and in finding practitioners who are co-workers. It can also be used to find a Type 1 (individual) provider's Type 2 (organization) records(s) (Type 1 to Type 2 crosswalk), and to find the workers at a given practice group (workplace to worker crosswalk).
- View all providers working at every workplace (Worker-to-Workplace linkage)
- Link physicians' Type 1 (individual) records with their Type 2 (business entity) records
- Find all Type 2 records of any given location of service (organization's multiple NPI records)
The CoLoCode is constructed using a proprietary algorithm that hyper-conforms practice location data for the 7.2 million US healthcare providers. Going well beyond the US Postal Service's Pub 28 standard, the CoLoCode accounts for addresses that may have been entered very differently. For example, "2525 North 33rd Street, Suite 205" is understood to be the same location as "2525 N 33 #205," as well as many other variations.
The CoLoCode is a great asset when merging lists of individuals with lists of their companies, or de-duplicating databases, and is key element of CarePrecise's Qo-Relate™ record-linkage processing.
People Working at the Same Locations: The CoLoCode Method
Here's how a query is structured that links all providers having the same CoLoCode, in a manner like this:
Click to view graphic full size...
Think of the table on the far left as the "input" table, and the table on the far right as the table(s) that you're gathering the "output" from. The "Names tables" are shown in the example, but you can also include address, phone/fax, and other tables. If you're using CP ListMaker, you might use the tSelect_Matches table on the left, and perhaps the tNames and tAddresses tables on the right. But in every case, the two iterations of tCoLoCode are what makes the query work.
The "CoLoCode table" is used twice in the query; the first one is the CoLoCode for the provider you want to find co-workers for, and the second one is going to pull all of the NPI numbers for people (and companies) listing that location for their place of practice. A count of the workers having an organization's CoLoCode can be an indicator of the size of the organization.
This method has a wide range of applications. For example, it can be used to infer a healthcare worker's coworker network for a range of purposes. Companies who utilize CarePrecise data and tools for marketing can identify prospective customers in the cohort of existing customers' coworkers. The method has been used in contact tracing efforts to identify potentially infected workers within healthcare service populations.
The same configuration works to locate the workplace (organization) name for a given worker. On the left side we specify Type 1 records (individuals), and on the right side we specify Type 2 records (organizations), to arrive at the "worker-to-workplace" connection, and a row that contains both the individual's record and the organization's record. Once again, the graphic shows only Names tables, but other tables would be included; such as address, phone/fax, and taxonomy/specialty/license.
Using the CoLoCode in Excel
If you have an Excel file containing a CoLoCode column, simply sort the file on the CoLoCode column. Where you see a cluster of the same CoLoCodes, these are records with the same practice location. If your Excel file also contains the EntityType field, then the record(s) of Type 2 identify the organization (workplace), and the records of Type 1 identify the individuals (workers). Like this:
Click to view graphic full size...
The example above illustrates a characteristic of the CoLoCode. Notice that both a physician group and a hospital have been captured as a single workplace. This can happen when the postal addresses are too similar, and do not have standard secondary identifiers. To determine one from the other, use the hospital Taxonomy Codes (in the Taxo table) to identify the hospitals. The same logic applies if you're looking for clinical pharmacy or lab workers.
There's an upside here; the fact that so many labs, groups and pharmacies are co-located with hospitals makes it possible to roll up some of these allied organizations to their hospital connection, and by extension, the health systems they're part of (using the hospital-to-health system connections in the Authoritative Hospital Database).
Creating an Accurate NPI Record Rollup for a particular Type of Organization
The CoLoCode can be used as shown above to locate all NPI records for every physical location. As explained, more than one business entity may occupy a given postal address. For example, a clinical lab or pharmacy may be co-located within a hospital, but may be externally owned.
To create the most accurate results in a rollup of hospitals' NPI records ("NPI Rollup"), we want to limit the Type 2/Organization side of the query by including only the records of hospitals, and not the records of any other kinds of co-located business. This can done by using our facility type codes in the tTaxo table (providers' reported taxonomy codes) to include only NPI records that report one or more of the hospital facility types. The same can be done for pharmacies, laboratories, primary care clinics, physical therapy locations, and all across the healthcare spectrum. This is the method used to produce the Hospital Rollup packaged with the CarePrecise Authoritative Hospital Database.
In another scenario, we may want to identify records belonging to a physician practice which may be co-located within a hospital. While the CoLoCode will usually, by itself, contain adequate suite-level data to disambiguate this circumstance, in some cases it will not. By removing Type 2 records that are associated with hospital taxonomy codes, and using data in the tTaxoGroup table, we are able to pinpoint the records associated with the practice group. A number of CarePrecise clients use these tables, along with the CoLoCode table, in just this way.
CoLoCoding Beyond Healthcare Provider Data
The powerful CoLoCode feature is available only from CarePrecise. It is present in CarePrecise datasets wherever location data is available, (with the exception of the OtherPracticeLocations table at this time). But it can also be used on any dataset that contains street address field(s) (line 1 only, or lines 1 and 2), Zip Code® (5-digit or Zip+4) fields, and a unique ID field (such as the NPI). CarePrecise makes its proprietary CoLoCoding system available for a wide range of applications, such as developing business intelligence on health plans, and populating web applications with location-related services.