Mapping makes it possible for Clingience to consume and analyze data from multiple providers, multiple practices and even multiple EHRs. It also makes it possible to produce meaningful reports without requiring the practice to use the exact terminology of an authority.
 

Mapping Methods

In order to detect instances where a term identified in a classification filter has been used by a provider, Clinigence must create a relationship between the items in the practice’s data (non-authoritative domains) and the items in the code lists (authoritative domains).  These relationships are called "Mappings".  There are several ways in which a mapping can be created:

  • Standard Mappings
  • Identity Mappings
  • Crowd Source Mappings

 

Identity Mappings

When patient data is loaded into the Clinigence system, many data items are automatically mapped to items in the Code Lists. This occurs when the match is simple and unambiguous, such as Birth dates from the patient profile, CPT codes matching numerically, and ICD codes matching numerically or by description. This is called Identity mapping.

The mapping administrator can specify an identity relationship between the domains of an EHR Vendor and the authoritative domains. After that, whenever a new practice provides us data from that EHR, the Clingence engine will recognize the identity relationship of the corresponding practice-specific domains.

Standard Mappings

Other items must be mapped manually from the patient data to an appropriate Code List. This is called Standard mapping. This is generally accomplished by selecting the appropriate Non-Authoritative Domain for the patient data item (Diagnosis, Treatment, Vital Signs, etc) and searching for an equivalent counterpart in one of the Authoritative Code Lists. The person manually mapping the data will determine when to create a relationship between the combination of the non-authoritative domain and the description of the patient data item and the available code list item.

mapping.png 

As you can see from the graphic, several items in the patient data can be mapped to a single item in the code list.
 

Crowd Source Mappings

Crowd sourcing is the ability to generate mappings in a practice’s data using mappings from other practices’ data.

Many providers use terms that are equivalent but not identical to authoritative domain items. Even though these provider-generated terms are not authoritative, they are likely to be common enough so that multiple providers use the same terms to mean the same thing. Because of this we can extract information from standard mappings and use that information to make assumptions about new data we receive.  For example, if 10 different providers have created a standard mapping from Sex=Boy to Gender=Male, then when another provider comes along and tells us that his patients have a property Sex=Boy, we can be pretty certain that the new provider also means Gender=Male.

Credibility of Crowd Source Mappings

How do we ensure that manual mappings are legitimate enough to be disseminated to other practices? Clinigence allows each practice to determine the credibility required for crowd source mappings:

  • Minimum number of votes.  The minimum number of standard mappings of a non-authoritative domain item (domain+code+description) needed for the practice to accept that mapping. Default is 2.
  • Minimum voting percentage. The minimum percentage of standard mappings of a non-authoritative domain item (domain+code+description) mapped to a specific authoritative code list item (domain+code+description) needed for this practice to accept a mapping. Default is 65%.