How Data is Standardized in UMPI

  • Data Standardization
  • UMPI

Attribute data values are standardized both to remove invalid values and to ensure consistent format for use in matching algorithm decisions. Universal MPI standardizes incoming data values on both web service requests and batch files. Standardized data values are persisted in Universal MPI in their changed or standardized form.

When those values are retrieved later (by retrieving the identity to which the attribute values belong), the values returned are the standardized values rather than the original values. Standardization can vary depending on the nature of each data element.

Data standardization in UMPI 

Attribute or attribute cluster Standardization performed by Universal MPI
All

For all attributes, extended ASCII characters with a logical ASCII equivalent are converted to their ASCII equivalent. For example, the extended ASCII character Ñ is converted to the ASCII character N.

All For all attributes, alphabetic characters are converted to uppercase. A subset of the Extended Attribute fields can optionally be kept in their original case, but by default all alphabetic characters are converted to uppercase.
Name
  • Numeric characters are not removed from name attribute values (though this can be changed with a configuration setting by contacting Verato support).
  • The position of name strings are not changed during normalization, but they are factored into matching decisions. For example, one patient record might have the name strings First Name=”JOHN PAUL” and Last Name=”SMITH”, while a second patient record might have the name strings First Name=”JOHN”, Middle Name=”PAUL”, and Last Name=”SMITH”. In this case, the position of the name strings are not changed or standardized, but the matching algorithm is designed to recognize that there is still a match between the names JOHN-JOHN and PAUL-PAUL.
  • Double-byte characters, if present, are NOT removed from name strings.
Birth Date
  • Birth date values are standardized to ISO format of YYYYMMDD.
  • Input birth date values are allowed with or without hyphen or slash characters. Input birth date values are also allowed in MMDDYYYY format, but they will be converted to YYYYMMDD format.
  • Input birth date values with only a 2-digit year (YYMMDD or MMDDYY) are converted to YYYYMMDD format only if the conversion is clear and unambiguous. For example, an input date value of 030405 is ambiguous – it could represent March 4, 2005, or April 5, 2003, so it is rejected in standardization.
  • Invalid dates and future dates are also rejected in standardization.
SSN

Alphabetic characters and punctuation symbols are removed.

Address

Address components are standardized in several ways consistent with US Postal Service conventions. Examples include:

  • Street direction values presented as a full text string are converted to their 1- or 2- character abbreviation (N, S, W, E, NW, NE, SW, SE)
  • Street type values presented as a full text string are converted to their USPS short format (Street becomes ST, Avenue becomes AVE, and so on)
  • Ordinal street names are converted to their numeric format (Second Street becomes 2 ND ST)
  • Unit, suite, or apartment information is parsed into address line 2, even if it is presented as part of address line 1
  • State names presented as a full text string are converted to their 2-character abbreviation (California becomes CA) Double-byte characters, if present, are NOT removed from city or state attributes, but they are removed from address line 1/address line 2 attributes.
Gender

Gender values are standardized to single-character gender codes of M, F, U, O, T, A, and N if the full string is provided as input. For example, an input gender of FEMALE is standardized to F. An input gender value that does not correspond to M, F, U, O, T, A, or N is not standardized – it is stored as-is.

Phone Phone Alphabetic characters and punctuation symbols are removed.