Method for linking de-identified patients using encrypted and unencrypted demographic and healthcare information from multiple data sources

• US 20060026156A1
• Filed: 07/28/2005
• Published: 02/02/2006
• Est. Priority Date: 07/28/2004
• Status: Active Grant

• Current Assignee: IMS Software Services Ltd. (IQVIA Holdings, Inc.)
• Original Assignee: IMS Software Services Ltd. (IQVIA Holdings, Inc.)
• Inventors: Zuleba, Heather
• Granted Patent: US 7,668,820 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 21/6227 where protection concerns the structure of data, e.g. records, types, queries
• CPC Class Codes: G06F 21/6254 by anonymising data, e.g. decorrelating personal data from the owner's identification
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records

Abstract

A longitudinal database of de-identified patient healthcare transaction data records linked by longitudinal linking tags (IDs) is provided. A new healthcare transaction data record, which may include alphanumeric identification code attributes, third party attributes and/or demographic attributes, is assigned an linking ID associated with a previous healthcare transaction data record based upon successful comparison of either a designated set of identification code attributes or a designated set of demographic attributes. The longitudinal data base is assembled by a matching process in which a new data record is compared level by level with previous healthcare transaction data records through a hierarchy of a first series of matching levels each defined by a designated set of alphanumeric identification code attributes and a second series of matching levels each defined by a designated set of attributes including demographic attributes and then assigned the ID associated with a successfully matched reference data record.

Claims

1. A matching algorithm for assigning a longitudinal linking tag to a de-identified patient healthcare transaction data record (subject data record) whose data attributes include at least some of the alphanumeric identification code attributes, third party attributes and/or demographic attributes related to the patient healthcare transaction, the matching algorithm comprising the steps of:

(a) defining a hierarchy of matching levels, wherein each matching level is defined by a designated set of data attributes for comparison with reference data records associated with known longitudinal linking tags, and wherein the hierarchy of matching levels comprises a first series of matching levels each defined by a designated set of alphanumeric identification code attributes and a second series of matching levels each defined by a designated set of attributes including demographic attributes; and

(b) comparing values of the designated sets of data attributes in subject data record and the reference data records level by level through the first series of matching levels in an attempt to find a matching reference data record;

(c) if no matching reference data record is found at step (b), determining that at least a relevant demographic attribute is present in the subject data record;

(d) if at least the relevant demographic attribute is present in the subject data record, comparing values of the designated sets of data attributes in subject data record and the reference data records level by level through the second-series of matching levels in an attempt to find a matching reference data record;

(e) if a matching reference data record is found, assigning the longitudinal linking tag associated with that reference data record to the subject data record.

2. The matching algorithm of claim 1 further comprising step (f) of generating a new a longitudinal linking tag and assigning that tag to the subject data record when no reference data records are successfully matched to the subject data record at steps (b) through (e).

3. The matching algorithm of claim 1, wherein at least a matching level in the first series of matching levels is defined by a set of attributes selected from the group of Supplier Patient ID, Supplier ID, Outlet ID, Rx No., NCPDP Qualifier, NCPDP ID attributes and any combination thereof.

4. The matching algorithm of claim 1, wherein in step (c) the relevant attribute is the DOB attribute.

5. The matching algorithm of claim 1, wherein at least a matching level in the second series of matching levels is defined by a set of demographic attributes selected from the group of Insurance Card No., DOB, Gender, Patient Zip Code, Last Name, First Name. Address No., and any combination thereof.

6. The matching algorithm of claim 1 further comprising step (g) of comparing values of demographic attributes in the subject data record and a matching reference data record found at step (b) to confirm the matching of the two records.

7. The matching algorithm of claim 6 wherein step (g) comprises comparing the values of DOB and Gender attributes in the subject data record and the matching reference data record to confirm the matching of the two records.

8. The matching algorithm of claim 6 wherein step (g) comprises comparing values of Last Name and Zip Code attributes in the subject data record and the matching reference data record to confirm the matching of the two records.

9. The matching algorithm of claim 1 wherein the matching levels are arranged in an hierarchical sequence according to empirically determined matching effectiveness ranks of the designated sets of data attributes defining each match level.

10. Computer readable media comprising instructions for performing steps (b)-(e).

11. A longitudinal database of de-identified patient healthcare transaction data records (subject data records) linked by longitudinal linking tags (IDs), wherein each subject record includes at least some of alphanumeric identification code attributes, third party attributes and/or demographic attributes, wherein each new subject data record linked in the longitudinal data base is first compared with existing reference data records level by level through a hierarchy of a first series of matching levels each defined by a designated set of alphanumeric identification code attributes and a second series of matching levels each defined by a designated set of attributes including demographic attributes and then assigned the ID associated with a successfully matched reference data record and assigned a new ID when not matched to a existing reference data record.

12. The longitudinal database of claim 11 wherein at least one of the subject data records is successfully matched to a reference data record only at a level in the second series of matching levels upon comparison of a designated set of demographic attributes and is assigned the ID associated with the successfully matched reference data record.