Database schema to simplify processing and reduce false positive matches in ambiguous data

A method performed by a computing system, the method comprising:

• identifying a master record that incorrectly groups a plurality of source records in a database, the plurality of source records containing demographic information for multiple people and obtained over a network from a plurality of different databases, each of the different databases corresponding to a database of a medical provider or pharmacy;
• creating, according to a database schema that includes the database and a relations table that prevents linking between two records of the database, a first record containing demographic information for a first person, a second record containing demographic information for a second person, and exclusion data stored in the relations table of the database that indicates that the first record should not be linked to the second record, wherein the relations table does not identify a false positive relation between the plurality of source records;
• adding the first record and the second record to a collection of source records that includes the plurality of source records that are incorrectly grouped together;
• executing a matching algorithm based at least on the collection of source records and processing the exclusion data stored in the relations table to identify a set of connections in the collection of source records, each connection of the set of connections indicating a match between two records based at least on a matching criteria, wherein the exclusion data prevents the first record and the second record from being linked to each other;
• storing an indication of each connection in the database;
• building the first record and the second record based at least on the set of connections;
• and
• providing access to the database that includes the first record and the second record over the network to at least one medical provider or pharmacy.

A computer-readable storage medium having program instructions encoded thereon that are executable by one or more processors to perform a computer-implemented method, the method comprising:

• identifying a master record that incorrectly groups a plurality of source records in a database, the plurality of source records containing demographic information for multiple people and obtained over a network from a plurality of different databases, each of the different databases corresponding to a database of a medical provider or pharmacy;
• creating, according to a database schema that includes the database and a relations table that prevents linking between two records of the database, a first record containing demographic information for a first person, a second record containing demographic information for a second person, and exclusion data stored in the relations table of the database that indicates that the first record should not be linked to the second record, wherein the relations table does not identify a false positive relation between the plurality of source records;
• adding the first record and the second record to a collection of source records that includes the plurality of source records that are incorrectly grouped together;
• executing a matching algorithm based at least on the collection of source records and processing the exclusion data stored in the relations table to identify a set of connections in the collection of source records, each connection of the set of connections indicating a match between two records based at least on a matching criteria, wherein the exclusion data prevents the first record and the second record from being linked to each other;
• storing an indication of each connection in the database;
• building the first record and the second record based at least on the set of connections;
• and
• providing access to the database that includes the first record and the second record over the network to at least one medical provider or pharmacy.