RECORD-LINKING SYSTEM FOR GENETICS AND EPIDEMIOLOGY

MANY GENETIC AND EPIDEMIOLOGIC STUDIES RELY ON LINKING RECORDS IN SEPARATE FILES THAT REFER TO THE SAME INDIVIDUAL. THE PURPOSE OF THIS PROJECT IS TO DESIGN AND BEGIN IMPLEMENTATION OF A PROTOTYPE OF A SYSTEM THAT FACILITATES LINKING RECORDS WHERE THE IDENTIFYING INFORMATION BETWEEN RECORDS IS NOT IDENTICAL. ALTHOUGH A GENERAL THEORY OF RECORD LINKING HAS EVOLVED OVER THE LAST SEVERAL DECADES, NO GENERAL SYSTEM EXISTS THATALLOWS RECORD LINKING OF DATA SETS TO BE DONE WITHOUT AN EXTENSIVE AD HOC PROGRAMMING EFFORT. THIS PROJECT WILL DEMONSTRATE A RECORD-LINKING SYSTEM THAT IS INTEGRATED WITH A DATABASE MANAGEMENT SYSTEM AND ALLOWS RECORD-LINKING PROJECTS TO BE PERFORMED WITHOUT THE USER BEING INVOLVED IN DETAILS OF DATA MANAGEMENT. DURING PHASE I, DMS SYSTEMS, INC., WILL (1) DESIGN THE USER INTERFACE FOR THE RECORD-LINKING SYSTEM AND (2) IMPLEMENT A PROTOTYPE OF PART OF THE SYSTEM, FOCUSING ON THE STEPS OF SCORING AND CHOOSING FINAL LINKS. SOCIAL SECURITY RECORDS AND DEATH CERTIFICATES FROM CALIFORNIA WILL BE UTILIZED AS ATEST BED FOR THE SYSTEM. THESE RECORDS WERE USED AS PART OF AN EARLIER PROJECT, AND BOTH INTERMEDIATE RESULTS AND "TRUE" LINKS ARE AVAILABLE FOR COMPARISON WITH RESULTS OBTAINED WITH THE NEW SYSTEM.