|
Activity Number:
|
9
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Sunday, August 8, 2004 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #301066 |
|
Title:
|
Use of Probabilistic Record Linkage for the Canadian Cancer Registry
|
|
Author(s):
|
Brad W. Thomas*+ and Jocelyne Marion
|
|
Companies:
|
Statistics Canada and Statistics Canada
|
|
Address:
|
R.H. Coats Bldg., 16th Floor, Ottawa, ON, K1A 0T6, Canada
|
|
Keywords:
|
probabilistic record linkage ; internal linkage ; unduplication ; external linkage
|
|
Abstract:
|
Cancer is a major cause of deaths in Canada. Both its detection and treatment place great burdens on the Canadian Health Care system. It is vital for the Canadian government to be able to estimate accurately the incidence of cancer in Canada each year, as well as be able to detect trends which indicate increases or decreases, in order for sufficient health care funds to be made available. Cases of cancer incidence in Canada are reported to the Provincial/Territorial Cancer Registries (PTCRs), who in turn pass this information onto Statistics Canada, for inclusion onto the Canadian Cancer Registry (CCR), a patient-oriented database that has cancer incidence in Canada recorded since 1992. This presentation focuses on estimating the incidence of cancer in Canada and the use of probabilistic record linkage to reduce cases of both underestimation and overestimation of cancer incidence. There are two linkages done in annual CCR production: (1) an internal linkage of the CCR to itself to detect possible duplicates of patients or tumours; (2) an external linkage of the CCR to the Canadian Mortality DataBase (CMDB) to find cases of cancer that were not detected until later.
|
