The Canadian Cancer Registry at Statistics Canada provides a rich source of data on newly diagnosed cases of cancer in Canada. This data set is population based and national in scope. The wide variation in cancer rates, both regionally and over time, offer insight on how successful we have been in prevention and control of cancer and help assess the need for future programs.
In this exercise, data on female cancer have been provided in the attached tables, with the exception of regional breast cancer incidence, which is available online at Health Canada's Disease Surveillance web site. This web site provides examples of standard analyses and charts, as well as a glossary of terms.
The data provided in this exercise provide an opportunity to become familiar with statistical methods used in epidemiological research. Epidemiology is the study of patterns in disease. There are many methods available for analysis and presentation. The data set includes population, number of deaths attributed to breast cancer and the number of new cases of breast cancer by age group at the national level since 1969 and at the regional level since 1984.
The main cancer control initiative has been breast cancer screening in women over 50 years of age, and in some provinces in women over 40 years of age where a family history of breast cancer exists. Initiation of a screening program will typically increase the number of new cases for a short period of time, then the number of new cases will level off, or even return to previous levels.
Cancer Incidence Data
Newly diagnosed cases of cancer are registered with the provincial and territorial cancer registries. Cancer incidence data were provided to Health Canada from the Canadian Cancer Registry, formerly the National Cancer Incidence Reporting System, at Statistics Canada. [1, 2]
Significant under-reporting of cancer cases occurred in the earlier years of the NCIRS. Under-reporting in Quebec prior to 1977 has been estimated at approximately 30%, followed by over-reporting, particularly in 1981 and 1982, with further details available in The Making of the Canadian Cancer Registry. 
Statistics Canada conducts the National Population census every five years. The counts obtained from the census are slightly under estimated, so adjusted population estimates provided by Statistics Canada were used. Annual estimates for the intercensal years by 5-year age groups, sex, and census division are provided by Statistics Canada in publications [4, 5] and in electronic format.
Cause-specific mortality statistics are collected by registrars of vital statistics in the provinces and territories of Canada and published annually by Statistics Canada. 
Small counts: To preserve the anonymity of individuals, Statistics Canada has rules concerning the reporting of small counts. In the attached Tables, counts of 1 or 2 have been grouped and appear as "<=2" but null counts are reported as "0".
Canadian Institute for Health Information: The Graduate Student Data Access Program (GSDAP) at the the Canadian Institute for Health Information (CIHI) may be of interest. This program provides quality health data to eligible graduate students at no cost. CIHI is a leader in developing Canada's health information system. In that role, CIHI is responsible for the provision of accurate and timely health data needed for policy making, health care management, and public awareness of the factors affecting good health. More detailed information about the GSDAP (including application forms) is available from their web site at www.cihi.ca/wedo/gsdap.htm.
What does "STD CANADA 1971" mean? If we want to compare rates over time, and the population is changing, then we use age standardized rates. That is, we use a weighted average of the age specific rates, and in this case, the table was produced using the 1971 population for weighting. You should be able to reproduce the age standardized rates from the counts and population given. The usual standard today is the 1991 population, and the use of the 1971 standard was an oversight, though it should not impact the analysis of the data.
- Table 1: Female Population for CANADA, YEAR 1969 TO 1996, STD CANADA 1971
- Table 2: Number of Cases of female breast cancer, for CANADA, YEAR of diagnosis 1969 TO 1994, STD CANADA 1971
- Table 3: INCIDENCE RATES /100,000 FOR CANADA, YEAR 1969 TO 1994, STD CANADA 1971
- Table 4: Number of deaths for female breast cancer, for CANADA, YEAR 1969 to 1996, STD CANADA 1971
- Table 5: MORTALITY RATES /100,000 FOR CANADA, YEAR 1969 TO 1996, STD CANADA 1971
- Table 6: Number of Deaths for Female Breast Cancer, for CANADA and PROVINCES, YEAR 1984 TO 1996, STD CANADA 1971
- Table 7: Female Population for CANADA and PROVINCES, YEAR 1984 to 1996
- Table 8: MORTALITY RATES /100,000 FOR CANADA AND PROVINCES, YEAR 1984 TO 1996, STD CANADA 1971
- Get all eight tables in an Excel 97/98 & 5.0/97 Workbook [138K binary file]
- Get all eight tables in an Excel 97/98 & 5.0/97 Workbook [138K binary file - by ftp]
Epidemiology is the identification of patterns in disease: spatially, temporally and among various sub-populations. The research problem is to discern this pattern. There are many methods available for the analyses and presentation of the data. The researcher will have to consider different levels of aggregation, as well as the choice of statistical method.
Comparison of regional rates: Does the risk of dying from breast cancer differ across provinces/regions, after standardizing for population differences?
Comparison of regional trends: Are some provinces/regions doing a better job with cancer control?
Comparison of age-specific trends: Are risks increasing for some age groups and not others?
Age-period-cohort analyses: To attempt to discriminate between trends due to changes in risk factor exposure and changes in diagnosis and cancer registration procedures, Age-Period-Cohort (APC) analyses can be used. APC analyses can be used to discriminate statistically between trends due to period effects and trends due to cohort effects. What evidence do we have of a change in the risk of dying from breast cancer?
Age, socioeconomic status, nulli parity, first pregnancy after 30, late age at menopause, early age at menarche, family history of breast cancer, chest radiation in high doses are consistently associated with increased risk of breast cancer. Possible risk factors include sedentary lifestyle, hormone replacement therapy, oral contraceptive pills, alcohol, smoking, diet, obesity, and therapeutic abortion.
 Statistics Canada. Cancer in Canada. Catalogue 82-218, Annual to 1991. Ottawa: Health Statistics Division.
 Gaudette, Leslie A. and Judy Lee. 1997. Cancer Incidence in Canada, 1969-1993. Catalogue 82-566-XPB. Ottawa: Statistics Canada, Health Statistics Division.
 Band, Pierre R., Leslie A. Gaudette, Gerry B. Hill, Eric J. Holowaty, Shirley A. Huchcroft, Grace M. Johnston, Eva M. Makomaski Illing, Yang Mao, and Robert M. Semenciw. 1993. The Making of the Canadian Cancer Registry: Cancer Incidence in Canada and Its Regions, 1969 to 1988. Ottawa: Ministry of Supply and Services Canada.
 Statistics Canada. 1994. Revised Intercensal Population and Family Estimates, July 1, 1971-1991, Catalogue No. 91-537. Ottawa: Demography Division, Statistics Canada.
 Statistics Canada. 1996. Annual Demographic Statistics, 1995, Catalogue No. 91-213-XPB. Ottawa: Statistics Canada (received electronically, January 1996).
 Statistics Canada. Causes of Death. Catalogue 84-208-XPB, Annual. Ottawa: Health Statistics Division.