You are using an outdated browser. We suggest you update your browser for a better experience. Click here for update.
Close this notification.
Skip to main content Skip to search

Methodology 1.1.c

Cancer in Ontario Methodology
Short description of Indicator Five-year relative survival ratio: the proportion of people alive five years after their diagnosis, divided by the survival expected among people of the same age, sex, and time period in the general population.
Rationale for Measurement Relative survival compares the survival experience of individuals with cancer to that of the general population of the same age and sex. It shows the extent to which a diagnosis of cancer shortens a life span. As an indicator of both disease severity and prognosis, relative survival is a popular indicator for several aspects of cancer control, including monitoring progress in treatment and early detection (i.e., screening).
Evidence/references for rationale This is a well-established cancer burden indicator, described and reported on by numerous provincial, national and international organizations, including:

Additional evidence supporting the utility of this measure is described in:

  • Dickman PW, Adami HO. Interpreting trends in cancer patient survival. J Intern Med. 2006; 260(2):103–17.

Mariotto AB, Noone AM, Howlader N, Cho H, Keel GE, Garshell J, et al. Cancer survival: an overview of measures, uses, and interpretation. J Natl Cancer Inst Monogr. 2014; 2014(49):145-86.

Calculations for the indicator Five-year relative survival ratio (age-standardized):
  • For cases diagnosed in a historic five-year time period, with at least five years of follow-up data available, five-year relative survival is calculated using the cohort method. The cohort method follows the survival of a group of patients diagnosed in a single calendar year or span of years.1
  • For cases diagnosed in the most recent five-year time period, for which five years of follow-up data are not available for all cases, five-year relative survival is estimated using period method. The period method observes, in the most recent time period possible, the survival of patients diagnosed in different calendar years, to estimate the survival expected for recently diagnosed cases.1
  • Cases are censored when they reach 100 years of age, to adjust for potential data limitations, such as the fact that many people die in other jurisdictions without notification to Ontario.
  • Cases with unknown age at diagnosis, diagnosed on the basis of an autopsy only, and those whose date of diagnosis and date of death are the same (that is, who were only diagnosed at or following death) are excluded.
  • Relative survival ratios are age-standardized by first calculating survival for 5 age groups (15-44, 45-54, 55-64, 65-74, 75-99). The age-specific survival estimates are then weighted using the International Cancer Survival Standard weights.2 The exception is prostate cancer, for which 5 different age groups are used (15-54, 55-64, 65-74, 75-84, 85-99) and the standard weights are summed for the 15-44 and 45-54 age groups.
  • Adjustment for the expected mortality of people of the same age in the general population of Ontario is carried out using Ontario life tables.
  • Relative survival estimates are based on one cancer case per SEER recode site group, per person. Multiple primary cancers that are captured by the Ontario Cancer Registry's new National Cancer Institute Surveillance Epidemiology and End Results (NCI SEER) program standards for counting multiple primary cancers, which were adopted by the Ontario Cancer Registry for cases diagnosed in 2010 and beyond, were excluded. If a person still has more than one cancer case per site after multiple primary cancers are removed, the following criteria are used to determine which cancer case should be selected:
    • If the cases were diagnosed within 30 days of each other, the case with the highest stage is selected. If none of the cancer cases had a valid stage, then the earliest case is selected. For colorectal cancer cases only, if none of the cancer cases has a valid stage, the topography of the cases is also considered before selecting the earliest case; if one case has a ICD-O-3 topography code of C18 (colon excluding rectum) or C20 (rectum) and the other case(s) has a topography code of C19 (rectosigmoid junction), the C18/C20 case is selected.
    • If the cases were diagnosed greater than 30 days apart, the earliest case is selected.

    Analysis:

    • Cases diagnosed in two time periods: 1999-2003 and 2009-2013.
    • Calculated for 14 common cancers, Ontario: thyroid, prostate, melanoma, female breast, uterus, kidney, bladder and urothelial cancer, non-Hodgkin lymphoma, colorectal, oral cavity and pharynx, leukemia, stomach, lung and pancreas.
    • Calculated for the 4 most common cancers (Prostate [ICD-O-3 code: C61.9], female breast [ICD-O-3 code: C50.0–C50.9], colorectal [ICD-O-3 code: C18.0–C20.9, C26.0] and lung [ICD-O-3 code: C34.0–C34.9]) by Local Health Integration Network (LHIN).
    • Survival was measured from the date of diagnosis to: the date of death; the date of the 100th birthday; or December 31, 2008 (for cases diagnosed in 1999-2003) or December 31, 2013 (for cases diagnosed in 2009-2013), if alive.
    • For Ontario estimates, statistically significant differences in the five-year relative survival estimates between 1999-2003 and 2009-2013 for a given cancer were tested by comparing the absolute difference between the two estimates with the square root of the sum of the squared margin of error (i.e., the upper 95% confidence limit minus the estimate) for the two estimates. If the difference between the two estimates was greater than the square root of the sum of the squares of their margin of errors, then the estimates were considered significantly different (approximately p<0.05).
    • For the LHIN estimates, five-year relative survival estimates for 2009-2013 for a given cancer type were compared between LHINs by first ranking the estimates from lowest to highest. The LHINs with the lowest four estimates and the highest four estimates were noted, although these estimates may not necessarily be significantly higher or lower than the Ontario estimate.
  • References:
    1. Brenner H, Gefeller O. Deriving more up-to-date estimates of long-term patient survival J Clin Epidemiol 1997;40:211–6.

    Corazziari I, Quinn M, Capocaccia R. Standard cancer patient population for age standardising survival ratios. Eur J Cancer 2004;40:2307–16.

Standardized Rate Calculation N/A
Unit N/A
Data sources Ontario Cancer Registry, 2016 (Cancer Care Ontario)
Time Frame N/A
Geographic Scale Provincial, LHIN
Denominator description N/A
Numerator description N/A
Considerations 
  • On Oct. 29, 2014, CCO’s Ontario Cancer Registry Information System (OCRIS) was formally decommissioned and replaced with the new Ontario Cancer Registry (OCR). The new registry brings the OCR in line with current Canadian and U.S. standards for tracking cancer incidence. The OCR now conforms to specific standards as set out by the National Cancer Institute’s Surveillance, Epidemiology and End Results (NCI SEER) program for counting multiple primary cancer sites, which most Canadian provinces and U.S. states now use. The adoption of specific NCI SEER standards with the new OCR has resulted in an increase in the incidence number of certain types of cancer reported in Ontario. To mitigate the impact of this change on survival estimates, relative survival estimates are based on one cancer case per SEER recode site group, per person as noted in the calculations section above.
  • Cancers were defined using U.S. Surveillance, Epidemiology and End Results (SEER) Recode definitions: see http://seer.cancer.gov/siterecode/icdo3_dwhoheme/index.html. Cancer definitions using SEER Recode may differ, especially for cancers of the colon and rectum and lung cancer, from definitions in other published analyses of survival.
  • For most cancer types (except urinary bladder and kidney), the full site grouping name outlined by the SEER site recode variable definitions was used. Short titles for certain cancers, however, were, used for graphing purposes. Cancer types for which short titles were displayed are as follows:
    • Lung and bronchus (ICD-O-3 code C34.0–C34.9): short title “Lung”
    • Melanoma of the skin (ICD-O-3 code C44.0 with histology codes 8720–8790): short title “Melanoma”
    • Corpus and uterus, NOS (ICD-O-3 code C54.0–C54.9, C55.9): short title “Uterus”
    • Colon and rectum (ICD-O-3 code C18.0–C18.9, C19.9, C20.9, C26.0): short title “Colorectal”
    • Urinary bladder (ICD-O-3 code C67.0–C67.9): short title “Bladder”
    • Kidney and renal pelvis (ICD-O-3 code C64.9, C65.9): short title “Kidney”
  • Age-standardizing five-year relative survival ratios and conditional five-year relative survival ratios allows for comparison across time periods and jurisdictions by adjusting for differences in the age-distributions of the populations of interest.
  • Due to changes to the Ontario Cancer Registry, which were implemented for cases diagnosed in 2010 and beyond, comparisons of the Ontario five-year survival estimates for 1999-2003 and 2009-2013 should be interpreted with caution. For the LHIN estimates, comparisons over time should not be made.
  • Relative survival is a useful population-based indicator of the burden of cancer and the variation in severity of different types of cancer. It does not necessarily reflect a person’s chance of surviving for a given time after diagnosis.
  • The Ontario Cancer Registry does not actively follow cases and so deaths may be missed. This, and the exclusion of individuals whose date of diagnosis is their date of death, may lead to overestimates of survival.
  • Relative survival estimates can be biased and can over- or under-estimate survival if there is a mismatch between the life table and the cancer patient cohort.1 In some cases, age-standardized relative survival estimates in excess of 100% are possible, suggesting that survival in the cancer patient cohort was higher than the expected survival in a comparable group from the general population, as measured by national life tables. This can happen when information on deaths is missing in the registry or when life tables do not reflect the background mortality experience of the population from which the cancer patient originated, or when cancer patients are successfully treated for their cancer and other comorbidities and, as a result, adopt a healthier lifestyle.

References:
Ellison LF. Estimating relative survival for cancer: An analysis of bias introduced by outdated life tables. Health Rep 2014; 25:13-19.

Data availability & limitations Five-year relative survival ratios are shown through 2013, the most recent year for which the Ontario Registry has received complete data at the time of analysis.
CSQI Year 2018