Understanding Cancer Incidence and Mortality Rates
All cancer registries use the International Classification of Diseases for Oncology, Third Edition (ICD-O-3) to code the anatomic site and morphology. Cancer incidence statistics include invasive cancers only, with the exception of in situ cancer of the bladder. Mortality rates are based on the underlying cause of death coded using the International Classification of Diseases, Tenth Edition (ICD-10). Cancer incidence rates in WISH represent the number of new cases of cancer per 100,000 population. Cancer mortality rates represent the number of cancer deaths per 100,000 population during a specific time period. Cancer incidence and mortality rates can be adjusted for demographic variables such as race, age, and sex. The most commonly used adjustment for cancer rates is age.
Crude rates are helpful in determining the cancer burden and specific needs for services for a given population, compared with another population, regardless of size. Crude rates are calculated as follows:
A crude incidence rate equals the total number of new cancer cases diagnosed in a specific year in the population category of interest, divided by the at-risk population for that category and multiplied by 100,000. A crude death rate equals the total number of cancer deaths during a specific year in the population category of interest, divided by the at-risk population for that category and multiplied by 100,000.
Crude Rates vs. Age-Adjusted Rates. Crude rates are influenced by the underlying age distribution of the state's (or other locality's) population. Even if two states have the same age-adjusted rates, the state with the relatively older population generally will have higher crude rates because incidence or death rates for most cancers increase with age. The age distribution of a population (i.e., the proportion of people in particular age categories) can change over time and can be different in different geographic areas. Age-adjusting the rate ensures that differences in incidence or deaths from one year to another, or between one geographic area and another, are not due to differences in the age distribution of the populations being compared.
Older age groups generally have higher cancer rates than younger age groups. To address this issue for purposes of analysis, most cancer incidence and mortality rates in major publications have been age-adjusted. This removes the effect of different age distributions between populations and allows for direct comparison of those populations. Age-adjustment also allows for the comparison of rates within a single population over time. The direct standardization method of age adjustment weights the age-specific rates for a given gender, race, or geographic area by the age distribution of the standard 2000 U.S. population.
There are three major components used to calculate age-adjusted rates: the number of cases or deaths reported, the population, and a "standard" population. A rate (new cases or deaths per 100,000 population) is first computed for each age group, then each of these age-specific rates is weighted by multiplying it by the proportion of the 2000 U.S. standard population for that same age group. The results from each age group are added to arrive at the age-adjusted rate for the total population.
An age-adjusted rate should only be compared with another age-adjusted rate that was calculated by the same method, using the same U.S. standard population. Starting with all 1999 data, the National Center for Health Statistics (NCHS) and the National Cancer Institute (NCI) began using the year 2000 U.S. standard million population age distribution reported by the Census Bureau. Cancer incidence increases with age, and because the 2000 population was older than the 1970 population, the change to the 2000 U.S. standard population resulted in apparently higher rates for most cancers. Caution should be used when comparing the data in this report with cancer incidence rates adjusted to standard populations other than the 2000 U.S. standard population.
The population estimates used in the WISH cancer modules are based on SEER population data (exit DHS) that also incorporate new intercensal bridged single-race estimates derived from the original multiple race categories as specified by the Office of Management and Budget for the collection of data on race and ethnicity. The bridged single-race estimates and a description of the methodology used to develop them are on the National Center for Health Statistics Web site (exit DHS).
Age-adjusted incidence and mortality rates are grouped by primary cancer site (site of origin) per 100,000 population. For cancers that occur only in one sex (prostate, uterine, cervical, female breast), sex-specific population denominators are used to calculate incidence and mortality rates. Incidence rates are for invasive cancers; the only exception is the incidence rate for urinary bladder, which includes both in situ and invasive cancers. Cancer incidence rates may include multiple primary cancers that occur in single patients; each cancer is counted as a separate case if a patient has more than one primary cancer.
Confidence intervals for the age-adjusted rates were calculated with a method based on the gamma distribution (modified by Tiwari, et al., 2006). This method produces valid confidence intervals even when the number of cases is very small. When the number of cases is large, the confidence intervals produced with the gamma method are equivalent to those produced with more traditional methods. The formulas for computing the confidence intervals can be found in the report, Tiwari RC, Clegg LX, Zou Z. Efficient interval estimation for age-adjusted cancer rates. Stat Methods Med Res 2006 Dec;15(6):547-69.
The cancer incidence and mortality counts derived from cancer registries and vital records are complete enumerations of information rather than samples (used in most research studies) and are, therefore, not subject to sampling error. The rates based on those counts are, however, subject to what is termed "random error," which arises from random fluctuations in the number of cases over time or between different communities. The 95 percent confidence intervals are an easily understood way to convey the stability of the rates. A stable rate is one that would be close to the same value if the measurement were repeated. An unstable rate is one that would vary from one year to the next due to chance alone. A wider confidence interval in relation to the rate itself indicates instability. On the other hand, a narrow confidence interval in relation to the rate tells you that the rate is relatively stable, and you would not expect to see large fluctuations from year to year. If differences are observed between stable rates (those with narrow confidence intervals), it is likely that the differences represent true variations rather than random fluctuations in the number of cases.
Last Revised: February 12, 2014