Methods and Software: Cancer Incidence & Mortality

A cancer incidence or mortality rate is the number of newly diagnosed cancers or number of reported cancer deaths of a specific site/type occurring in a specified population during a year (or group of years), usually expressed as the number of cancers per 100,000 population at risk.

Several statistical methods and software tools have been developed for the analysis and reporting on cancer incidence and mortality statistics.

Delay Model: Cancer Incidence Rates Adjusted for Reporting Delay

Timely and accurate calculation of cancer incidence rates is hampered by reporting delay, the time elapsed before a diagnosed cancer case is reported to the NCI. The idea behind modeling reporting delay is to adjust the current case count to account for anticipated future corrections (both additions and deletions) to the data. See Cancer Incidence Rates Adjusted for Reporting Delay for more information.

SEER*Stat Statistical Methods

SEER*Stat is statistical software for the analysis of SEER and other cancer-related databases. The following methods associated with the reporting of basic cancer incidence statistics are added directly to SEER*Stat.

• Age-Adjusted Rates - An age-adjusted incidence or mortality rate is a weighted average of the age-specific incidence rates, where the weights are the proportions of persons in the corresponding age groups of a standard million population.
• Gamma Confidence Intervals for Age-Adjusted Rates: This method provides confidence intervals with good statistical properties even when the cancer is rare, the population is small, or the population of interest is very different from the standard.
  Reference: Fay MP, Feuer EJ. Confidence intervals for directly standardized rates: A method based on the Gamma distribution. Stat Med 1997;16:791-801.
• Trends in Rates - Trends over time based on frequencies (percent change, annual percent change). The average annual percent change is used to measure trends or the change in rates over time. For information on how this is calculated, go to Trend Algortihms in the SEER*Stat Help system.
• Risk-adjusted Incidence Rates - The usual SEER incidence rates which are reported count multiple instances of primaries of the same cancer in the numerator, and use the total population as the denominator. However, in many situations, interest focuses instead on first instances of a particular cancer type and the population who has never had the cancer as the denominator. This type of incidence rate is sometimes called "Risk-Adjusted Incidence" since it represents the risk estimate that would be used in a population model for the transition rate from the healthy population to the population with cancer. First instances of a particular cancer can be derived directly from SEER data (under certain assumptions), while the size of the cancer-free population must be derived indirectly by subtracting out the population with prevalent cancers. These rates are sometimes lower than the standard rates (e.g. prostate cancer with high prevalence and almost no instances of multiple cancers), sometimes higher than the standard rates (e.g. melanoma with low prevalence and many instances of multiple cancers), and sometimes very close to the standard rates (e.g. breast cancer with relatively high prevalence and relatively high instances of multiple cancers). For more information on this topic, see: Merrill RM, Feuer EJ. Risk-adjusted cancer-incidence rates (United States).Cancer Causes Control 1996 Sep;7(5):544-52.
• Multiple Primary - Standardized Incidence Ratios (MP-SIR): These ratios can be used to compare incidence of cancer in a defined cohort of persons previously diagnosed with cancer to the incidence of cancer in the general population.
• Incidence-based Mortality - The incidence-based mortality rate allows a partitioning of mortality by variables associated with the cancer onset.

Joinpoint Trend Analysis Software

Joinpoint is statistical software for the analysis of trends using joinpoint models, that is, models with several different lines that are connected at the "joinpoints." The software takes trend data (e.g., cancer rates) and fits the simplest joinpoint model that the data allows. This enables the user to test whether an apparent change in trend is statistically significant. See Joinpoint Regression Program for more information and to download the software.

DevCan - Probability of Developing or Dying of Cancer

The DevCan software uses lifetable methods to compute the lifetime and age-conditioned probability of developing cancer and dying of cancer in the general population. Input data for the computations include cancer incidence rates, cancer mortality rates, and all-cause mortality rates. Data sets are supplied to estimate risks of developing and dying of cancer for over 20 cancer sites by race and sex. The Probability of Developing or Dying of Cancer Web site provides access to the DevCan software as well as links to published statistics and documentation of the methodology. In addition, the Cancer Query Systems: Probability of Developing or Dying from Cancer provides access to a database of pre-calculated statistics.

Predicting Current Year Rates

Extension of the spatial statistical model to also model rate changes over time can be used to fill in the data gaps to provide estimates of cancer incidence in all U.S. counties for each year. These spatio-temporal estimates can be used to project rates or counts ahead in time. This method will be used by the American Cancer Society to estimate the numbers of new cancer cases in the coming calendar year beginning in 2007.

Pickle LW, Hao Y, Jemal A, Zou Z, Tiwari RC, Ward E, Hachey M, Howe HL, Feuer EJ. A new method of estimating United States and state-level cancer incidence counts for the current calendar year. CA Cancer J Clin 2007 57:30-42. [Full Text] [PDF]