US cancer mortality data derived from information recorded on death certificates are frequently relied upon as an indicator of progress against cancer. A limitation of this measure is the lack of information pertaining to the onset of disease, such as year-of-diagnosis, age-at-diagnosis, stage of disease at diagnosis, and histology of lesions. For example, death certificate mortality represents patients who were diagnosed at any year in the past, regardless of how long ago it was. If a new treatment affects patients diagnosed starting in a certain year, mortality for patients diagnosed after that point should show potential improvements associated with the treatment more clearly than overall mortality. It is also of interest to partition mortality by stage at diagnosis, in order to determine the proportion of cancer deaths associated with each stage at diagnosis.
Population-based cancer registries collect these types of data and allow the calculation of an incidence-file based mortality rate. This incidence-based mortality rate (IBM) allows a partitioning of mortality by variables associated with cancer onset. Accurately measuring IBM requires high quality population-based cancer registry data and high quality follow-up of cancer patients for vital status including cause of death. In analyzing incidence based mortality, one must be cautious in interpreting the results, since factors like lead-time bias can influence these analyses, whereas they generally will not influence overall death certificate mortality.
Examples of analyses that can be conducted using IBM include:
- Measure of the efficacy of an intervention based on comparing IBM subsequent to the introduction of an intervention with appropriate historical patterns of IBM.
- Assess the impact of an intervention on IBM partitioned by stage at diagnosis.
An introduction and overview of incidence-based mortality can be found in:
Chu KC, Miller BA, FeuerEJ, Hankey BF. A method for partitioning cancer mortality trends by factors associated with diagnosis: An application to female breast cancer. J Clinical Epidemiology 1994:47(12);1451-61.
Using SEER*Stat to Calculate Incidence-Based Mortality
Incidence-based mortality rates can now be calculated in SEER*Stat. These calculations require databases containing cancer incidence data which are linked to populations via age and year of death. Only patients who have died by the study cut-off date are included in the databases. You must use SEER*Stat in client-server mode to access SEER's incidence-based mortality databases. These databases include "incidence-based mortality" in the database name.
The following is a sample exercise for using SEER*Stat to create two tables showing incidence based mortality.
Incidence-Based Mortality References
Chu KC, Tarone RE, Freeman HP. Trends in prostate cancer mortality among black men and white men in the United States. Cancer 2003 Mar 15;97(6):1507-16.
Merrill RM, Lyon JL. Explaining the difference in prostate cancer mortality rates between white and black men in the United States. Urology 2000 May;55(5):730-5.
Merrill RM, Stephenson RA. Trends in mortality rates in patients with prostate cancer during the era of prostate specific antigen screening. J Urol 2000 Feb;163(2):503-10.
Feuer EJ, Merrill RM, Hankey BF. Cancer surveillance series: interpreting trends in prostate cancer--part II: Cause of death misclassification and the recent rise and fall in prostate cancer mortality. J Natl Cancer Inst 1999 Jun 16;91(12):1025-32.