This report presents various statistics on net cancer survival among Pennsylvanian adults and is divided into the following chapters:
Explanation of key terms and concepts
A brief overview of how cancer affects survival in the state
Patterns and changes in net cancer survival from 2001 to 2014
The effect of stage at diagnosis on survival
The effects of poverty, race and insurance coverage on net cancer survival
Net cancer survival by county and health district
Details on the collection, analysis and presentation of data in the report
Tables of net cancer survival estimates for each primary cancer site by year of diagnosis, race, poverty level, etc.
Additional information on cancer survival
Statisticians use survival analysis to measure what proportion of a group will be alive a certain amount of time after an event, otherwise known as a survival rate. Survival rates are key measures for gauging the health of various segments in society.
Survival analysis first estimates the risk of dying, called the hazard, for small intervals such as individual days. These hazards are then combined to calculate the total hazard for a longer period, such as the five years after a cancer diagnosis. The survival rate is an easily understood way to show this total hazard.
Net survival rates show how much a condition, such as cancer, increases a group’s hazard of death. The underlying assumption is a person’s hazard can be split into two parts:
Some notes on interpreting net survival in this report:
Those diagnosed with cancer differ from the general population in ways that were not, and maybe could not be, accounted. Income, lifestyle and health habits are believed to be major unaccounted differences.
A population’s net survival rate comes from a weighted average of its members’ cumulative hazards. It is assumed there is a true average for the group that one can only estimate. A confidence interval shows the uncertainty of an estimate. Estimates for larger populations have more certainty.
Throughout this report, net survival estimates are usually presented with their 95% confidence intervals. This means, for each confidence interval, there is a 95% chance the true rate falls inside.
If a difference is said to be statistically significant, or just “significant,” that means the stated number lies outside the estimate’s confidence interval. To compare two estimates, their difference is considered an estimate. If the confidence interval about the difference does not include zero, then the difference is statistically significant.