Please contact Health Statistics & Data Management with any questions you may have. 

National cancer incidence estimates can be found at the National Cancer Institute's Surveillance, Epidemiology and End Results program; data from other states' cancer registries can be found at the North American Association of Central Cancer Registries.

The Centers for Disease Control and Prevention's BRFSS national program site has data from other states and medians for the states.

A 95% confidence interval represents the range of values that, with 95% certainty, include the true value for the entire population.  An estimate of a rate is only truly useful and applicable if the 95% CI is also considered.  This is especially true of sample data like that from the Behavioral Risk Factor Surveillance System (BRFSS),  that estimates a population value from a sample of people interviewed.  The need for a confidence interval also applies to rates based on NH event data, like birth or cancer data.  Even though nearly complete data is collected on event data (i.e., it is not a sample), it is still subject to random variation.  When the sample or number of events is very small, the 95% CI will be wide.  This is because the small sample is less likely to accurately represent the whole population or the small number of events is likely to be highly variable.

A 95% CI can be used to determine whether the difference between two groups, areas or time periods is statistically significant.  If the range of values in the 95% CIs do not overlap between groups, the difference between the groups is "statistically significant."  If the 95% CIs do overlap (i.e., if the CIs share any of the same values), there is no detected statistically significant difference.  This could mean that, in fact, no difference exists between the groups, or it could mean that a difference does exist, but was not detected due to small sample size or number of events.  If very few people or events were included in the group being considered, the 95% CI will likely be very wide.  It is possible that with a better measure of that group—more people in the sample or an aggregate of events over multiple years—producing a more accurate estimate, will detect differences between groups.

Rates need to be calculated with at least 20 events in the numerator.  Calculating a rate based on less than 20 events in the numerator creates an unstable estimate that is not statistically reliable and varies greatly from year to year by chance alone.  Therefore, generating rates based on small numbers can lead to misinterpretation.  For this reason it is often helpful to combine years of data to create a more stable rate.

To compare groups where proportions of age groups are different, an adjustment needs to be made to make the age groups comparable.  For example, the rate of lung cancer in NH may look higher than that of the US.  However, if NH has fewer younger people than the US as a whole, its rate will appear higher.  Once the two rates are adjusted to control for the differences in age distribution, the rates may not look different at all.  Rates for the entire population that are not age-adjusted are called crude rates.

For individual copies of birth, death, and marriage certificates, contact the Division of Vital Records Administration or visit the National Center for Health Statistics.