Many diseases are rare, but occur in large numbers when considering national, state, and county levels. However, disease counts at a city, municipality level may be quite small, especially when evaluating case counts by gender or race and ethnicity.

Many agencies do not report case counts for indicators when the fewer than 5 cases occurred; nor are rates calculated. Suppression (or not reporting) these cases and rates are done for many reasons.

First, the case count is suppressed due to privacy considerations, where it may be possible to determine the identity of the person(s) affected.

Second, by suppressing case counts, rates cannot be calculated by data users.

Rates calculated with less than 5 events are considered "unstable", meaning that these few cases would not fairly reflect the the underlying population. 

The National Center for Health Statistics recommends that rates should not be calculated when this happens (less than 5 cases per time period per area.) Not only are these rates unstable, but privacy considerations must be considered, especially when rates are applied to a very small geographic area such as a block or census tract.

One method to avoid having these rare rates suppressed is to combine counts across several years and calculate the average annual rates. This methods has several advantages:

• Rates of rare events can be calculated as an average annual rate.
• These rates are more "robust" (a statistical term), meaning that the rates are more stable and are less sensitive to year-to-year variation.
• Suppression of counts of these rare events is not needed.
• Trends can be analyzed with greater confidence.