Some data sources that provide an estimate for an indicator may provide a confidence interval with a lower or upper limit that are outside logical bounds. For example, if a population estimate is near zero, the calculated value of the lower confidence limit may be negative. However, a negative number of people does not make sense. The negative number should be interpreted as zero. A similar instance is seen when estimates are close to or at 100% and the upper confidence limit is over 100%. These upper confidence limits should be interpreted as 100%. These data sources may be using methods to calculate confidence intervals meant for estimates based on larger samples of a population and are not adjusting confidence intervals to be within more logical or "natural" limits. These estimates with illogical confidence intervals are likely to be found in geographies with small overall or subgroup populations. Estimates based on small sample sizes are less reliable. Exercise caution when using or interpreting these estimates.