Why should you use caution when using crude rates to compare disease frequencies between populations? Give an example of how a systematic factor (e.g., gender, age) could affect crude rates in different populations.