Following is a response from DHS staff member, Tom Pullum:

Variables with a prefix s in the IR and KR files, or a prefix sh in the PR files, are "special" or "survey-specific". Because they are not standard, they don't have the same variable names wherever they appear. Often such questions eventually do become standardized. For example, questions about early child development are now standard but for several years they appeared with variable names and even coding structures that varied considerable from one survey to another.

Questions about chronic diseases are going through a similar evolution. I predict that eventually they will have a standard wording, coding, and names, but until then, you have to do some searching to find them. In Stata you can open a file and then, for example, enter "lookfor cancer" and you will find the variables that have "cancer" in the variable label.

If you are looping through surveys and want to do the same kind of analysis for each of them, but the variable names are different in the different surveys, you will need to include a segment for each survey to rename and/or recode the variables into consistent names and codes.