First, proxy reporting about health conditions by one household informant for other household members was eliminated to improve data quality. Instead, one sample adult and one sample child (for households containing children) are now randomly chosen to respond to questions about health conditions. As has always been the practice for the NHIS, a knowledgeable adult, usually a parent, responds in proxy for children under age 18 years.

A second change involved streamlining data collection on health conditions. Prior to the redesign, over 100 health conditions were divided among six separate lists. Each list was presented to one sixth of the survey sample. In the redesigned NHIS, these separate lists were eliminated and each sample child and sample adult receive all the health condition questions, although the number of conditions asked about in the survey is now more limited. These changes resulted in a larger annual sample of children and adults with responses to the asthma questions and thus a smaller sampling error for asthma prevalence estimates.

A third change involved the question wording for chronic conditions. Respondents are now asked if they have ever received a physician diagnosis before they are included in the case definition of having that condition. This was done to improve validity. But it also created a break in the trend because the new case definitions of diseases are different than those used prior to 1997.