US childhood asthma prevalence estimates: The impact of the 1997 National Health Interview Survey redesign

The 1997 redesign of the National Health Interview Survey (NHIS) affected US childhood asthma prevalence estimates. The 1997 asthma attack prevalence estimate for children 0-17 years was 5.4%. Pre-redesign NHIS childhood asthma period prevalence estimates peaked in 1995 at 7.5%. It is unclear whether the difference reflects the change in survey methodology or changing asthma prevalence. To examine the impact of the NHIS redesign on childhood asthma prevalence estimates, the authors analyzed the 1988 NHIS that contained two sets of asthma questions: the core survey used until 1996 and the Child Health Supplement (CHS) with questions more similar to those in the redesigned 1997 NHIS. The authors measured the difference between 1988 core and CHS childhood asthma prevalence estimates to calculate an inflation factor for 1997-2000 NHIS estimates. The 1988 CHS questions produced asthma prevalence estimates 19-34% lower than the 1988 core question, depending on the methodology used to assess the difference. Inflating the 1997 asthma attack prevalence estimate by these differences yielded modified 1997 estimates ranging from 6.5% (95% confidence interval: 5.6%, 7.5%) to 7.3% (95% confidence interval: 6.4%, 8.2%). The change in the 1997 NHIS asthma questions likely explains much of the difference in asthma prevalence estimates between 1995 and 1997.