The objective of the National Health Interview Survey (NHIS) is to provide data that can be used to monitor the health of the US population.

In this study, we evaluate whether the disability questions currently used in the NHIS – the Washington Group questions – threaten the ability of this survey to fulfil its stated objective for disabled people.

Data were from the 2011–2012 NHIS with linkage to mortality status through 2019. We examined whether people who reported a disability in the American Community Survey (ACS) disability questions had their disability counted in the Washington Group (WG) questions. We then examined the consequence of use of the WG, as opposed to the ACS questions, on estimates of disability prevalence and comparative mortality risk.

We find that when compared to their predecessor, the American Community Survey disability questions, the Washington Group questions accounted for less than half of disabled people, primarily counting disabled people with more than one disability status, but not counting many disabled people with only one disability status. As a result of this undercount, disability prevalence rates based on the Washington Group questions underestimate the size of the disabled population in the US, and overestimate the comparatively higher mortality risk associated with disability status.

These results underscore the need to re-evaluate the disability questions used in the NHIS, and invest in the development of improved and expanded disability questionnaires for use in national surveys.