Information About The Proposed Elimination of the Family Questionnaire

The primary goals of the NHIS redesign are to ensure that the health topics being measured are consistent with the needs of the Department of Health and Human Services to monitor the nation’s health over the coming decades, and to reduce the respondent burden by significantly shortening the length of the questionnaire. Toward these ends, a design was proposed that features questions focused on the health of one randomly selected sample adult and one randomly selected sample child per household. The proposed approach will move critical family-level content from the family questionnaire to the updated sample adult and sample child questionnaires, such as family income, family poverty level, housing subsidies, household program participation, and financial burden of medical care. A page on the Impact of the Proposed NHIS Redesign on Family Questionnaire Content has more detail on how existing family-level content will (or will not) change in the redesign.

The proposed design retains other health-related topics addressed in the family questionnaire but collects this information about sample adults only. These include general health status, activity limitations, medical care that has been delayed or not received, and health insurance. Additional demographic and socioeconomic information that are proposed for annual data collection from the randomly sampled adult are: race, Hispanic origin, sexual orientation, English language proficiency, educational attainment, veteran status, marital status, nativity, citizenship, and employment.

The length of the current NHIS interview is contributing substantially to the erosion of the survey response rate and, if not shortened, could erode the quality of the resulting data. It is not possible to reduce the length of the questionnaire without cutting content. With the proposed elimination of the family questionnaire, however, several concerns have been expressed, and we appreciate having the chance to clarify.

Family context

The proposed design includes variables such as marital status of the sample adult (legal marital status, cohabitation, and same-sex marriage or partnership), family size, and the presence of children. We appreciate the concern that other types of information on family context may be needed for analyses of social determinants of health. Additional information is in the proposed sample child questionnaire, which is currently under review and will be available later this year. This additional information may include parents’ education, marital status, race, and employment status, among other topics. In the proposed design, we are also exploring ways to appropriately link and analyze the data from the sample child and the sample adult when those two persons form a parent/child dyad. This type of analysis is not currently recommended with NHIS public-use data files.

Some researchers may find that the family context measures in the proposed design are still not sufficient for their work, especially those researchers who examine family context and adult health. In this case, we welcome specific, focused, and evidence-based comments on these data needs (including how the data items have been used) and suggestions for ways to capture the relevant information within the sample adult or sample child interview.

Health insurance

We also appreciate the concern that analyses of the availability of coverage within the family unit will be limited by the elimination of the family questionnaire. The proposed design includes existing questions on the source of coverage, how it was obtained, and who helps pay for it, which may include family members within or outside the household. A new question was proposed on whether the coverage is self or family coverage and we are planning to add it to the redesigned survey. If these questions are not sufficient, we welcome additional specific, focused, and evidence-based comments on other insurance-related data needs.


We also appreciate the concern that the proposed elimination of the family questionnaire will reduce the annual sample size available to study health disparities for relatively small population groups, such as same-sex couples, people with disabilities, or immigrants. These special populations can still be identified in the proposed design. Moreover, most researchers who currently study detailed health outcomes for these groups with NHIS data were already working with the sample adult file and its associated smaller annual sample size. It is true, however, that researchers who only used the proxy-reported basic health information in the family section as their primary outcomes will require multiple years of data from the proposed design to achieve the annual sample sizes found in the 1997-2017 design.

Again, we appreciate the input that we have received and will take it into consideration as we plan next steps for the survey. Within a few months, we anticipate that we will distribute another iteration of the proposed questionnaire and will again request comments. However, if you have any additional comments or questions, please do not hesitate to contact us now at  We continue to invite the research community, federal partners, and NHIS data users to provide evidence-based comments using the guidance provided on the NHIS redesign website for consideration in the next round of revision of the NHIS content.