On March 19, 2020, NHIS temporarily became a telephone survey due to the coronavirus pandemic. Where possible, sample addresses were matched to telephone numbers using commercial lists and additional searches. As a result, the household response rate varied substantially by quarter due to the impact of the pandemic on in-person data collection. The household response rate declined from 60% in Quarter 1 to 43% in Quarter 2, and the sample skewed toward older and more affluent households. Personal visits to households resumed in selected areas in July and in all areas of the country in September. However, cases were still attempted by telephone first. Personal visits were used only to follow-up on nonresponse, to deliver recruitment materials, and to conduct interviews when telephone numbers were unknown (approximately 30% of interviews in Quarters 3 and 4). The household response rate was 49% in Quarter 3 and approximately 54% in Quarter 4. This telephone-first approach will continue for the foreseeable future.
NCHS chose to also field the 2020 NHIS questionnaire with a parallel sample that had known representativeness and nearly complete telephone contact information. That is, starting in August and continuing through the end of December, a subsample of about 20,000 adult respondents who completed the NHIS in 2019 were recontacted by telephone and asked to participate again. These follow-up interviews will also permit data users to look at health, health care, and well-being from before and during the pandemic, for the same individuals. The completion rate for this longitudinal follow-up survey was between 50% and 60%.
As a result, NHIS fielded four designs in 2020: Normal production in Quarter 1, telephone-only in Quarter 2, telephone-first in Quarters 3 and 4, and the longitudinal follow-up. The Q3 and Q4 samples were reduced in size by approximately half in order to free up resources for the longitudinal follow-up. Our ongoing challenges in 2021 include figuring out how to use weighting and estimation techniques to produce official 2020 estimates from the disparate pieces, each with its own coverage and nonresponse issues. The 2020 NHIS public-use data files and documentation are tentatively planned for release in Fall 2021.
