Guidelines for High Quality Analyses of NHANES Data

First published: December 12, 2025

Analyses based on NHANES—a unique source of interview, dietary, exam and laboratory data on the health and nutritional status of adults and children in the United States—are widely published in scientific and medical journals. Recently, an exponential rise in NHANES-based studies has raised concerns about the publication of low-quality studies driven by freely available data and new artificial intelligence tools. NCHS remains committed to providing public access to data and transparent methods on data acquisition and processing.

Once NHANES data are released, oversight of the scientific integrity of analyses using NHANES falls to the analysts who conduct analyses and the institutions that publish these studies. Publications by CDC authors undergo rigorous scientific review prior to journal submission. For other authors and journal editors and reviewers, NCHS provides guidance about correctly analyzing NHANES data and assessing studies based on NHANES. NCHS will continue with the policies and support for dissemination of data and health estimates, including the following:

1. NHANES data are publicly available:
• NCHS is one of 13 principal Federal Statistical Agencies charged with disseminating data and statistical information to guide policies and action to improve the health of the American people. Releasing public use data files is an essential part of this mission.
• The Foundations for Evidence-based Policymaking Act of 2018, or "the Evidence Act," was established to improve access to data and expand evaluation capacity. This act requires agencies to make Federal data publicly available by default.
2. The National Center for Health Statistics provides detailed information about NHANES public use data files and analytic guidelines. Analysts should thoroughly review these resources.
• NHANES Questionnaires, Datasets, and Related Documentation includes information about survey weights, design variables, skip patterns, bridging equations and other details.
• Survey instruments and manuals:
  • Questionnaire instruments provide verbatim wording and skip patterns.
  • Laboratory manuals detail methods, materials, and instrumentation.
  • Procedure manuals provide interview, laboratory, and examination methods.
• A comprehensive online tutorial covers the basics in conducting accurate analyses. including sample computer code.
• Analytic guidelines and reports on survey methodology are provided. The most recent Guidelines are here: Brief Overview of Sample Design, Nonresponse Bias Assessment, and Analytic Guidelines for NHANES August 2021-August 2023.
• Published guidelines outline the steps to assess statistical reliability of estimates.
• The NHANES protocol undergoes ethical review by the NCHS Ethics Review Board including the brochures and consent documents used each cycle to inform participants of the survey content and seek their consent to participate.
3. There are basic elements of analyses that editors and reviewers should look for in manuscripts using NHANES data. This may require consulting NHANES documentation materials.
• Skip patterns: Failing to account for skip patterns in questionnaires and lab testing algorithms can result in inappropriate denominators for estimates. Researchers should specify the questions used in their analyses, include definitions of denominators in proportions, and assess whether sample sizes and estimates appear plausible.
• Bridging equations: Analysts presenting trend analyses of lab results should assess and document whether lab methods remained constant for the period studied. When lab equipment changed, NCHS conducted cross over studies to evaluate potential impact on results. If an effect was detected, bridging equations were provided in the documentation to account for the equipment change.
• Survey weights: NHANES is a population-based study and is designed to make inferences about the health of the US civilian noninstitutionalized population.
  • Analyses should use appropriate survey weights and specify which weights were used.
  • Using NHANES for unweighted analyses or as a convenience sample is generally not recommended given the clustered design (see next point). Inferences made using unweighted analyses are not generalizable to the US population.
• Survey design variables: Instead of choosing participants completely at random, NHANES data are collected using a complex design that involves visiting selected areas each year. Participants travel to a medical exam center which is only feasible if it is close to where they live. As a result, analysts must account for the clustered sample to correctly calculate measures of precision (e.g., confidence intervals). If the survey design variables are not incorporated using specific statistical software (e.g., SUDAAN or survey procedures in SAS or R), analysts will overestimate precision and could inaccurately conclude that differences between subgroups are significant.
• Statistical Reliability: NCHS has published guidelines outlining when estimates are considered statistically reliable. NCHS authors are required to abide by these guidelines. It is recommended that all authors consider statistical reliability when making inferences.
• Weighted population counts: It is not recommended to use the sum of the final sample weights for sample persons with the health condition of interest to calculate the number of people with the health condition. If there are exclusions or missing data for a health condition, summing the weights will underestimate the population estimate. (See NHANES tutorial for more details).

These resources are specific to correct analysis of NHANES data and do not include broader considerations for sound data analysis, such as defining a valid research question to fill knowledge gaps. However, the above resources can help authors, reviewers and editors provide high quality public health science.