In complex survey designs such as NHANES, samples are not selected through simple random sampling. Instead, they are obtained through stratified, multistage, and oversampling designs. As a result, each survey participant represents a different number of individuals in the population. This difference must be corrected using sample weights. The NHANES program provides multiple types of weights for different survey components and subsamples, including interview weights, Mobile Examination Center (MEC) weights, subsample weights, and dietary recall weights. Correct use of these weights is essential to ensure that analysis results accurately represent the U.S. non-institutionalized civilian population. Otherwise, systematic bias or even incorrect conclusions may occur. In NHANES analysis, the selection of weights strictly depends on the variable with the smallest coverage in the analysis. In other words, once a specific subsample or survey module is involved, the corresponding weight must be used. This is the most fundamental principle in NHANES data analysis and also one of the most commonly overlooked aspects by beginners.

Read more

Some people simply take the sample they have and start calculating indicators right away. That’s understandable—but if you ignore the sampling design and the survey weights, your results can end up pretty far from the truth.

NHANES is a classic example. It doesn’t sample people purely at random. Some subgroups are intentionally oversampled so that we actually have enough data to analyze them. The problem is, if you analyze the raw, unweighted data, those groups end up having way more influence than they should.

Read more

In modern public health research, reliable data are the cornerstone of scientific analysis and policy making. The National Health and Nutrition Examination Survey (NHANES) of the United States is a globally renowned health survey program whose data are widely used to monitor the health status, nutritional levels, and disease prevalence of the U.S. population. However, NHANES does not rely on simple random sampling. Instead, it adopts a complex, multistage sampling design. This leads to a critically important concept—sample weights.

Read more

NHANES (National Health and Nutrition Examination Survey) is a comprehensive health survey project focusing on U.S. residents. Its data covers demographics, health status, laboratory tests, and more. By understanding the NHANES data structure and learning how to download, read, append, and merge its datasets, we can conduct better research and analysis using NHANES data.

1. Understanding NHANES Data Structure

1.1 Data Categorization

To analyze NHANES (National Health and Nutrition Examination Survey) data, understanding its structure is essential. Since 1999, NHANES has entered a “continuous survey cycle,” releasing data every two years. The data are divided into five major components, each containing multiple related files, facilitating efficient download and analysis. Detailed explanation of NHANES data structure:

Continuous NHANES refers to the two-year survey cycles released since 1999, such as 2015–2016 and 2017–2018. This continuity allows researchers to analyze trends over multiple years and perform detailed studies on specific cycles.

Each survey cycle’s data are divided into five main components:

1. Demographics

   Includes survey design variables (weights, stratification, primary sampling units) and demographic variables (gender, age, race/ethnicity, income, etc.).

2. Dietary

   Contains participants’ dietary intake data, including foods, beverages, and dietary supplements. Useful for nutritional analysis and dietary pattern research.

3. Examination

   Contains physical and dental examination data such as hearing tests, blood pressure, anthropometrics (height, weight, waist circumference), muscle strength, and vision tests.

4. Laboratory

   Includes lab results from blood, urine, hair, air, TB tests, household dust, and water samples (e.g., heavy metals, blood glucose, cholesterol, triglycerides).

5. Questionnaire

   Collected through household interviews and mobile exam centers, covering alcohol and drug use, social support, weight history, chronic disease diagnoses (diabetes, hypertension), and medication use.

Tip: NHANES data structure applies to each survey cycle, composed of the five components above. Comprehensive analysis often requires combining multiple components. For example, a cardiovascular disease study may need:

• Age and gender (Demographics)

• Blood pressure (Examination)

• Cholesterol (Laboratory)

• Hypertension diagnosis and medication use (Questionnaire)

By integrating these sources, researchers can perform holistic health and nutrition analyses, informing public health policy and interventions.

Read more

NHANES is an important public health data source and is widely used in epidemiological research, public health policy formulation, and clinical analysis. To help researchers efficiently obtain and process NHANES data, this article introduces several methods for downloading and reading NHANES data using R, including manually downloading data from the official website, reading local files with R, and directly downloading and processing online data using R. Combining these approaches can greatly improve the convenience and reproducibility of data analysis.

Downloading NHANES Data and Reading It with R

Each year, NHANES collects data from thousands of participants and makes it publicly available worldwide. The data are widely used in academic research, policymaking, and evaluation of public health interventions. Whether studying disease prevalence trends, analyzing relationships between nutrition and health, or assessing the effectiveness of public health policies, NHANES data provide rich support for researchers.

Although NHANES data are relatively easy to access, beginners may still find it challenging to efficiently download and manage these datasets. Below are several main download methods.

Why Choose NHANES Data?

NHANES data have several notable advantages:

1. Comprehensiveness: NHANES covers health, nutrition, behavior, environmental exposure, and many other indicators, meeting the needs of multidisciplinary research.

2. Authority: The data are collected and managed by the CDC and are highly credible and authoritative.

3. Public Accessibility: NHANES data are freely available, lowering the barrier to research.

4. Standardization: Data collection and processing strictly follow standardized procedures, ensuring high quality and comparability.

1. Manually Downloading NHANES Data from the Official Website

First, it is important to know that since the late 1990s, all publicly available NHANES data and related documentation have been released and updated in a single centralized location: the official NHANES website. NHANES datasets are labeled by survey cycle year. The website provides all publicly available data files and documentation, from the earliest National Health Examination Survey (NHES I) to the most recent datasets. Each data file includes codebooks and related documentation to help users understand and use the data.

To obtain NHANES data, visit the official NHANES website:

NHANES – National Health and Nutrition Examination Survey Homepage, which is part of the U.S. Centers for Disease Control and Prevention (CDC).

Read more

This article aims to introduce the application of secondary data analysis in medical and pharmaceutical research. It begins by outlining the importance and usage of primary data in medical research, followed by an explanation of the concept and advantages of secondary data analysis. The article then discusses research methodologies for joint mining of multiple secondary datasets, as well as approaches for data integration and analysis using clinical public databases. Special emphasis is placed on NHANES (National Health and Nutrition Examination Survey), a major public health database in the United States, with a detailed review of its historical development and data collection methods.

I. Understanding Secondary Data Analysis

1. Primary Data in Medical Research

In medical research, data form the foundation for answering research questions, testing hypotheses, and understanding specific health-related issues. Broadly speaking, data acquisition methods can be divided into two types: primary data and secondary data.

Primary data refer to data that researchers collect directly for a specific research purpose, whereas secondary data refer to data that have already been collected, organized, and made available by other researchers or institutions for reuse and reanalysis. Each data type has its own advantages and limitations, and researchers select the appropriate type based on the research question, budget, and available resources.

Primary data research in medicine refers to the process by which researchers directly collect and generate data themselves. The main advantage of this approach is that data collection can be tailored precisely to the researcher’s objectives, enabling highly targeted answers to specific research questions. Primary data can be collected through various methods, including questionnaires, interviews, laboratory tests, and field observations. For example, when studying the epidemiological characteristics of a disease, researchers may need to conduct field visits, interview patients, or perform direct clinical experiments to obtain relevant data.

The advantages of primary data lie in their high specificity and flexibility. Researchers can choose suitable collection methods according to their needs and ensure the accuracy and relevance of the data. However, primary data collection is often time-consuming and costly. First, it requires substantial investments of time, labor, and financial resources. Second, the data collection process may face challenges such as ethical approval and quality control. In particular, medical research involves patient privacy and sensitive information, which necessitates strict adherence to ethical standards and may be subject to legal and moral constraints.

Although primary data analysis can provide precise and direct answers, many researchers choose to rely on existing secondary datasets due to the complexity and high cost of primary data collection, especially when research resources are limited.

Read more