SAS Programs to Analyze NHANES Accelerometer Data

Abstract

The SAS Programs to Analyze NHANES Accelerometer Data were developed by the National Cancer Institute (NCI) to facilitate the standardized processing and analysis of physical activity data collected during the National Health and Nutrition Examination Survey (NHANES), specifically from the 2003–2004 cycle. These programs are essential tools for researchers seeking to derive meaningful summary variables, such as per-day and per-person activity metrics, from complex, raw accelerometer data. A key feature of this resource is the inclusion of a utility program capable of converting ActiGraph data files into the specific layout required for compatibility with the NHANES processing methodology, thereby broadening the applicability of the NCI’s analytical framework. Users should note that intermediate-level SAS programming skills are necessary for modifying the provided code, and technical support from the NCI is limited.

Keywords

NHANES, SAS Programs, Accelerometer Data, Physical Activity Monitoring, Data Processing, National Cancer Institute, ActiGraph, Public Health Data.

Authors

National Cancer Institute (NCI)

Purpose

The primary purpose of these SAS Programs is to standardize the often complex process of preparing and summarizing large datasets derived from accelerometers used in major public health surveys like NHANES. These programs automate crucial data edits, including identifying and managing non-wear time and cleaning artifacts, which are necessary steps before calculating meaningful summary variables. The resulting per-day and per-person variables are vital for epidemiological research concerning physical activity and health outcomes.

Furthermore, the programs offer flexibility. Users possessing intermediate-level SAS programming skills are explicitly encouraged to modify the code. This adaptability allows researchers to implement alternative definitions for valid wear time, monitor specific activity bouts differently, or apply custom data filtering criteria, ensuring the methodology aligns precisely with their specific research hypotheses. The inclusion of a program to convert data from other devices, specifically ActiGraph accelerometers, ensures the NCI preparation programs can be utilized across various datasets employing similar measurement devices.

Construct

As a set of data processing tools, these SAS Programs do not directly measure a psychological construct. Instead, they facilitate the measurement and quantification of Physical Activity (PA) and sedentary behavior—behavioral constructs derived from objective data. The programs process raw accelerometer counts into metrics that represent the intensity, duration, and frequency of human movement. Specifically, the programs focus on operationalizing key components of PA measurement, including valid wear periods, total activity counts, and time spent in various intensity levels (e.g., sedentary, light, moderate-to-vigorous physical activity), thereby providing standardized variables for downstream analyses.

Validity

Psychometric validity metrics are not applicable to this resource, as it is a suite of programming code designed for data manipulation and variable creation, rather than a psychometric assessment instrument itself. The validity of the output relies on the underlying quality and calibration of the accelerometer device used in the NHANES 2003–2004 survey and the appropriateness of the specific algorithms and cut-points implemented within the SAS Programs for defining activity and non-wear time. Researchers must ensure that the methodological definitions embedded in the code (e.g., criteria for a valid day or non-wear time) align with contemporary research standards.

Reliability

Traditional psychometric reliability statistics (e.g., test-retest reliability, internal consistency) are not applicable to this technical tool. The concept of reliability, in this context, refers to the consistency and robustness of the SAS Programs execution. When executed correctly and consistently, the programs reliably apply standardized edits and calculations to the input data, ensuring the replicability of the defined summary measures across different researchers or instances of data processing, provided the input data remains identical.

Factor Analysis

Factor analysis is not applicable. This statistical technique is typically used to identify underlying dimensions or latent factors within a set of questionnaire items or observed variables. The NCI SAS Programs serve the upstream function of preparing and summarizing objective physical activity variables from raw data. These standardized variables may subsequently be used as input variables in factor analytic studies conducted by researchers investigating patterns of physical activity behavior.

Instrument

Test Type: Data Processing and Standardization Tool

Format: Statistical Programming Code (SAS Language)

Language Available: English (Code Comments and Documentation)

Population Group: Researchers and analysts working with public health survey data, specifically related to physical activity measurement.

Age Group: Data derived from participants of the NHANES 2003–2004 survey cycle, which includes various age groups depending on the specific accelerometer subsample.

Population Details: The programs are specifically tailored to process objective physical activity data collected from a nationally representative sample of the U.S. population participating in NHANES.

Test Methodology: The methodology involves the application of standardized algorithms and editing rules coded in SAS Programs. This includes converting raw ActiGraph counts into standard NHANES data format, identifying non-wear time based on prolonged periods of zero counts, and calculating time spent in defined activity intensity categories to create standardized summary variables.

Keywords

Physical Activity Measurement, Data Editing, Statistical Software, NHANES 2003-2004, ActiGraph, Wear Time, Epidemiology, SAS Programming.

Authors

Author ORCID Identifier: Not applicable (Organizational Author: National Cancer Institute)

Affiliation Email addresses: Contact information is typically provided on the official NCI resource page.

Correspondence Address: National Cancer Institute (NCI), Division of Cancer Control and Population Sciences (DCCPS), Bethesda, MD, USA.

Permissions & Fee and Test Year

The SAS Programs are provided by the National Cancer Institute as a public service resource. They are generally available for free download and use by the research community, although users must adhere to any specific terms of use outlined on the NCI website regarding citation and modification. Technical support from NCI is explicitly stated as limited, necessitating intermediate-level SAS programming skills for effective utilization and customization.

The programs were created to analyze data collected during the NHANES 2003–2004 cycle, serving as the foundational reference period for the specific editing rules embedded within the code. While the code is adaptable, its original parameters are benchmarked against this specific survey period.

Reference’s

The official source and documentation for the programs are maintained by the National Cancer Institute.

• National Cancer Institute. (n.d.). SAS Programs to Analyze NHANES Accelerometer Data. Retrieved from: https://epi.grants.cancer.gov/nhanes-pam/

Items of the SAS Programs to Analyze NHANES Accelerometer Data

IMPORTANT: The following scale items must be preserved in their original language and must not be changed in any way.

These are not conventional scale items but rather the functional components provided by the software package, available as executable files at the NCI link provided above. The programs include:

• Programs designed to perform specific data edits on raw NHANES 2003-2004 accelerometer data.

• Programs for the creation of per-day summary variables.

• Programs for the creation of per-person summary variables.

• A utility program that converts data downloaded from ActiGraph accelerometers to the data layout used in NHANES, enabling the preparation programs to be used on ActiGraph data as well.