Database Open Access
Minute level step counts and physical activity data from the National Health and Nutrition Examination Survey (NHANES) 2011-2014
Published: Jan. 28, 2025. Version: 1.0.0
When using this resource, please cite:
(show more options)
Koffman, L., & Muschelli, J. (2025). Minute level step counts and physical activity data from the National Health and Nutrition Examination Survey (NHANES) 2011-2014 (version 1.0.0). PhysioNet. https://doi.org/10.13026/ah0j-3z47.
Please include the standard citation for PhysioNet:
(show more options)
Goldberger, A., Amaral, L., Glass, L., Hausdorff, J., Ivanov, P. C., Mark, R., ... & Stanley, H. E. (2000). PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals. Circulation [Online]. 101 (23), pp. e215–e220.
Abstract
The National Health and Nutrition Examination Survey (NHANES) is a nationally representative study that collects demographic, socioeconomic, dietary, and health-related information from 10,000 Americans annually. In Wave G (2011-2012) and Wave H (2013-2014), participants wore an ActiGraph GT3X+ on the non-dominant wrist for seven consecutive days. Publicly available data include minute-level wear predictions (wake wear, sleep wear, unknown, nonwear), data quality flags, and monitor-independent movement summary units (MIMS). Additionally, raw triaxial accelerometer data recorded at 80 Hz are provided, enabling detailed analyses beyond MIMS.
We applied five step counting algorithms and the ActiGraph Activity Count (AC) algorithm to the raw data to create a dataset of minute-level step counts. The dataset also includes minute level AC, MIMS, wear predictions, and wear flags for all participants who wore accelerometers in NHANES 2011–2014. These data facilitate detailed investigations of accelerometry-measured physical activity patterns in a nationally representative U.S. population. Potential analyses include examination of within-day activity patterns and their association with health outcomes.
Background
The National Health and Nutrition Examination Survey (NHANES) is an ongoing study of the health and nutritional status of adults and children in the United States. Data from the study are publicly available and include information from interviews, physical examinations, and laboratory tests from a nationally representative sample of Americans [1].
Objective physical activity measurement was first introduced in NHANES from 2003-2006 cycles through hip-worn accelerometers, enabling detailed insights into physical activity patterns. In 2011, NHANES reintroduced physical activity monitoring with the wrist-worn ActiGraph GT3X+, a triaxial accelerometer that measures acceleration in three dimensions at a frequency of 80 Hz (every 1/80th of a second). Participants were given the devices after their visit at the NHANES Mobile Examination Center and were asked to wear the devices for seven full days. On the morning of the ninth day, they mailed the devices back to NHANES [2]. The raw 80 Hz accelerometer data was released in 2022 [3], which enabled the application of algorithms specifically developed for raw accelerometry. The availability of raw data facilitates comparability and reproducibility across studies utilizing accelerometry.
Methods
Study participants and protocol
Participants aged six and older (2011-2012) and aged three and older (2013-2014) received an ActiGraph GT3X+ accelerometer at the Mobile Examination Center visit and were asked to wear the device on the wrist for 7 consecutive days following the visit. The devices collect acceleration in gravitational units (g, 9.81 ) in three dimensions every 1/80th of a second (80Hz). 6,917 individuals in 2011-2012 and 7,776 individuals in 2012-2013 received the accelerometers and had raw data files.
Raw data processing
Raw data in .tar.bz2 format were downloaded from the NHANES website [3]. Information related to physical activity (i.e. 80Hz acceleration and log files) was extracted and converted to CSV files.
Algorithm deployment
Four open source step counting algorithms: ADEPT [4], Oak [5], stepcount [6] (self-supervised learner [SSL] and random forest versions), and Verisense [7] (original and revised versions), and one proprietary step counting algorithm – Actilife [8] – were deployed on the raw data to obtain minute level step counts. Additionally, ActiGraph's Activity Counts [9] (AC) algorithm was deployed on the raw data.
Monitor Independent Movement Summary [10] (MIMS) units, wear prediction [11] (one of wake wear, sleep wear, unknown, or non-wear), and wear flags (yes/no) were provided by NHANES at the minute level in subject-specific files [2].
Algorithm notes
1. The parameters used for ADEPT are:
sim_MIN = 0.6
dur_MIN = 0.8
dur_MAX = 1.4
ptp_r_MIN = 0.5
ptp_r_MAX = 2
vmc_r_MIN = 0.05
vmc_r_MAX = 0.5
mean_abs_diff_med_p_MAX = 0.7
mean_abs_diff_med_t_MAX = 0.2
mean_abs_diff_dur_MAX = 0.3
2. The parameters used for Oak are:
sample_rate_analysis = 10
min_amplitude = 0.3
step_frequency = c(1.4, 2.3)
alpha = 0.6
beta = 2.5
min_duration_peak = 3
delta = 20
4. Default parameters were used for both the SSL and random forest versions of stepcount.
3. Before deploying Verisense, data were resampled to 15Hz using linear interpolation because the Verisense parameters are hard coded assuming a sample rate of 15Hz. The parameters used for Verisense original are:
window_size = 3
periodicity_range = c(5, 15)
similarity_threshold = -0.5
continuity_window_size = 4
continuity_threshold = 4
variance_threshold = 0.001
vm_threshold = 1.2
global_vm_threshold = 0.025
The parameters used for verisense revised are:
window_size = 4
periodicity_range = c(4, 20)
similarity_threshold = -1
continuity_window_size = 4
continuity_threshold = 4
variance_threshold = 0.01
vm_threshold = 1.25
Data Description
Data contain observations for 14,693 individuals (6,917 from 2011-2012 and 7,776 from 2013-2014). Individuals are 51% female, mean (SD) age 35.8 (23.2).
Data description is as follows and can also be found in the readme.
Overview
The project files include:
subject-info.csv
: NHANES subject ID (SEQN) and other demographic variables provided by NHANES including gender, age at screening, and survey weights.-
13 files, each named
nhanes_1440_<varname>.csv.xz
-
<varname>
is one of:actisteps
,adeptsteps
,oaksteps
,scrfsteps
,scsslsteps
,vssteps
,vsrevsteps
,AC
,log10AC
,PAXMTSM
,log10PAXMTSM
,PAXPREDM
,PAXFLGSM
-
Each row in each file is one day for one participant. Each file contains the following columns:
-
SEQN
: NHANES participant ID, a character scalar -
PAXDAYM
: NHANES day of physical activity measurements for the participant, integer between 1 and 9. Note: days 1 and 9 will not have complete data. -
PAXDAYWM
: day of the week, integer between 1 and 7, where 1 corresponds to Sunday, 2 to Monday, …, and 7 to Saturday. -
min_x
forx = 1, 2, ..., 1440
: the value of<varname>
for minutex
. Foractisteps
,adeptsteps
,oaksteps
,scrfsteps
,scsslsteps
,vssteps
,vsrevsteps
,AC
,log10AC
,PAXMTSM
,log10PAXMTSM
, the column values are floats. ForPAXPREDM
they are integers, where1
= wake wear,2
= sleep wear,3
= unknown wear, and4
= nonwear. ForPAXFLGSM
the values are logical, whereTRUE
corresponds to any wear flags andFALSE
corresponds to no wear flags.
Key
-
actisteps
= Actilife steps -
oaksteps
= Oak steps -
scrfsteps
= stepcount random forest steps -
scsslsteps
= stepcount SSL steps -
vssteps
= Verisense steps -
vsrevsteps
= Verisense revised steps -
AC
= ActiGraph activity counts -
log10AC
= log10 of 1 + activity counts -
PAXMTSM
= MIMS -
log10PAXMTSM
= log10 of 1 + MIMS -
PAXPREDM
= wear prediction (1 = Wake wear, 2 = Sleep wear, 3 = Non wear, 4 = unknown) -
PAXFLGSM
= data quality flags (TRUE if flag, FALSE if otherwise)
Usage Notes
Reuse potential and limitations: Minute level physical activity data can be used along with the other rich demographic and health information provided by NHANES to investigate the association between various summaries of physical activity and age, sex, mortality, and other comorbidities, as has been done in NHANES 2003-2006 [12]. The data are also provided in a format that lends itself to functional data analysis (FDA) [13]; FDA methods have been previously used on minute-level MIMS but not step counts. Researchers should be sure to define inclusion criteria for the accelerometry data (i.e. define a valid day as 10 hours of wake wear and require at least 3 valid days for inclusion) and to take into account NHANES complex survey structure.
Code, tools, and manuscript: In an accompanying paper [14], we use this data to investigate the distribution of step counts by age, agreement between different step counting algorithms, and the association of step counts with mortality. Our code to perform this analysis is publicly available on GitHub [15].
Ethics
The authors declare no ethical concerns. The data abide by the data user agreement from the CDC/NCHS.
Conflicts of Interest
The authors declare no conflicts of interest that could influence this work.
References
- National Center for Health Statistics. National Health and Nutrition Examination Survey. Atlanta, GA: Centers for Disease Control and Prevention; 2023. Available from: https://www.cdc.gov/nchs/nhanes/index.htm
- National Center for Health Statistics. NHANES 2011-2012: Physical Activity Monitor (PAXMIN_G) Data Documentation. Atlanta, GA: Centers for Disease Control and Prevention; 2017. Available from: https://wwwn.cdc.gov/Nchs/Data/Nhanes/Public/2011/DataFiles/PAXMIN_G.htm
- National Center for Health Statistics. NHANES 2011-2012: Physical Activity Monitor (PAX80_G) Raw Accelerometer Data Documentation. Atlanta, GA: Centers for Disease Control and Prevention; 2017. Available from: https://wwwn.cdc.gov/Nchs/Data/Nhanes/Public/2011/DataFiles/PAX80_G.htm
- Karas M, Straczkiewicz M, Fadel W, Harezlak J, Crainiceanu CM, Urbanek JK. Adaptive empirical pattern transformation (ADEPT) with application to walking stride segmentation. Biostatistics. 2021;22(2):331–47. Available from: https://doi.org/10.1093/biostatistics/kxz033
- Straczkiewicz M, Huang EJ, Onnela JP. A “one-size-fits-most” walking recognition method for smartphones, smartwatches, and wearable accelerometers. NPJ Digit Med. 2023;6:29. Available from: https://doi.org/10.1038/s41746-022-00745-z
- Small SR, Chan S, Walmsley R, et al. Development and validation of a machine learning wrist-worn step detection algorithm with deployment in the UK Biobank. medRxiv. 2023. Available from: https://doi.org/10.1101/2023.02.20.23285750
- Rowlands AV, Maylor B, Dawkins NP, Dempsey PC, Edwardson CL, Soczawa-Stronczyk AA, et al. Stepping up with GGIR: validity of step cadence derived from wrist-worn research-grade accelerometers using the Verisense step count algorithm. J Sports Sci. 2022;40(19):2182–90. Available from: https://doi.org/10.1080/02640414.2022.2147134
- ActiLife Software. (2015). ActiGraph LLC.
- Neishabouri A, Nguyen J, Samuelsson J, et al. Quantification of acceleration as activity counts in ActiGraph wearable. Sci Rep. 2022;12:11958. Available from: https://doi.org/10.1038/s41598-022-16003-x
- John D, Tang Q, Albinali F, Intille S. An open-source monitor-independent movement summary for accelerometer data processing. J Meas Phys Behav. 2019;2(4):268–81. Available from: https://doi.org/10.1123/jmpb.2018-0068
- Thapa-Chhetry B, Arguello DJ, John D, Intille S. Detecting sleep and nonwear in 24-h wrist accelerometer data from the National Health and Nutrition Examination Survey. Med Sci Sports Exerc. 2022;54(11):1936–46. Available from: https://doi.org/10.1249/MSS.0000000000002973
- Leroux A, Di J, Smirnova E, McGuffey EJ, Cao Q, Bayatmokhtari E, et al. Organizing and analyzing the activity data in NHANES. Stat Biosci. 2019;11(2):262–87. Available from: https://doi.org/10.1007/s12561-018-09229-9
- Crainiceanu CM, Goldsmith J, Leroux A, Cui E. Functional Data Analysis with R. 1st ed. Boca Raton (FL): Chapman and Hall/CRC; 2024. doi:10.1201/9781003278726.
- Koffman L, Crainiceanu C, Muschelli J. Comparing step counting algorithms for high-resolution wrist accelerometry data in NHANES 2011-2014. Med Sci Sports Exerc. 2024; Available from: https://doi.org/10.1249/MSS.0000000000003616
- Koffman L, Muschelli J. NHANES 2011-2014 step counts [computer software]. 2024. Available from: https://github.com/lilykoff/nhanes_steps_mortality
Access
Access Policy:
Anyone can access the files, as long as they conform to the terms of the specified license.
License (for files):
Creative Commons Zero 1.0 Universal Public Domain Dedication
Discovery
DOI (version 1.0.0):
https://doi.org/10.13026/ah0j-3z47
DOI (latest version):
https://doi.org/10.13026/5htm-7515
Topics:
accelerometry
physical activity
nhanes
steps
Project Website:
https://github.com/lilykoff/nhanes_steps_mortality
Corresponding Author
Files
Total uncompressed size: 2.3 GB.
Access the files
- Download the ZIP file (2.3 GB)
-
Download the files using your terminal:
wget -r -N -c -np https://physionet.org/files/minute-level-step-count-nhanes/1.0.0/
Name | Size | Modified |
---|---|---|
LICENSE.txt (download) | 6.5 KB | 2025-01-28 |
SHA256SUMS.txt (download) | 1.4 KB | 2025-01-28 |
data_README.md (download) | 2.1 KB | 2024-12-10 |
nhanes_1440_AC.csv.xz (download) | 696.2 MB | 2024-09-16 |
nhanes_1440_PAXFLGSM.csv.xz (download) | 843.3 KB | 2024-09-16 |
nhanes_1440_PAXMTSM.csv.xz (download) | 248.7 MB | 2024-09-16 |
nhanes_1440_PAXPREDM.csv.xz (download) | 9.5 MB | 2024-09-16 |
nhanes_1440_actisteps.csv.xz (download) | 70.2 MB | 2024-09-16 |
nhanes_1440_adeptsteps.csv.xz (download) | 104.6 MB | 2024-09-16 |
nhanes_1440_log10AC.csv.xz (download) | 668.2 MB | 2024-09-16 |
nhanes_1440_log10PAXMTSM.csv.xz (download) | 314.2 MB | 2024-09-16 |
nhanes_1440_oaksteps.csv.xz (download) | 110.4 MB | 2024-09-16 |
nhanes_1440_scrfsteps.csv.xz (download) | 36.1 MB | 2024-09-16 |
nhanes_1440_scsslsteps.csv.xz (download) | 30.0 MB | 2024-09-16 |
nhanes_1440_vsrevsteps.csv.xz (download) | 49.9 MB | 2024-09-16 |
nhanes_1440_vssteps.csv.xz (download) | 57.5 MB | 2024-09-16 |
subject-info.csv (download) | 926.3 KB | 2024-09-16 |