Consequences of Choosing Different Settings When Processing Hip-Based Accelerometry Data From Older Adults: A Practical Approach Using Baseline Data From the SITLESS Study

Jason Wilson, Mathias Skjødt, Ilona Mc Mullan, Nicole Blackburn, Maria Giné-Garriga, Oriol Sansano-Nadal, Marta Roqué i Figuls, Jochen Klenk, Dhayana Dallmeier, Emma McIntosh, Manuela Deidda, Mark Tully, Paolo Caserotti

Research output: Contribution to journalArticle

Abstract

Accurately measuring older adults’ physical activity (PA) and sedentary behavior (SB) using accelerometers is essential, as both are important markers of health. This study aimed to highlight how steps taken during data processing may affect key hip-based accelerometry outcomes in older adults, using a selection of baseline accelerometry data (n = 658) from the SITLESS study. Different analytical parameters tested included wear-time algorithms, use of low-frequency extension (LFE) filter, epoch length, and minimum and maximum daily wear-time thresholds. These were compared against vertical axis counts per minute (CPM), vector magnitude (VM) CPM, SB, light PA, moderate-to-vigorous PA, step counts, and wear-time percentage. Differences in settings across the analytical parameters were assessed using paired sample t-tests and repeated measures ANOVAs using Bonferroni correction. Using the “Choi” versus “Troiano” wear-time algorithm resulted in a higher percentage wear-time. Most SB and PA outcomes were significantly different across wear-time algorithms (p < .001). This was similar when using the LFE filter versus normal filter (p < .001). Using 10-second epoch length increased daily SB time (between +75.7 and +79.2 minutes) compared to 60-second. Most SB and PA outcomes significantly changed comparing minimum-wear-time thresholds of 360, 480, 600, and 720 minutes per day (p < .001). Applying a log-diary with a ≥1140-minute threshold had a significant impact on vertical axis CPM, VM CPM, SB, and light PA outcomes (p < .001). This study demonstrates the potential variability in the number of participants being included in studies and reported SB and PA levels when processing older adults’ accelerometry data dependent on the analytical procedures utilized.
LanguageEnglish
Pages1
Number of pages11
JournalJournal for the Measurement of Physical Behaviour
Volume3
Issue number2
Early online date29 Jan 2020
DOIs
Publication statusE-pub ahead of print - 29 Jan 2020

Fingerprint

Accelerometry
Hip
Light
Analysis of Variance

Keywords

  • Accelerometer
  • Actigraph
  • Measurement
  • Methodology
  • Physical activity
  • Sedentary behavior

Cite this

Wilson, Jason ; Skjødt, Mathias ; Mc Mullan, Ilona ; Blackburn, Nicole ; Giné-Garriga, Maria ; Sansano-Nadal, Oriol ; Roqué i Figuls, Marta ; Klenk, Jochen ; Dallmeier, Dhayana ; McIntosh, Emma ; Deidda, Manuela ; Tully, Mark ; Caserotti, Paolo. / Consequences of Choosing Different Settings When Processing Hip-Based Accelerometry Data From Older Adults: A Practical Approach Using Baseline Data From the SITLESS Study. In: Journal for the Measurement of Physical Behaviour. 2020 ; Vol. 3, No. 2. pp. 1.
@article{532cc5d7352e4dd99a82b17f41a16ea8,
title = "Consequences of Choosing Different Settings When Processing Hip-Based Accelerometry Data From Older Adults: A Practical Approach Using Baseline Data From the SITLESS Study",
abstract = "Accurately measuring older adults’ physical activity (PA) and sedentary behavior (SB) using accelerometers is essential, as both are important markers of health. This study aimed to highlight how steps taken during data processing may affect key hip-based accelerometry outcomes in older adults, using a selection of baseline accelerometry data (n = 658) from the SITLESS study. Different analytical parameters tested included wear-time algorithms, use of low-frequency extension (LFE) filter, epoch length, and minimum and maximum daily wear-time thresholds. These were compared against vertical axis counts per minute (CPM), vector magnitude (VM) CPM, SB, light PA, moderate-to-vigorous PA, step counts, and wear-time percentage. Differences in settings across the analytical parameters were assessed using paired sample t-tests and repeated measures ANOVAs using Bonferroni correction. Using the “Choi” versus “Troiano” wear-time algorithm resulted in a higher percentage wear-time. Most SB and PA outcomes were significantly different across wear-time algorithms (p < .001). This was similar when using the LFE filter versus normal filter (p < .001). Using 10-second epoch length increased daily SB time (between +75.7 and +79.2 minutes) compared to 60-second. Most SB and PA outcomes significantly changed comparing minimum-wear-time thresholds of 360, 480, 600, and 720 minutes per day (p < .001). Applying a log-diary with a ≥1140-minute threshold had a significant impact on vertical axis CPM, VM CPM, SB, and light PA outcomes (p < .001). This study demonstrates the potential variability in the number of participants being included in studies and reported SB and PA levels when processing older adults’ accelerometry data dependent on the analytical procedures utilized.",
keywords = "Accelerometer, Actigraph, Measurement, Methodology, Physical activity, Sedentary behavior",
author = "Jason Wilson and Mathias Skj{\o}dt and {Mc Mullan}, Ilona and Nicole Blackburn and Maria Gin{\'e}-Garriga and Oriol Sansano-Nadal and {Roqu{\'e} i Figuls}, Marta and Jochen Klenk and Dhayana Dallmeier and Emma McIntosh and Manuela Deidda and Mark Tully and Paolo Caserotti",
year = "2020",
month = "1",
day = "29",
doi = "https://doi.org/10.1123/jmpb.2019-0037",
language = "English",
volume = "3",
pages = "1",
journal = "Journal for the Measurement of Physical Behaviour",
issn = "2575-6613",
number = "2",

}

Consequences of Choosing Different Settings When Processing Hip-Based Accelerometry Data From Older Adults: A Practical Approach Using Baseline Data From the SITLESS Study. / Wilson, Jason; Skjødt, Mathias; Mc Mullan, Ilona; Blackburn, Nicole; Giné-Garriga, Maria; Sansano-Nadal, Oriol; Roqué i Figuls, Marta; Klenk, Jochen; Dallmeier, Dhayana; McIntosh, Emma; Deidda, Manuela; Tully, Mark; Caserotti, Paolo.

In: Journal for the Measurement of Physical Behaviour, Vol. 3, No. 2, 30.06.2020, p. 1.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Consequences of Choosing Different Settings When Processing Hip-Based Accelerometry Data From Older Adults: A Practical Approach Using Baseline Data From the SITLESS Study

AU - Wilson, Jason

AU - Skjødt, Mathias

AU - Mc Mullan, Ilona

AU - Blackburn, Nicole

AU - Giné-Garriga, Maria

AU - Sansano-Nadal, Oriol

AU - Roqué i Figuls, Marta

AU - Klenk, Jochen

AU - Dallmeier, Dhayana

AU - McIntosh, Emma

AU - Deidda, Manuela

AU - Tully, Mark

AU - Caserotti, Paolo

PY - 2020/1/29

Y1 - 2020/1/29

N2 - Accurately measuring older adults’ physical activity (PA) and sedentary behavior (SB) using accelerometers is essential, as both are important markers of health. This study aimed to highlight how steps taken during data processing may affect key hip-based accelerometry outcomes in older adults, using a selection of baseline accelerometry data (n = 658) from the SITLESS study. Different analytical parameters tested included wear-time algorithms, use of low-frequency extension (LFE) filter, epoch length, and minimum and maximum daily wear-time thresholds. These were compared against vertical axis counts per minute (CPM), vector magnitude (VM) CPM, SB, light PA, moderate-to-vigorous PA, step counts, and wear-time percentage. Differences in settings across the analytical parameters were assessed using paired sample t-tests and repeated measures ANOVAs using Bonferroni correction. Using the “Choi” versus “Troiano” wear-time algorithm resulted in a higher percentage wear-time. Most SB and PA outcomes were significantly different across wear-time algorithms (p < .001). This was similar when using the LFE filter versus normal filter (p < .001). Using 10-second epoch length increased daily SB time (between +75.7 and +79.2 minutes) compared to 60-second. Most SB and PA outcomes significantly changed comparing minimum-wear-time thresholds of 360, 480, 600, and 720 minutes per day (p < .001). Applying a log-diary with a ≥1140-minute threshold had a significant impact on vertical axis CPM, VM CPM, SB, and light PA outcomes (p < .001). This study demonstrates the potential variability in the number of participants being included in studies and reported SB and PA levels when processing older adults’ accelerometry data dependent on the analytical procedures utilized.

AB - Accurately measuring older adults’ physical activity (PA) and sedentary behavior (SB) using accelerometers is essential, as both are important markers of health. This study aimed to highlight how steps taken during data processing may affect key hip-based accelerometry outcomes in older adults, using a selection of baseline accelerometry data (n = 658) from the SITLESS study. Different analytical parameters tested included wear-time algorithms, use of low-frequency extension (LFE) filter, epoch length, and minimum and maximum daily wear-time thresholds. These were compared against vertical axis counts per minute (CPM), vector magnitude (VM) CPM, SB, light PA, moderate-to-vigorous PA, step counts, and wear-time percentage. Differences in settings across the analytical parameters were assessed using paired sample t-tests and repeated measures ANOVAs using Bonferroni correction. Using the “Choi” versus “Troiano” wear-time algorithm resulted in a higher percentage wear-time. Most SB and PA outcomes were significantly different across wear-time algorithms (p < .001). This was similar when using the LFE filter versus normal filter (p < .001). Using 10-second epoch length increased daily SB time (between +75.7 and +79.2 minutes) compared to 60-second. Most SB and PA outcomes significantly changed comparing minimum-wear-time thresholds of 360, 480, 600, and 720 minutes per day (p < .001). Applying a log-diary with a ≥1140-minute threshold had a significant impact on vertical axis CPM, VM CPM, SB, and light PA outcomes (p < .001). This study demonstrates the potential variability in the number of participants being included in studies and reported SB and PA levels when processing older adults’ accelerometry data dependent on the analytical procedures utilized.

KW - Accelerometer

KW - Actigraph

KW - Measurement

KW - Methodology

KW - Physical activity

KW - Sedentary behavior

U2 - https://doi.org/10.1123/jmpb.2019-0037

DO - https://doi.org/10.1123/jmpb.2019-0037

M3 - Article

VL - 3

SP - 1

JO - Journal for the Measurement of Physical Behaviour

T2 - Journal for the Measurement of Physical Behaviour

JF - Journal for the Measurement of Physical Behaviour

SN - 2575-6613

IS - 2

ER -